{"initiatives": [{"ID": "AI-001", "Company": "Chevron", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "APOLO prospectivity and development optimization", "Extracted_finding": "Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.", "Stage_at_disclosure": "Launched Aug 2024 (pre-window); substantive Investor Day update 2025-11-12", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Model launch date Aug 2024 is before window; Nov 2025 investor day is qualifying update. No attributable APOLO production uplift or ROI disclosed.", "Source_IDs": "S01", "Source_date": "2025-11-12", "Source_type": "investor presentation", "Source_locator": "Slides 39–40; transcript pp. 21–22", "Source_URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-11-12; see chronology for operating dates.", "Context": "Traditional type curves produce coarse, binary prospectivity views from limited variables, and development choices interact across spacing, completion and timing.", "Implementation": "APOLO analyzes more than 50,000 wells to map prospectivity at finer resolution and recovery variation. Chevron was training it to optimize well spacing, completion design and development timing; decision support spans exploration screening and shale/tight development.", "Assessment": "Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.", "Detail_words": 71, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-001", "Initiative_ID": "AI-001", "Title": "APOLO well planning", "Date": "2025-11-12", "Finding": "Machine learning guides prospectivity; development extends to spacing, completions and timing.", "Metric": "Training data: >50,000 wells", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Chevron described training on more than 50,000 wells and extending the approach to spacing, completion design and development timing.", "Context": "Well planning must account for interacting geology and development choices. APOLO uses machine learning to make prospectivity assessment more granular than conventional type-curve groupings.", "Assessment": "A larger training set establishes scale of inputs. It does not establish forecast accuracy on new acreage or incremental recovery. Development decisions still require economic and geological judgment."}], "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-001"], "Research_search": " {\"key\": \"chevron-apolo\", \"existing_ids\": [\"AI-001\"], \"related_existing_ids\": [\"AI-002\"], \"company\": \"Chevron\", \"title\": \"APOLO prospectivity and development optimization\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Launched Aug 2024 (pre-window); substantive Investor Day update 2025-11-12\", \"practical_relevance\": \"Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.\", \"business_problem\": \"Traditional type curves produce coarse, binary prospectivity views from limited variables, and development choices interact across spacing, completion and timing.\", \"workflow\": \"APOLO analyzes more than 50,000 wells to map prospectivity at finer resolution and recovery variation. Chevron was training it to optimize well spacing, completion design and development timing; decision support spans exploration screening and shale/tight development.\", \"technology_data\": \"Multivariate machine learning on well data; Chevron does not name training/validation design or integration details.\", \"integration_human_role\": \"Subsurface and development specialists use mapped candidates; company says still training/scaling. Human review/approval is not specified in disclosure.\", \"deployment_scope\": \"Chevron Permian/shale and tight use described; scale across portfolio is a goal. This is an in-window update about a model launched Aug 2024, so launch predates window.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No quantified recovery uplift. Management’s broad “billions” annual AI value by decade-end is potential, not APOLO realized savings.\", \"timeline\": [{\"date\": \"2025-11-12\", \"event\": \"APOLO prospectivity and development optimization documented or disclosed.\", \"source_ids\": [\"S01\"]}], \"limitations\": [\"Model launch date Aug 2024 is before window; Nov 2025 investor day is qualifying update.\", \"No attributable APOLO production uplift or ROI disclosed.\"], \"assessment\": \"Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.\", \"sources\": [{\"id\": \"S01\", \"title\": \"2025 Chevron Investor Day edited transcript\", \"url\": \"https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4\", \"publisher\": \"Chevron\", \"date\": \"2025-11-12\", \"type\": \"investor presentation\", \"locator\": \"Slides 39–40; transcript pp. 21–22\", \"review_status\": \"full text reviewed\", \"supports\": [\"APOLO launched Aug 2024 (pre-window); Nov 2025 update describes >50,000 wells, finer prospectivity mapping and model training for spacing, completion and development timing.\", \"Chevron described an AI team and prioritization of high-value workflows including shale/tight and exploration.\", \"Chevron’s “billions” annual value is a potential decade-end ambition, not realized savings.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-001\"}], \"field_sources\": {\"business_problem\": [\"S01\"], \"workflow\": [\"S01\"], \"technology_data\": [\"S01\"], \"integration_human_role\": [\"S01\"], \"deployment_scope\": [\"S01\"], \"economics\": [\"S01\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-001\", \"source_ids\": [\"S01\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-001\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-001", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. Chevron's investor-day transcript returned HTTP 403 on 5 Oct 2026.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-002", "Company": "Chevron", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Exploration workflows", "Extracted_finding": "AI supports seismic interpretation, inventory assessment and prospect decisions.", "Stage_at_disclosure": "In use", "Reported_metric": "Not separately disclosed", "Interpretation_limit": "Portfolio-wide financial benefit not isolated.", "Source_IDs": "S01", "Source_date": "2025-11-12", "Source_type": "Investor day transcript", "Source_locator": "PDF pp. 21–23, 27–28, 33–34; slides 38–39; Jeff Gustavson / Mark Nelson", "Source_URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Chevron places AI within exploration work, where interpreting seismic and evaluating inventory inform which prospects receive further capital.", "Implementation": "The investor-day discussion links AI to interpretation and prospect decisions. High-performance computing and proprietary imaging also support the broader workflow.", "Assessment": "AI is one element of a technical workflow. Seismic computing capacity alone cannot demonstrate a better discovery rate, and the source does not isolate AI’s financial contribution.", "Detail_words": 65, "Primary_department": "Exploration & geoscience", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-002", "Initiative_ID": "AI-002", "Title": "Exploration workflows", "Date": "2025-11-12", "Finding": "AI supports seismic interpretation, inventory assessment and prospect decisions.", "Metric": "Not separately disclosed", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The investor-day discussion links AI to interpretation and prospect decisions. High-performance computing and proprietary imaging also support the broader workflow.", "Context": "Chevron places AI within exploration work, where interpreting seismic and evaluating inventory inform which prospects receive further capital.", "Assessment": "AI is one element of a technical workflow. Seismic computing capacity alone cannot demonstrate a better discovery rate, and the source does not isolate AI’s financial contribution."}], "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-002", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. Chevron's investor-day transcript returned HTTP 403 on 5 Oct 2026.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-003", "Company": "Chevron", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Workforce and workflows", "Extracted_finding": "Employee tools and workflow redesign support productivity and decisions, including abandonment preparation.", "Stage_at_disclosure": "In use", "Reported_metric": "Over half of employees have tool access", "Interpretation_limit": "Access is not active use.", "Source_IDs": "S01", "Source_date": "2025-11-12", "Source_type": "Investor day transcript", "Source_locator": "PDF pp. 21–23, 27–28, 33–34; slides 38–39; Jeff Gustavson / Mark Nelson", "Source_URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Chevron is combining employee tools with workflow redesign as part of a standardized operating model. More than half of employees had access at the disclosure date.", "Implementation": "A well-abandonment example connects an India-based team with Australian operations to prepare information, permit applications and orders overnight.", "Assessment": "The case illustrates cross-team preparation work. Access and a faster example do not establish sustained usage, a measured reduction in headcount, or enterprise-wide savings.", "Detail_words": 68, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-003", "Initiative_ID": "AI-003", "Title": "Workforce and workflows", "Date": "2025-11-12", "Finding": "Employee tools and workflow redesign support productivity and decisions, including abandonment preparation.", "Metric": "Over half of employees have tool access", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "A well-abandonment example connects an India-based team with Australian operations to prepare information, permit applications and orders overnight.", "Context": "Chevron is combining employee tools with workflow redesign as part of a standardized operating model. More than half of employees had access at the disclosure date.", "Assessment": "The case illustrates cross-team preparation work. Access and a faster example do not establish sustained usage, a measured reduction in headcount, or enterprise-wide savings."}], "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Workforce productivity / abandonment preparations / annual billions in AI value as concrete realized initiatives.", "reason": "Investor Day establishes broad AI ambition and value potential but does not quantify realized AI-only returns or provide a distinct named operational workflow for each legacy entry. APOLO is profiled separately.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}, {"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-003", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "over half of employees", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. Chevron's investor-day transcript returned HTTP 403 on 5 Oct 2026.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-004", "Company": "Chevron", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "AI value ambition", "Extracted_finding": "Management sees annual AI value beyond existing forecasts by decade-end.", "Stage_at_disclosure": "Forward-looking ambition", "Reported_metric": "Potential billions annually", "Interpretation_limit": "Not realized savings or guidance.", "Source_IDs": "S01", "Source_date": "2025-11-12", "Source_type": "Investor day transcript", "Source_locator": "PDF pp. 21–23, 27–28, 33–34; slides 38–39; Jeff Gustavson / Mark Nelson", "Source_URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Management described AI as an additional source of future value beyond the company’s existing forecasts.", "Implementation": "The ambition is potential annual value in the billions by the end of the decade. The disclosure does not provide a project-by-project bridge to that amount.", "Assessment": "Treat this as strategic upside. Adding it to a valuation would require timing, implementation costs, attribution and a definition of value. It should not be presented as realized savings.", "Detail_words": 70, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-004", "Initiative_ID": "AI-004", "Title": "AI value ambition", "Date": "2025-11-12", "Finding": "Management sees annual AI value beyond existing forecasts by decade-end.", "Metric": "Potential billions annually", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The ambition is potential annual value in the billions by the end of the decade. The disclosure does not provide a project-by-project bridge to that amount.", "Context": "Management described AI as an additional source of future value beyond the company’s existing forecasts.", "Assessment": "Treat this as strategic upside. Adding it to a valuation would require timing, implementation costs, attribution and a definition of value. It should not be presented as realized savings."}], "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Workforce productivity / abandonment preparations / annual billions in AI value as concrete realized initiatives.", "reason": "Investor Day establishes broad AI ambition and value potential but does not quantify realized AI-only returns or provide a distinct named operational workflow for each legacy entry. APOLO is profiled separately.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}, {"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-004", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. Chevron's investor-day transcript returned HTTP 403 on 5 Oct 2026.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-005", "Company": "Chevron", "Segment": "Integrated", "Theme": "AI infrastructure", "Initiative": "West Texas data-center power", "Extracted_finding": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Stage_at_disclosure": "PPA signed; FID expected end-2026", "Reported_metric": "~2.67 GW; 20-year PPA; first power targeted 2028", "Interpretation_limit": "PPA signed; FID and operation not established.", "Source_IDs": "S01; S44", "Source_date": "2025-11-12 / 2026-06-22", "Source_type": "Investor day transcript", "Source_locator": "PDF pp. 21–23, 27–28, 33–34; slides 38–39; Jeff Gustavson / Mark Nelson", "Source_URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Implementation": "June 2026 update: approximately 2.67 GW, developed with Engine No. 1, using GE Vernova and Solar turbines. Chevron expected FID by year-end 2026 and first power in 2028.", "Assessment": "The signed PPA advances commercial commitment, while investment approval and construction remain separate milestones. The 2028 date supersedes the 2027 first-power target in the November 2025 investor discussion.", "Detail_words": 84, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-005", "Initiative_ID": "AI-005", "Title": "West Texas data-center power", "Date": "2025-11-12 / 2026-06-22", "Finding": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Metric": "~2.67 GW; 20-year PPA; first power targeted 2028", "Source_IDs": "S01; S44", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "June 2026 update: approximately 2.67 GW, developed with Engine No. 1, using GE Vernova and Solar turbines. Chevron expected FID by year-end 2026 and first power in 2028.", "Context": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Assessment": "The signed PPA advances commercial commitment, while investment approval and construction remain separate milestones. The 2028 date supersedes the 2027 first-power target in the November 2025 investor discussion."}], "Latest_disclosure": "2025-11-12 / 2026-06-22", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [{"claim": "West Texas data-center power was presented in original register alongside AI adoption.", "reason": "This is energy supply/infrastructure for AI demand, not company AI adoption; classify separately as AI-demand infrastructure, outside internal-use initiative count.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}, {"claim": "West Texas data-center power project classified as AI adoption.", "reason": "This is power infrastructure to supply AI demand; infrastructure for AI is not use of AI in Chevron operations.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}, {"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-005", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Chevron’s cited Project Kilby page could not be reopened; the direct URL returned 403.", "correction": "Retain prior verified project information only as previously reviewed; mark this supplemental source not reverified, not disproved. Its category remains AI-demand infrastructure, not AI adoption.", "source_url": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-006", "Company": "Devon Energy", "Segment": "E&P", "Theme": "Operational AI", "Initiative": "Autonomous artificial-lift optimization at operating wells", "Extracted_finding": "Operator reports 1,000 live closed-loop AI lift wells in August 2026. Prior pilot benefits do not establish fleet-wide results.", "Stage_at_disclosure": "Scaled. Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "Reported_metric": "1,000 live wells as of August 2026; >2,000 additional wells prospective. Separate 2025 gas-lift pilot reported 2–3% uplift in May 2026; not a measured result for the full August fleet.", "Interpretation_limit": "Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale. Do not attribute the company-wide business optimization program to AI.", "Source_IDs": "RS034; S02; R040; R041; R042; FU001; SW8-001", "Source_date": "2026-08-05", "Source_type": "Issuer-hosted earnings transcript", "Source_locator": "Prepared remarks p. 4; Q&A pp. 12–14", "Source_URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Evidence_status": "Issuer-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Latest operator disclosure 5 August 2026; older disclosures retained in history.", "Context": "Devon is expanding AI-enabled artificial lift optimization and analytics from trials into wider operations. The business objective is sustaining production and reducing operating intervention.", "Implementation": "Closed-loop AI adjusts lift settings in real time, 24 hours a day. Devon reports 1,000 operating wells, including first legacy Coterra wells; more than 2,000 additional Permian wells are prospective deployment opportunities. The Q2 disclosure does not provide a lift-type breakdown.", "Assessment": "Lift optimization concerns how existing wells produce. Devon’s broader $1 billion business-optimization program also includes other actions, so its savings cannot be assigned wholly to this AI initiative.", "Detail_words": 81, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Devon Energy", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-040", "REF-041", "REF-042"], "Related_IDs": ["AI-007"], "Disclosures": [{"ID": "REF-040", "Initiative_ID": "AI-006", "Title": "AI-enabled lift and maintenance drive $1B optimization plan", "Date": "2026-02", "Finding": "On the Q4 2025 call, Devon said it is taking AI-enabled artificial lift optimization and advanced analytics beyond pilots, and adding condition-based maintenance and faster drilling cycle times. Management said investments in AI and platforms built over the prior year were 'coming to fruition' in production. The work sits inside a $1B pre-tax business optimization program.", "Metric": "85% of $1B optimization target captured by YE2025; more than 100 active workstreams; downtime cut from about 7% to under 5%; base production about 5,000 BOE/d ahead of plan", "Source_IDs": "R040", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Track AI as named workstreams inside a cost program with a dollar target, so the impact shows up in guidance and not only in anecdotes.", "Partners": "", "Original_date": "2026-02", "Related_IDs": ["AI-007"]}, {"ID": "ORIG-AI-006", "Initiative_ID": "AI-006", "Title": "Artificial lift optimization", "Date": "2026-02-18", "Finding": "Expanding AI-enabled lift optimization and advanced analytics beyond pilots.", "Metric": "No AI-only savings disclosed", "Source_IDs": "S02", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The February 2026 call connects second-half 2025 trials, including gas lift, with broader rollout in the first and second quarters of 2026. Named vendors and fleet size are absent.", "Context": "Devon is expanding AI-enabled artificial lift optimization and analytics from trials into wider operations. The business objective is sustaining production and reducing operating intervention.", "Assessment": "Lift optimization concerns how existing wells produce. Devon’s broader $1 billion business-optimization program also includes other actions, so its savings cannot be assigned wholly to this AI initiative."}, {"ID": "REF-041", "Initiative_ID": "AI-006", "Title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "Date": "2026-05", "Finding": "Devon's Smart Gas Lift program pairs AI models with a physics-based calculation to set gas-lift injection rates in a closed loop directly at the wells. A 2025 pilot delivered a 2-3% uplift. Devon then moved to full implementation in the Delaware and is piloting similar AI models on ESPs and rod pumps.", "Metric": "Pilot uplift 2-3%; more than 850 wells live, heading to about 1,500; full-scale uplift said to exceed the pilot", "Source_IDs": "R041", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Closed-loop setpoint control on artificial lift, a physics model plus ML, is a proven first 'agentic' use case in production ops that scales well by well.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-042", "Initiative_ID": "AI-006", "Title": "AI speeds Coterra integration; in-house plus vendor model", "Date": "2026-08", "Finding": "CTO Trey Lowe said Devon 'leaned in really hard' on enterprise AI after ChatGPT launched and is merging its operations-focused AI stack with Coterra's ML well-prediction models. Closed-loop gas lift went from a few wells to about 1,000, and was extended to legacy Coterra wells. Service-company partners supply pieces of the solution, but Lowe called 'empowering our employees' the 'secret sauce'. AI tools also benchmark every rig and crew against a best-performance curve to find D&C gaps.", "Metric": "Closed-loop gas-lift wells rose from 'a few' to a couple of hundred (March 2026) to about 1,000; more than 2,000 more Permian candidates; new data tables built in 2 days instead of a couple of months", "Source_IDs": "R042", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "In M&A, a mature AI and data stack becomes a synergy engine: onboarding the acquired staff onto the tools is itself a value lever.", "Partners": "ChatDVN, Smart Gas Lift (in-house); unnamed service-provider partners", "Original_date": "2026-08", "Related_IDs": ["AI-007"]}, {"ID": "FU-DVN-01", "Initiative_ID": "AI-006", "Title": "Issuer confirmation of Smart Gas Lift deployment", "Date": "2026-05-06", "Finding": "Physics-informed AI calculates gas-lift injection settings and controls wells in a closed loop. Management reported more than 850 Delaware wells operating, with ESP and rod-pump models still in pilot.", "Metric": "2025 gas-lift pilot: approximately 2–3% production uplift; 850+ Delaware wells live by May 2026 (management-reported).", "Source_IDs": "FU001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "SW7-D-AI-006", "Initiative_ID": "AI-006", "Title": "Autonomous artificial-lift optimization at operating wells", "Date": "2026-05-06", "Finding": "AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.", "Metric": "May 2026 CEO remarks report >850 fully autonomous lift wells without a new quantified productivity result. Earlier 2025 pilot uplift remains in the original profile and is not remeasured here.", "Source_IDs": "FU001", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "OCT4-AI-006", "Initiative_ID": "AI-006", "Title": "Autonomous artificial-lift optimization at operating wells", "Date": "2026-08-05", "Finding": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "Metric": "1,000 live wells as of August 2026; >2,000 additional wells prospective. Separate 2025 gas-lift pilot reported 2–3% uplift in May 2026; not a measured result for the full August fleet.", "Source_IDs": "SW8-001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "October 4 evidence review"}], "Latest_disclosure": "2026-08-05", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct4-ai-006", "swarm7-ai-006", "north_america-001"], "Research_search": "{\"key\": \"devon-closed-loop-lift-ai\", \"existing_ids\": [\"AI-006\"], \"company\": \"Devon Energy\", \"title\": \"Closed-loop AI adjusts gas-lift settings at operating wells\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"850+ Delaware wells live as of May 2026; ESP and rod-pump pilots\", \"practical_relevance\": \"Closed-loop operating control at a disclosed multi-well scale; management-reported pilot results remain separate from full-rollout economics.\", \"business_problem\": \"Well conditions require recurring lift adjustment.\", \"workflow\": \"Physics-informed AI calculates gas-lift injection settings and controls wells in a closed loop. Management reported more than 850 Delaware wells operating, with ESP and rod-pump models still in pilot.\", \"technology_data\": \"Model, input data and validation are undisclosed.\", \"integration_human_role\": \"Control limits and override procedures were not disclosed.\", \"deployment_scope\": \"850+ Delaware wells; 1,500 portfolio wells targeted.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Gas-lift pilot production uplift\", \"value\": \"approximately 2–3%\", \"period\": \"2025 pilot; May 2026 disclosure\", \"attribution\": \"Management-reported\", \"source_ids\": [\"FU001\"]}, {\"metric\": \"Operating gas-lift scope\", \"value\": \"850+ Delaware wells\", \"period\": \"May 2026\", \"attribution\": \"Management-reported\", \"source_ids\": [\"FU001\"]}], \"economics\": \"The full-rollout uplift and AI-only cash benefit were not quantified.\", \"timeline\": [{\"date\": \"2026-05-06\", \"event\": \"Issuer call distinguishes pilot uplift, live gas-lift scope and additional lift pilots.\", \"source_ids\": [\"FU001\"]}, {\"date\": \"2026-08\", \"event\": \"CEO discussed 1,000-well scale and future rollout.\", \"source_ids\": [\"RS034\"]}], \"limitations\": [\"Pilot uplift is not the result for all 850+ wells. Broader rollout performance was not quantified; 1,500 wells was a target. Exclude the company-wide optimization program from AI-only benefits. Later secondary-hosted disclosures retain their source limits.\"], \"assessment\": \"Scaled control is established as a management claim; pilot economics cannot be extrapolated mechanically.\", \"sources\": [{\"id\": \"RS034\", \"title\": \"Devon Energy Q2 2026 Earnings Call Transcript\", \"url\": \"https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html\", \"publisher\": \"Yahoo Finance\", \"date\": \"2026-08\", \"type\": \"secondary\", \"locator\": \"CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"CEO reports 1,000 wells live; >2,000 more were potential rollout.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-003\"}, {\"id\": \"S02\", \"title\": \"Devon Q4 2025 Earnings Q&A Webcast Commentary Transcript\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"publisher\": \"Devon Energy\", \"date\": \"2026-02\", \"type\": \"earnings call\", \"locator\": \"Existing register cites second-half trials and 2026 rollout; PDF not independently opened in this pass\", \"review_status\": \"inaccessible\", \"supports\": [\"Existing profile claims second-half 2025 lift trials and Q1–Q2 2026 broader rollout; those details remain unverified in this review.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-024\"}, {\"id\": \"FU001\", \"title\": \"Devon Q1 2026 earnings call — Smart Gas Lift\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"publisher\": \"Devon Energy\", \"date\": \"2026-05-06\", \"type\": \"Issuer-hosted transcript\", \"locator\": \"Prepared remarks p. 4; Q&A pp. 12–14\", \"review_status\": \"full text reviewed\", \"supports\": [\"2025 gas-lift pilot: approximately 2–3% production uplift; 850+ Delaware wells live by May 2026 (management-reported).\"], \"accessed\": \"2026-09-28\"}], \"field_sources\": {\"business_problem\": [\"RS034\"], \"workflow\": [\"FU001\"], \"technology_data\": [\"RS034\"], \"integration_human_role\": [\"RS034\"], \"deployment_scope\": [\"FU001\"], \"economics\": [\"FU001\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Follow-up on 28 September 2026 reviewed the issuer-hosted Q1 2026 transcript, resolving the earlier secondary-only provenance limitation for the May gas-lift scope and pilot claims. Later secondary updates and the earlier Q4 source retain their separate access limitations. AI-007, AI-086 and AI-087 describe distinct workflows.\", \"related_existing_ids\": [\"AI-007\", \"AI-086\", \"AI-087\"], \"cohort\": \"north_america\", \"packet_id\": \"north_america-001\", \"source_ids\": [\"RS034\", \"S02\", \"FU001\"], \"report_date\": \"2026-09-28\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-006\"]} {\"key\": \"swarm7-ai-006\", \"packet_id\": \"swarm7-ai-006\", \"record_ids\": [\"AI-006\"], \"existing_ids\": [\"AI-006\"], \"related_existing_ids\": [], \"company\": \"Devon Energy\", \"title\": \"Autonomous artificial-lift optimization at operating wells\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Operating and scaled, as of May 6, 2026. >850 wells on fully autonomous artificial-lift optimization, per CEO on earnings call.\", \"practical_relevance\": \"AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.\", \"business_problem\": \"\", \"workflow\": \"AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.\", \"technology_data\": \"CEO describes AI translating drilling, completion and production data flows into actionable intelligence, and AI-created tools performing real-time artificial-lift optimization. No model architecture, lift types, control guardrails, or connectivity details disclosed.\", \"integration_human_role\": \"Engineering teams build/use tools; the reported 850-well subset is described as fully autonomous. The source does not specify exception handling or human approval steps.\", \"deployment_scope\": \"Operating and scaled, as of May 6, 2026. >850 wells on fully autonomous artificial-lift optimization, per CEO on earnings call.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"May 2026 CEO remarks report >850 fully autonomous lift wells without a new quantified productivity result. Earlier 2025 pilot uplift remains in the original profile and is not remeasured here.\", \"period\": \"Period as described in source; metric measurement period not separately disclosed.\", \"attribution\": \"Issuer-reported; no independent validation.\", \"source_ids\": [\"FU001\"]}], \"economics\": \"CEO separately linked AI value to drilling/completion time and lower capital costs, and cited business optimization savings; no amount is attributed specifically to AI or the lift tool.\", \"timeline\": [{\"date\": \"2026-05-06\", \"event\": \"Operating and scaled, as of May 6, 2026.\", \"source_ids\": [\"FU001\"]}], \"limitations\": [\"Company-level CEO assertion; no baseline, period, control group, production uplift, uptime, savings, lift mix, well list or attribution. Do not allocate Devon's $1bn business-optimization target/achievement to AI.\"], \"assessment\": \"AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.\", \"sources\": [{\"title\": \"Devon Energy Q1 2026 Earnings Call Corrected Transcript\", \"date\": \"2026-05-06\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"locator\": \"pp. 3-4 (PDF pp. 3-4), CEO Clay Gaspar prepared remarks, lines 108-128; fully read 20-page transcript\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"FU001\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"FU001\"], \"field_sources\": {\"workflow\": [\"FU001\"], \"technology_data\": [\"FU001\"], \"deployment_scope\": [\"FU001\"], \"practical_relevance\": [\"FU001\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Issuer-reported; no independent validation.\"} {\"key\": \"oct4-ai-006\", \"packet_id\": \"oct4-ai-006\", \"record_ids\": [\"AI-006\"], \"existing_ids\": [\"AI-006\"], \"related_existing_ids\": [], \"company\": \"Devon Energy\", \"title\": \"Autonomous artificial-lift optimization at operating wells\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.\", \"practical_relevance\": \"Operator reports 1,000 live closed-loop AI lift wells in August 2026. Prior pilot benefits do not establish fleet-wide results.\", \"business_problem\": \"\", \"workflow\": \"Closed-loop AI adjusts lift settings in real time, 24 hours a day. Devon reports 1,000 operating wells, including first legacy Coterra wells; more than 2,000 additional Permian wells are prospective deployment opportunities. The Q2 disclosure does not provide a lift-type breakdown.\", \"technology_data\": \"Closed-loop AI artificial-lift optimization; detailed model architecture and the lift-type mix are not disclosed in the Q2 update.\", \"integration_human_role\": \"The Q2 call describes autonomous 24/7 optimization. Human exception handling, override rates and control guardrails are not specified.\", \"deployment_scope\": \"Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Live deployment\", \"value\": \"1,000 wells; >2,000 additional wells prospective\", \"period\": \"August 2026\", \"attribution\": \"Devon management\", \"source_ids\": [\"SW8-001\"]}], \"economics\": \"CEO separately linked AI value to drilling/completion time and lower capital costs, and cited business optimization savings; no amount is attributed specifically to AI or the lift tool.\", \"timeline\": [{\"date\": \"2026-08-05\", \"event\": \"Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.\", \"source_ids\": [\"SW8-001\"]}], \"limitations\": [\"Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale.\"], \"assessment\": \"Latest issuer disclosure establishes deployment reach, not fleet-wide production benefit.\", \"sources\": [{\"id\": \"SW8-001\", \"title\": \"Devon Q2 2026 earnings call transcript\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"date\": \"2026-08-05\", \"type\": \"Issuer-hosted earnings call\", \"locator\": \"Printed/PDF pp. 5 and 17; CEO prepared remarks and CTO Q&A\", \"review_status\": \"Relevant passages reviewed\", \"accessed\": \"2026-10-04\"}], \"source_ids\": [\"SW8-001\"], \"field_sources\": {\"workflow\": [\"SW8-001\"], \"technology_data\": [\"SW8-001\"], \"deployment_scope\": [\"SW8-001\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"October 4 evidence review\", \"report_date\": \"2026-10-04\", \"evidence_note\": \"Operator-reported deployment; no independent validation.\"}", "Research_status": "Supplemental source added · October gap review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-006", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf"], "findings": [{"severity": "low", "field": "Follow-up review · 28 September 2026", "issue": "The issuer-hosted Q1 2026 transcript now supports the May gas-lift deployment and pilot claims. This supersedes the earlier secondary-only limitation for those claims.", "correction": "Pilot uplift is not the result for all 850+ wells. Broader rollout performance was not quantified; 1,500 wells was a target. Exclude the company-wide optimization program from AI-only benefits. Later secondary-hosted disclosures retain their source limits.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Followup_review": {"date": "2026-09-28", "decision": "Enriched existing profile", "note": "May call date and metrics checked against issuer-hosted transcript. No new initiative counted."}, "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-006", "stage": "Scaled", "stage_basis": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Resolve stale 'Reference pending review' flag: primary operator-hosted call verifies active scaled deployment. Replace 850-well Q1 scale with 1,000-well Q2 status; do not claim all are gas-lift or Delaware wells. The specific workflow title is gas-lift; Q2 calls this closed-loop AI artificial lift and later names smart gas lift.", "limits": "Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale.", "source_id": "SW8-001"}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "1,000 wells", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Operator-reported. 1,000 live wells (Aug 2026) is deployment scale; the 2–3% uplift is from the 2025 gas-lift pilot, not a measured fleet result.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "Issuer-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-007", "Company": "Devon Energy", "Segment": "E&P", "Theme": "Operational AI", "Initiative": "AI agents for drilling and completions monitoring", "Extracted_finding": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "Stage_at_disclosure": "Operating. Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "Reported_metric": "Company reported 85% of broader $1bn program achieved", "Interpretation_limit": "No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only.", "Source_IDs": "S03; R043; SW8-001", "Source_date": "2026-08-05", "Source_type": "Issuer-hosted earnings-call transcript", "Source_locator": "Q2 2026 call, printed/PDF pp. 5 and 17; CEO prepared remarks and CTO Q&A", "Source_URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Evidence_status": "Issuer-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Latest operator disclosure 5 August 2026; older disclosures retained in history.", "Context": "Devon’s drilling and completion work uses models and agents to identify unproductive time and repeat better operating practices.", "Implementation": "The Q4 2025 earnings presentation places this work inside the broader optimization program, which had achieved 85% of its $1 billion target.", "Assessment": "Faster drilling cycles can reduce required capital, but the presentation does not isolate an AI contribution. The relevant comparison would hold geology, well design and service conditions constant.", "Detail_words": 68, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Devon Energy", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-043"], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-007", "Initiative_ID": "AI-007", "Title": "D&C monitoring and AI agents", "Date": "2026-02", "Finding": "Models identify unproductive time and practices to improve drilling and completion cycles.", "Metric": "Company reported 85% of broader $1bn program achieved", "Source_IDs": "S03", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The Q4 2025 earnings presentation places this work inside the broader optimization program, which had achieved 85% of its $1 billion target.", "Context": "Devon’s drilling and completion work uses models and agents to identify unproductive time and repeat better operating practices.", "Assessment": "Faster drilling cycles can reduce required capital, but the presentation does not isolate an AI contribution. The relevant comparison would hold geology, well design and service conditions constant."}, {"ID": "REF-043", "Initiative_ID": "AI-007", "Title": "AI agents for well design and completions sequencing", "Date": "2026-05", "Finding": "At IADC Drilling Onshore, Devon's VP of D&C described ChatDVN plus the proprietary InnDrill and InnFrac data applications, which were built over about 2.5 years. Uses include well-design optimization agents, AI-driven completions sequencing, basin-wide pre-drill performance prediction, and supply-chain and contract optimization. These tools support the post-Coterra synergy plan.", "Metric": "$350M of the $1B merger synergy target comes from D&C capital optimization", "Source_IDs": "R043", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Owning the D&C data layer (InnDrill/InnFrac) is what makes design agents and pre-drill prediction possible; build the data apps before the agents.", "Partners": "ChatDVN, InnDrill, InnFrac (in-house)", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "OCT4-AI-007", "Initiative_ID": "AI-007", "Title": "AI agents for drilling and completions monitoring", "Date": "2026-08-05", "Finding": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "Metric": "Company reported 85% of broader $1bn program achieved", "Source_IDs": "SW8-001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "October 4 evidence review"}], "Latest_disclosure": "2026-08-05", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct4-ai-007", "north_america-019"], "Research_search": "{\"key\": \"devon-dc-monitoring-agents\", \"existing_ids\": [\"AI-007\"], \"company\": \"Devon Energy\", \"title\": \"AI agents for drilling and completions monitoring\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Scaling; issuer reported 1,000 autonomous wells as of Q2 2026\", \"practical_relevance\": \"This is the cohort’s most concrete AI disclosure: a named control workflow, current reported scale, and integration into core well operations. Evidence of deployment is company-reported; incremental production, costs and economics remain unisolated.\", \"business_problem\": \"Prior-register claim (not reverified): D&C monitoring models may identify unproductive time and repeat effective practices.\", \"workflow\": \"Prior-register claim (not reverified): the Q4 2025 presentation was said to describe D&C monitoring models and AI agents; the cited presentation could not be reopened.\", \"technology_data\": \"Not disclosed in verified source text.\", \"integration_human_role\": \"Not disclosed in verified source text.\", \"deployment_scope\": \"Not disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific economics verified; broader business optimization savings cannot be attributed to this workflow.\", \"timeline\": [{\"date\": \"2026-02\", \"event\": \"Existing register points to Q4 2025 earnings presentation; not reverified.\", \"source_ids\": [\"S03\"]}], \"limitations\": [\"The cited investor presentation was not readable/reopened during this pass. Treat as unverified, not disproved.\"], \"assessment\": \"The workflow would be practical if confirmed, but available evidence here is only the existing register citation.\", \"sources\": [{\"id\": \"S03\", \"title\": \"Devon Q4 2025 Earnings Presentation\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf\", \"publisher\": \"Devon Energy\", \"date\": \"2026-02\", \"type\": \"investor presentation\", \"locator\": \"Existing profile cites D&C monitoring / AI agents; exact slide not reopened\", \"review_status\": \"inaccessible\", \"supports\": [\"Existing profile says models and agents identify unproductive time and repeat better drilling/completion practices; claim is not reverified in this pass.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-023\"}], \"field_sources\": {\"business_problem\": [], \"workflow\": [], \"technology_data\": [], \"integration_human_role\": [], \"deployment_scope\": [], \"economics\": []}, \"verification\": \"unable to verify\", \"corrections_to_existing\": \"AI-007 is a distinct D&C workflow and should not be merged into lift optimization. It remains unverified pending review of Devon’s presentation.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-019\", \"source_ids\": [\"S03\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-007\"]} {\"key\": \"oct4-ai-007\", \"existing_ids\": [\"AI-007\"], \"company\": \"Devon Energy\", \"title\": \"AI agents for drilling and completions monitoring\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.\", \"practical_relevance\": \"Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.\", \"business_problem\": \"Prior-register claim (not reverified): D&C monitoring models may identify unproductive time and repeat effective practices.\", \"workflow\": \"Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.\", \"technology_data\": \"AI analysis of drilling/completion performance; engineers review alerts and take action. Detailed model implementation not disclosed.\", \"integration_human_role\": \"Teams review performance alerts and act on identified gaps. Autonomous execution is not established.\", \"deployment_scope\": \"Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Operational workflow\", \"value\": \"Live performance analysis and alerts; no AI-specific cash return established\", \"period\": \"August 2026\", \"attribution\": \"Devon management\", \"source_ids\": [\"SW8-001\"]}], \"economics\": \"No AI-specific economics verified; broader business optimization savings cannot be attributed to this workflow.\", \"timeline\": [{\"date\": \"2026-08-05\", \"event\": \"Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.\", \"source_ids\": [\"SW8-001\"]}], \"limitations\": [\"No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only.\"], \"assessment\": \"The Q2 call supports a live analytics workflow; no discrete AI-only financial outcome is isolated.\", \"sources\": [{\"id\": \"SW8-001\", \"title\": \"Devon Q2 2026 earnings call transcript\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"date\": \"2026-08-05\", \"type\": \"Issuer-hosted earnings call\", \"locator\": \"Printed/PDF pp. 5 and 17; CEO prepared remarks and CTO Q&A\", \"review_status\": \"Relevant passages reviewed\", \"accessed\": \"2026-10-04\"}], \"field_sources\": {\"workflow\": [\"SW8-001\"], \"technology_data\": [\"SW8-001\"], \"deployment_scope\": [\"SW8-001\"]}, \"verification\": \"Original source reviewed\", \"corrections_to_existing\": \"AI-007 is a distinct D&C workflow and should not be merged into lift optimization. It remains unverified pending review of Devon’s presentation.\", \"related_existing_ids\": [], \"cohort\": \"October 4 evidence review\", \"packet_id\": \"oct4-ai-007\", \"source_ids\": [\"SW8-001\"], \"report_date\": \"2026-10-04\", \"evidence_note\": \"Operator-reported deployment; no independent validation.\", \"record_ids\": [\"AI-007\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-007", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf"], "findings": [{"severity": "medium", "field": "Supplemental source review", "issue": "R043's IADC archive listing verifies only the article title/byline; the body could not be reopened, and the title concerns post-merger AI/synergies rather than the imported description of well-design/completions-sequencing agents.", "correction": "Do not use R043 as fresh verification of the named well-design workflow. Keep earlier presentation evidence as not reverified this pass rather than disproved; reconcile the R043 title/summary mismatch if retaining that citation.", "source_url": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-007", "stage": "Operating", "stage_basis": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Use later primary-source operating evidence; retain 'AI agents' only if Q4 2025 deck is reopened to verify that specific term/function. The Q2 source supports operating AI D&C monitoring/benchmarking.", "limits": "No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only.", "source_id": "SW8-001"}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Verified from Devon's issuer-hosted 5 Aug 2026 call transcript. The 85% of $1bn figure is the broader business-optimization program, not an AI-attributed result, so it is not counted as an AI outcome.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Issuer-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-008", "Company": "SLB", "Segment": "Oilfield services", "Theme": "AI products", "Initiative": "Tela agentic-AI product catalog and workflow capability", "Extracted_finding": "Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.", "Stage_at_disclosure": "Tela launched Nov 2025; current catalog advertises 33 ready workflows", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Product page is undated and mutable; accessed 2026-09-27. The current 33-workflow count may postdate original Nov 2025 launch.", "Source_IDs": "RS042; S45; S04; R070", "Source_date": "2025-11-03", "Source_type": "company product documentation", "Source_locator": "Tela feature and 33 workflow brochure sections", "Source_URL": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-11-03; see chronology for operating dates.", "Context": "Provide upstream teams with domain-aware assistance inside engineering and operations tools for technical interpretation and repetitive workflow execution.", "Implementation": "Catalog covers data discovery, reservoir/simulation queries, document ingestion for well planning, equipment recommendations, production forecasting/surveillance and other domains. SLB says some embedded agents recommend actions while more mature autonomous workflows can act within physics/domain guardrails.", "Assessment": "Analyst inference: Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.", "Detail_words": 79, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Industry software platform", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-070"], "Related_IDs": [], "Disclosures": [{"ID": "REF-070", "Initiative_ID": "AI-008", "Title": "Tela agentic AI assistant launched for upstream", "Date": "2025-11", "Finding": "SLB launched Tela, an agentic AI assistant built on the Lumi platform that combines LLMs with domain foundation models. SLB says it can interpret well logs, predict drilling issues and optimize equipment, working alongside humans or on its own. It is available across SLB's digital platforms, in the cloud or on premises.", "Metric": "Five-step agent loop: observe, plan, generate, act, learn", "Source_IDs": "R070", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The largest service company is now selling agents as a layer across its whole software estate, so independent agent vendors need either deep workflow niches or to interoperate with Tela.", "Partners": "Lumi, Delfi", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "ORIG-AI-008", "Initiative_ID": "AI-008", "Title": "Tela agentic assistant", "Date": "2026-01-23", "Finding": "Upstream assistant supports geoscience workflows and autonomous production applications.", "Metric": "More than a dozen customers engaging", "Source_IDs": "S04; S45", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Powered by Lumi, Tela combines language models with domain foundation models. Examples include well-log interpretation, drilling-issue prediction and equipment optimization. Customers can build or integrate agents.", "Context": "Tela is SLB’s upstream agentic assistant, launched in November 2025. It is intended to help technical users perform work within existing energy applications.", "Assessment": "The investor disclosure reports engagement with more than a dozen customers. Engagement is an adoption signal, but the source does not state how many are paying or running production-critical workflows."}], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-001"], "Research_search": " {\"key\": \"slb-tela-catalog\", \"existing_ids\": [\"AI-008\"], \"related_existing_ids\": [\"AI-009\"], \"company\": \"SLB\", \"title\": \"Tela agentic-AI product catalog and workflow capability\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Tela launched Nov 2025; current catalog advertises 33 ready workflows\", \"practical_relevance\": \"Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.\", \"business_problem\": \"Provide upstream teams with domain-aware assistance inside engineering and operations tools for technical interpretation and repetitive workflow execution.\", \"workflow\": \"Catalog covers data discovery, reservoir/simulation queries, document ingestion for well planning, equipment recommendations, production forecasting/surveillance and other domains. SLB says some embedded agents recommend actions while more mature autonomous workflows can act within physics/domain guardrails.\", \"technology_data\": \"Lumi data/AI platform; LLMs plus domain foundation models; deployed cloud or on-prem. Catalog currently covers in-app Tela, Canvas and Innovation Factori workflows.\", \"integration_human_role\": \"Product documentation describes human oversight and supports recommendations; autonomy is configurable/guardrailed. No claim every agent is autonomous.\", \"deployment_scope\": \"Commercial catalog offers 33 workflows; customer installation per workflow not established.\", \"partners\": [\"SLB internal product and Innovation Factori teams\"], \"disclosed_results\": [], \"economics\": \"No disclosed sales, usage, or quantified customer return for the catalog itself.\", \"timeline\": [{\"date\": \"2025-11-03\", \"event\": \"Tela launch announced\", \"source_ids\": [\"S45\"]}, {\"date\": \"2026-09-27\", \"event\": \"Current product page lists 33 deployment-ready workflows; access date, not publication date\", \"source_ids\": [\"RS042\"]}], \"limitations\": [\"Product page is undated and mutable; accessed 2026-09-27. The current 33-workflow count may postdate original Nov 2025 launch.\"], \"assessment\": \"Analyst inference: Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.\", \"sources\": [{\"id\": \"RS042\", \"title\": \"Tela agentic-AI assistant catalog\", \"url\": \"https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela\", \"publisher\": \"SLB\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company product documentation\", \"locator\": \"Tela feature and 33 workflow brochure sections\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB lists 33 deployment-ready workflows across subsurface, drilling, production, facilities and sustainability.\", \"Tela uses LLMs and domain foundation models on Lumi and is offered within SLB applications/cloud or on premises.\", \"The catalog is product capability and availability; individual workflows are not all evidence of customer deployment.\"], \"accessed_date\": \"2026-09-27\", \"original_id\": \"slb_tela\"}, {\"id\": \"S45\", \"title\": \"SLB Unveils Groundbreaking New Agentic AI Technology for the Energy Industry\", \"url\": \"https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai\", \"publisher\": \"SLB\", \"date\": \"2025-11-03\", \"type\": \"company announcement\", \"locator\": \"Tela architecture and deployment availability\", \"review_status\": \"full text reviewed\", \"supports\": [\"Tela announced as agentic AI assistant built on Lumi and using LLMs and domain foundation models.\", \"SLB described a five-step loop and says assistants can interpret logs, predict drilling issues and optimize equipment; autonomous actions depend on workflow and are not customer deployment evidence.\"], \"original_id\": \"slb_tela_launch\"}], \"field_sources\": {\"business_problem\": [\"RS042\", \"S45\"], \"workflow\": [\"RS042\", \"S45\"], \"technology_data\": [\"RS042\", \"S45\"], \"integration_human_role\": [\"RS042\", \"S45\"], \"deployment_scope\": [\"RS042\", \"S45\"], \"economics\": [\"RS042\", \"S45\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-001\", \"source_ids\": [\"RS042\", \"S45\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-008\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [{"claim": "33 deployment-ready Tela workflows", "reason": "Current SLB product page confirms 33 workflows; this is a catalog/capability count, not completed customer deployments. Product page is undated and mutable.", "source_url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela"}], "Audit_review": {"id": "AI-008", "company": "SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-009", "Company": "SLB", "Segment": "Oilfield services", "Theme": "AI products", "Initiative": "Lumi data and AI platform", "Extracted_finding": "Platform adoption connects geoscience, drilling and production workflows.", "Stage_at_disclosure": "Commercial adoption", "Reported_metric": "More than 50 customers", "Interpretation_limit": "Digital ARR is broader than AI revenue. 24 October thesis limit: SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Source_IDs": "S04; R069; SLBTH-Q3; SLBTH-FY25; SLBTH-MENON; SLBTH-BIGUET; SLBTH-TERMS; SLBTH-GENAI; SLBTH-MULLIN; SLBTH-RANNEN", "Source_date": "2026-01-23", "Source_type": "Earnings call transcript", "Source_locator": "PDF pp. 3, 10; Olivier Le Peuch", "Source_URL": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Lumi provides a common data and AI platform across geoscience, drilling and production applications. Its role is to connect data with the tools that technical teams already use.", "Implementation": "SLB reported more than 50 customers. Lumi also underpins Tela and the announced Shell collaboration, so those records represent related layers of the product architecture.\n\n24 October data and AI review: Menon describes a seismic model trained on public and SLB multiclient datasets, with customers able to fine-tune for their own assets. Lumi connects governed data; domain models handle technical formats; Tela provides interaction and workflow orchestration. SLB describes subscription upgrades, metered usage, tailored workflows and marketplace revenue sharing. These are commercial mechanisms, not disclosed AI-only earnings.", "Assessment": "Customer count supports commercial adoption. It gives no common denominator for seats, data volume or recurring AI spend. SLB’s digital recurring revenue includes products beyond this platform.", "Detail_words": 80, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Exploration & geoscience", "Reservoir engineering & field development", "Drilling & completions", "Production & field operations"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Industry software platform", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-069"], "Related_IDs": ["AI-008", "AI-047", "AI-105", "AI-301"], "Disclosures": [{"ID": "REF-069", "Initiative_ID": "AI-009", "Title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "Date": "2025-06", "Finding": "SLB said its Lumi data and AI platform would be available on Mistral Compute, Mistral AI's new European AI infrastructure. Mistral's LLMs have been the primary generative models inside Lumi since the partnership began in early 2024, and this deal lets the platform be hosted entirely within European borders for data residency.", "Metric": "Lumi launched Sept 2024 (SLB Digital Forum, Monaco)", "Source_IDs": "R069", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Where a buyer needs data sovereignty, offering models and hosting that stay in-region wins deals, so a sovereign or on-prem deployment option should be designed in from the start.", "Partners": "Mistral AI, Lumi, Delfi", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "ORIG-AI-009", "Initiative_ID": "AI-009", "Title": "Lumi data and AI platform", "Date": "2026-01-23", "Finding": "Platform adoption connects geoscience, drilling and production workflows.", "Metric": "More than 50 customers", "Source_IDs": "S04", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB reported more than 50 customers. Lumi also underpins Tela and the announced Shell collaboration, so those records represent related layers of the product architecture.", "Context": "Lumi provides a common data and AI platform across geoscience, drilling and production applications. Its role is to connect data with the tools that technical teams already use.", "Assessment": "Customer count supports commercial adoption. It gives no common denominator for seats, data volume or recurring AI spend. SLB’s digital recurring revenue includes products beyond this platform."}, {"ID": "SLBTH-AI-009", "Initiative_ID": "AI-009", "Title": "What holds up in SLB’s data and AI investment thesis?", "Date": "2026-10-23", "Finding": "SLB has a credible route to earning more from AI: it can put specialized models into software and operating workflows that customers already use. The public evidence supports a platform, engineering and governed-access advantage. It does not support treating every customer’s downhole data as an unrestricted SLB-owned training pool. Menon describes a seismic model trained on public and SLB multiclient datasets, with customers able to fine-tune for their own assets. Lumi connects governed data; domain models handle technical formats; Tela provides interaction and workflow orchestration. SLB describes subscription upgrades, metered usage, tailored workflows and marketplace revenue sharing. These are commercial mechanisms, not disclosed AI-only earnings.", "Metric": "FY2025 Digital revenue was $2.660bn and adjusted EBITDA $928m (34.9% margin). Q3 2026 Digital revenue was $693m, adjusted EBITDA $228m and ARR $1.04bn. These are division-wide figures, not AI revenue or customer ROI. The CFO’s 2030 $1.8–2.0bn adjusted EBITDA and approximately $2bn ARR are ambitions, not results.", "Source_IDs": "SLBTH-Q3; SLBTH-FY25; SLBTH-MENON; SLBTH-BIGUET; SLBTH-TERMS; SLBTH-GENAI; SLBTH-MULLIN; SLBTH-RANNEN", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "Evidence": "Original investor scripts, financial tables and public terms read", "Origin": "24 October editor SLB thesis review; no new initiative", "Assessment": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures."}], "Latest_disclosure": "2026-10-23", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_status": "Existing Lumi dossier expanded 24 October; company and investor claims kept separate", "Research_packets": ["slb-data-ai-thesis"], "Research_category": "Commercial AI product", "Audit_flags": [], "Audit_review": {"id": "AI-009", "company": "SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f"], "findings": [{"severity": "medium", "field": "economics", "issue": "SLB Digital ARR is a broad digital business metric and does not isolate Lumi revenue or adoption.", "correction": "Keep platform positioning; do not assign corporate digital ARR to Lumi.", "source_url": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Digital ARR is broader than AI revenue. 24 October thesis limit: SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-24", "Research_verification": "verified primary"}, {"ID": "AI-010", "Company": "Shell / SLB", "Segment": "Integrated / services", "Theme": "Operational AI", "Initiative": "Upstream data/AI infrastructure and agentic-solutions collaboration", "Extracted_finding": "Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.", "Stage_at_disclosure": "Agreement announced 2025-12-11; Q1 2026 SLB reporting repeats plans", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "2026-01 SLB earnings disclosure is a later update/repetition; agreement launch was 2025-12-11. Development plan only; no deployed agents or outcomes substantiated.", "Source_IDs": "S46; S05; R017", "Source_date": "2026-01-23", "Source_type": "company announcement", "Source_locator": "Shell collaboration agreement", "Source_URL": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-01-23; see chronology for operating dates.", "Context": "Upstream teams often work across distinct data and workflows from subsurface to well construction and production.", "Implementation": "Shell and SLB agreed to develop agentic AI solutions for upstream experts and an open infrastructure to unify data and workflows across the three domains using SLB Lumi.", "Assessment": "Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.", "Detail_words": 71, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Upstream workflow integration", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Shell", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-017"], "Related_IDs": [], "Disclosures": [{"ID": "REF-017", "Initiative_ID": "AI-010", "Title": "SLB and Shell co-develop agentic AI for upstream", "Date": "2025-12", "Finding": "SLB and Shell signed a strategic collaboration to develop agentic AI solutions for technical experts across subsurface, well construction and production. The work builds on SLB's Lumi data and AI platform and Shell's earlier deployment of Petrel across its global assets.", "Metric": "No quantified metric supplied", "Source_IDs": "R017", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Supermajors are co-developing agents with OFS vendors on shared data platforms; independent builders should interoperate with Lumi/OSDU instead of competing with them.", "Partners": "SLB (Lumi, Petrel)", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "ORIG-AI-010", "Initiative_ID": "AI-010", "Title": "Upstream agentic collaboration", "Date": "2026-01-23", "Finding": "Plans an open data and AI environment across subsurface, well construction and production.", "Metric": "Not disclosed", "Source_IDs": "S05; S46", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The planned open environment uses Lumi to connect subsurface, well construction and production data. Agentic applications would support technical experts and decision makers across these workflows.", "Context": "Shell and SLB signed a December 2025 agreement to develop digital and AI solutions for upstream operations and potentially the wider industry.", "Assessment": "The agreement defines a development direction. It does not establish a deployed Shell-wide system, a production schedule, contract economics or measured efficiency gains."}], "Latest_disclosure": "2026-01-23", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-003"], "Research_search": " {\"key\": \"shell-slb-collab\", \"existing_ids\": [\"AI-010\"], \"related_existing_ids\": [], \"company\": \"Shell / SLB\", \"title\": \"Upstream data/AI infrastructure and agentic-solutions collaboration\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Agreement announced 2025-12-11; Q1 2026 SLB reporting repeats plans\", \"practical_relevance\": \"Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.\", \"business_problem\": \"Upstream teams often work across distinct data and workflows from subsurface to well construction and production.\", \"workflow\": \"Shell and SLB agreed to develop agentic AI solutions for upstream experts and an open infrastructure to unify data and workflows across the three domains using SLB Lumi.\", \"technology_data\": \"SLB Lumi data/AI platform; intended open and secure data environment. Specific model, data schemas and production systems not detailed.\", \"integration_human_role\": \"Framed to augment technical experts and decision-makers; no human approval gates described.\", \"deployment_scope\": \"Joint development ambition for Shell and wider industry; no named deployed workflow or user/asset scale disclosed.\", \"partners\": [\"SLB\"], \"disclosed_results\": [], \"economics\": \"No contract value, realized savings or results reported.\", \"timeline\": [{\"date\": \"2025-12-11\", \"event\": \"Upstream data/AI infrastructure and agentic-solutions collaboration documented or disclosed.\", \"source_ids\": [\"S46\", \"S05\"]}], \"limitations\": [\"2026-01 SLB earnings disclosure is a later update/repetition; agreement launch was 2025-12-11.\", \"Development plan only; no deployed agents or outcomes substantiated.\"], \"assessment\": \"Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.\", \"sources\": [{\"id\": \"S46\", \"title\": \"SLB Enters Collaboration Agreement to Accelerate New Digital Solutions\", \"url\": \"https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab\", \"publisher\": \"SLB\", \"date\": \"2025-12-11\", \"type\": \"company announcement\", \"locator\": \"Shell collaboration agreement\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB and Shell agreed to develop upstream digital and AI solutions; roadmap aims for open data/AI infrastructure across subsurface, well construction and production on SLB Lumi.\", \"Announcement states ambition/aim, not a deployed customer workflow or realized results.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-003\"}, {\"id\": \"S05\", \"title\": \"SLB Q4 and full-year 2025 results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results\", \"publisher\": \"SLB\", \"date\": \"2026-01-23\", \"type\": \"earnings release\", \"locator\": \"Digital segment customer collaborations section\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q4 report repeats Shell collaboration aims and scope; this is an in-window disclosure/update, not evidence the collaboration was launched in January or deployed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-004\"}], \"field_sources\": {\"business_problem\": [\"S46\"], \"workflow\": [\"S46\", \"S05\"], \"technology_data\": [\"S46\"], \"integration_human_role\": [\"S46\"], \"deployment_scope\": [\"S46\", \"S05\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-003\", \"source_ids\": [\"S46\", \"S05\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-010\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-010", "company": "Shell / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The JPT/SPE report confirms a signed strategic collaboration aimed at developing agentic AI and unified upstream workflows; it does not establish an operating deployment.", "correction": "Keep the profile at collaboration/development stage and attribute it to secondary editorial corroboration; avoid wording that suggests Shell assets already use the joint agentic system.", "source_url": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-011", "Company": "ADNOC / SLB", "Segment": "NOC / services", "Theme": "Operational AI", "Initiative": "AiPSO production system optimization", "Extracted_finding": "A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.", "Stage_at_disclosure": "Scaled. ADNOC’s 2025-11-03 announcement says AiPSO launched with an initial deployment across eight fields and plans to roll out further through 2027; the existing multi-field footprint is deployed while additional expansion remains planned.", "Reported_metric": "complex tasks speed-up: up to 10x (Company-reported)", "Interpretation_limit": "Company-reported disclosure; no independent validation identified in reviewed sources.", "Source_IDs": "S06; S05", "Source_date": "2025-11-03", "Source_type": "company announcement", "Source_locator": "AiPSO deployment paragraphs", "Source_URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-11-03; see chronology for operating dates.", "Context": "Speed diagnosis of well-performance issues and production-system optimization that otherwise takes engineers days.", "Implementation": "Connects office and field work so engineers can identify issues and optimize wells in minutes; platform is intended to coordinate work across assets.", "Assessment": "Analyst inference: A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.", "Detail_words": 69, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ADNOC", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-011", "Initiative_ID": "AI-011", "Title": "AiPSO production optimization", "Date": "2026-01-23", "Finding": "Combines live data, proprietary ML and Cognite Data Fusion to optimize wells and facilities.", "Metric": "Eight fields", "Source_IDs": "S05; S06", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The solution combines Lumi, Cognite Data Fusion and ADNOC machine-learning models. It processes live operating information to help engineers diagnose and optimize production.", "Context": "AiPSO addresses fragmented production information across wells and facilities. ADNOC and SLB initially deployed it across eight fields.", "Assessment": "The company describes analysis moving from days to minutes. That is a workflow claim, with no disclosed field-level production uplift or independently verified cash return in these sources."}], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-003"], "Research_search": " {\"key\": \"adnoc-aipso\", \"existing_ids\": [\"AI-011\"], \"company\": \"ADNOC\", \"title\": \"AiPSO production system optimization\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Launched Nov 2025; initial deployment across eight fields; rollout plan through 2027\", \"practical_relevance\": \"A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.\", \"business_problem\": \"Speed diagnosis of well-performance issues and production-system optimization that otherwise takes engineers days.\", \"workflow\": \"Connects office and field work so engineers can identify issues and optimize wells in minutes; platform is intended to coordinate work across assets.\", \"technology_data\": \"AI-powered production system optimization platform; source does not disclose model architecture, sensors, workflow integrations or specific recommendations.\", \"integration_human_role\": \"Engineers use platform outputs for operational decisions; autonomous actuation is not claimed.\", \"deployment_scope\": \"Initial eight fields with planned onshore/offshore deployment across ADNOC by 2027.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"complex tasks speed-up\", \"value\": \"up to 10x\", \"period\": \"Nov 2025 launch disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S06\"]}], \"economics\": \"No monetary return disclosed; ADNOC claims complex work up to 10x faster.\", \"timeline\": [{\"date\": \"2025-11-03\", \"event\": \"Platform launched; initial deployment across eight fields\", \"source_ids\": [\"S06\"]}, {\"date\": \"2027\", \"event\": \"Target for all onshore and offshore fields\", \"source_ids\": [\"S06\"]}], \"limitations\": [\"Company-reported disclosure; no independent validation identified in reviewed sources.\"], \"assessment\": \"Analyst inference: A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.\", \"sources\": [{\"id\": \"S06\", \"title\": \"ADNOC and SLB Launch Transformative AI-Powered Solution to Boost Upstream Productivity\", \"url\": \"https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity\", \"publisher\": \"ADNOC\", \"date\": \"2025-11-03\", \"type\": \"company announcement\", \"locator\": \"AiPSO deployment paragraphs\", \"review_status\": \"full text reviewed\", \"supports\": [\"Production System Optimization platform launched with initial deployment across eight fields.\", \"Connects office and field workflows so engineers diagnose issues and optimize wells in minutes rather than days; planned deployment across all onshore/offshore fields by 2027.\", \"ADNOC says complex tasks can be completed up to ten times faster; this is a company claim, not independently validated.\"], \"original_id\": \"adnoc_aipso\"}], \"field_sources\": {\"business_problem\": [\"S06\"], \"workflow\": [\"S06\"], \"technology_data\": [\"S06\"], \"integration_human_role\": [\"S06\"], \"deployment_scope\": [\"S06\"], \"economics\": [\"S06\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-003\", \"source_ids\": [\"S06\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-011\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-011", "company": "ADNOC / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-011", "stage": "Scaled", "stage_basis": "ADNOC’s 2025-11-03 announcement says AiPSO launched with an initial deployment across eight fields and plans to roll out further through 2027; the existing multi-field footprint is deployed while additional expansion remains planned.", "source_url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "source_date": "2025-11-03", "evidence_origin": "Operator", "limits": "No field-level adoption, achieved production metric, or independent validation; planned 2027 reach is not achieved scale.", "source_id": "S06"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "scaled", "Scale_note": "8 fields", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "'Up to 10x' is a launch capability claim, not a measured realized result.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-012", "Company": "bp / SLB", "Segment": "Integrated / services", "Theme": "Operational AI", "Initiative": "Remote offshore drilling operations for bp Azerbaijan rigs", "Extracted_finding": "Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.", "Stage_at_disclosure": "Scaled. SLB’s 2026-01-23 earnings disclosure reports its remote offshore drilling operating arrangement deployed on multiple bp Azerbaijan rigs, with operational result reported in the same earnings update.", "Reported_metric": "offshore well-placement crew size: reduced 66% (Vendor-reported); personnel-on-board days saved: >1,500 (Vendor-reported)", "Interpretation_limit": "Vendor earnings disclosure; no operator corroboration reviewed.", "Source_IDs": "RS046; RS047; S05", "Source_date": "2026-01-23", "Source_type": "earnings release", "Source_locator": "Technology highlights, pp. 18-19", "Source_URL": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-01-23; see chronology for operating dates.", "Context": "Reduce offshore staffing and oversight burden.", "Implementation": "Performance Live remotely supports offshore platforms with DrillOps/Neuro; Neuro automates trajectory delivery and downlinking.", "Assessment": "Analyst inference: Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.", "Detail_words": 73, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "bp", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-012", "Initiative_ID": "AI-012", "Title": "Azerbaijan remote drilling", "Date": "2026-01-23", "Finding": "AI automation and remote operations support offshore drilling decisions.", "Metric": "66% smaller placement crew; >1,500 onboard days saved", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB names Performance Live, DrillOps advisory and Neuro autonomous directional drilling. It reports a 66% smaller placement crew and more than 1,500 personnel-on-board days avoided.", "Context": "bp’s Azerbaijan operations use a combination of remote operations and drilling automation to reduce the need for well-placement staff offshore.", "Assessment": "These results belong to an integrated digital operating model. They cannot be assigned solely to an AI algorithm. Personnel-on-board days measure offshore exposure, rather than direct cash savings."}], "Latest_disclosure": "2026-01-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-004"], "Research_search": " {\"key\": \"slb-bp-azerbaijan\", \"existing_ids\": [\"AI-012\"], \"related_existing_ids\": [], \"company\": \"bp\", \"title\": \"Remote offshore drilling operations for bp Azerbaijan rigs\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Deployed on multiple rigs; result reported in Jan 2026 earnings\", \"practical_relevance\": \"Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.\", \"business_problem\": \"Reduce offshore staffing and oversight burden.\", \"workflow\": \"Performance Live remotely supports offshore platforms with DrillOps/Neuro; Neuro automates trajectory delivery and downlinking.\", \"technology_data\": \"DrillOps models formation/drilling parameters; Neuro controls trajectory. Detailed model architecture undisclosed.\", \"integration_human_role\": \"Remote center specialists and cross-trained offshore drillers work together; override details unspecified.\", \"deployment_scope\": \"Several Western Asia offshore platforms; operator not named in SLB case study.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"offshore well-placement crew size\", \"value\": \"reduced 66%\", \"period\": \"Q4 2025 results\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS047\"]}, {\"metric\": \"personnel-on-board days saved\", \"value\": \">1,500\", \"period\": \"Q4 2025 results\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS047\"]}], \"economics\": \"SLB reports crew −66% and >1,500 offshore personnel-days saved; financial savings not stated.\", \"timeline\": [{\"date\": \"2026-01-16\", \"event\": \"SLB disclosed Azerbaijan remote drilling results in Q4 earnings release\", \"source_ids\": [\"RS046\"]}], \"limitations\": [\"Vendor earnings disclosure; no operator corroboration reviewed.\"], \"assessment\": \"Analyst inference: Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.\", \"sources\": [{\"id\": \"RS046\", \"title\": \"SLB Announces Fourth-Quarter and Full-Year 2025 Results\", \"url\": \"https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae\", \"publisher\": \"SLB\", \"date\": \"2026-01-23\", \"type\": \"earnings release\", \"locator\": \"Technology highlights, pp. 18-19\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q4 release also records the bp partnership update.\"], \"original_id\": \"slb_q4\"}, {\"id\": \"RS047\", \"title\": \"SLB digital solutions save 1,500 personnel-on-board days and boost efficiency\", \"url\": \"https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days\", \"publisher\": \"SLB\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company case study\", \"locator\": \"Summary, objective and results; excerpt reviewed\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Performance Live with DrillOps and Neuro reduced offshore well-placement crew size 66% and saved >1,500 POB days across several Western Asia platforms.\", \"Case attributes benefits to remote operations and autonomous directional drilling; exact operator not identified in the case page.\"], \"accessed_date\": \"2026-09-27\", \"original_id\": \"bp_azerbaijan_case\"}], \"field_sources\": {\"business_problem\": [\"RS047\"], \"workflow\": [\"RS047\"], \"technology_data\": [\"RS047\"], \"integration_human_role\": [\"RS047\"], \"deployment_scope\": [\"RS047\"], \"economics\": [\"RS047\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-004\", \"source_ids\": [\"RS046\", \"RS047\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-012\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-012", "company": "bp / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-012", "stage": "Scaled", "stage_basis": "SLB’s 2026-01-23 earnings disclosure reports its remote offshore drilling operating arrangement deployed on multiple bp Azerbaijan rigs, with operational result reported in the same earnings update.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration reviewed here, and 'multiple rigs' is not a count of wells or offshore fields.", "source_id": "RS046"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release; no bp-side corroboration.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-013", "Company": "OQ Exploration & Production / SLB", "Segment": "E&P / services", "Theme": "Operational AI", "Initiative": "Neuro autonomous directional drilling on OQ wells", "Extracted_finding": "This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.", "Stage_at_disclosure": "Operating. SLB’s 2026-01-23 earnings update reports Neuro autonomous directional drilling applied on four OQ wells. That is field use, but four wells alone do not establish broad scaling.", "Reported_metric": "ROP increase: up to 50% on three wells (Vendor-reported); ROP increase: 40% on one well (Vendor-reported)", "Interpretation_limit": "Primary vendor earnings evidence; operator-side corroboration and well-by-well denominators are not disclosed.", "Source_IDs": "RS046; S05", "Source_date": "2026-01-23", "Source_type": "earnings release", "Source_locator": "Technology highlights, pp. 18-19", "Source_URL": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-01-23; see chronology for operating dates.", "Context": "Improve drilling speed in Oman.", "Implementation": "Neuro/PowerDrive results on three wells; PowerDrive One on another.", "Assessment": "Analyst inference: This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.", "Detail_words": 69, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "OQ Exploration & Production", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-013", "Initiative_ID": "AI-013", "Title": "Oman autonomous drilling", "Date": "2026-01-23", "Finding": "AI optimization improves rotary steerable drilling.", "Metric": "Up to 50% faster penetration on three wells", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB reports up to 50% faster rate of penetration across three wells using Neuro and PowerDrive. A separate PowerDrive One application reported a 40% improvement.", "Context": "OQ Exploration & Production applied autonomous directional drilling in Oman to improve drilling performance.", "Assessment": "The three-well result is a limited field comparison. Rate of penetration measures drilling speed, and should not be treated as a same-sized reduction in total well cost or full-cycle duration."}], "Latest_disclosure": "2026-01-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-005"], "Research_search": " {\"key\": \"slb-oq-neuro\", \"existing_ids\": [\"AI-013\"], \"related_existing_ids\": [], \"company\": \"OQ Exploration & Production\", \"title\": \"Neuro autonomous directional drilling on OQ wells\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Drilled four reported wells; results published Jan 2026\", \"practical_relevance\": \"This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.\", \"business_problem\": \"Improve drilling speed in Oman.\", \"workflow\": \"Neuro/PowerDrive results on three wells; PowerDrive One on another.\", \"technology_data\": \"Neuro, RSS; model details absent.\", \"integration_human_role\": \"Supervision details absent.\", \"deployment_scope\": \"Four wells.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"ROP increase\", \"value\": \"up to 50% on three wells\", \"period\": \"reported Jan 2026\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS046\"]}, {\"metric\": \"ROP increase\", \"value\": \"40% on one well\", \"period\": \"reported Jan 2026\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS046\"]}], \"economics\": \"SLB: ROP up to +50% on three, +40% on one; less tripping/casing, no cost.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"OQ Neuro and PowerDrive well campaigns conducted; exact well dates not disclosed\", \"source_ids\": [\"RS046\"]}, {\"date\": \"2026-01-16\", \"event\": \"Results disclosed by SLB\", \"source_ids\": [\"RS046\"]}], \"limitations\": [\"Primary vendor earnings evidence; operator-side corroboration and well-by-well denominators are not disclosed.\"], \"assessment\": \"Analyst inference: This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.\", \"sources\": [{\"id\": \"RS046\", \"title\": \"SLB Announces Fourth-Quarter and Full-Year 2025 Results\", \"url\": \"https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae\", \"publisher\": \"SLB\", \"date\": \"2026-01-23\", \"type\": \"earnings release\", \"locator\": \"Technology highlights, pp. 18-19\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile.\"], \"original_id\": \"slb_q4\"}], \"field_sources\": {\"business_problem\": [\"RS046\"], \"workflow\": [\"RS046\"], \"technology_data\": [\"RS046\"], \"integration_human_role\": [\"RS046\"], \"deployment_scope\": [\"RS046\"], \"economics\": [\"RS046\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-005\", \"source_ids\": [\"RS046\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-013\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [{"claim": "Autonomous drilling project in Oman", "reason": "Existing cited source is SLB earnings material; no OQ primary-source verification completed. Vendor-reported partnership, not disproved.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"}, {"claim": "Seven AI use cases executed and two agentic AI cases developed; >$10m return from digital portfolio.", "reason": "Search excerpt of OQ 2025 report exposes aggregate figures but not named cases or AI-attributable portion of overall digital economics; existing Neuro Oman is separately audited.", "source_url": "https://assets.oq.com/-/media/oq/files/sustainability/sustainability-at-oq/oq-sustainability-report--2025--english.pdf?rev=-1"}], "Audit_review": {"id": "AI-013", "company": "OQ Exploration & Production / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-013", "stage": "Operating", "stage_basis": "SLB’s 2026-01-23 earnings update reports Neuro autonomous directional drilling applied on four OQ wells. That is field use, but four wells alone do not establish broad scaling.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration or well-by-well outcome denominator reviewed.", "source_id": "RS046"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release; results cover three wells plus one well.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-014", "Company": "Kuwait Oil Company / SLB", "Segment": "NOC / services", "Theme": "Operational AI", "Initiative": "AI-assisted seismic processing contract for Kuwait Oil Company", "Extracted_finding": "The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.", "Stage_at_disclosure": "Five-year award reported Jan 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Vendor-reported contract; no KOC source or specific AI performance metric reviewed.", "Source_IDs": "RS046; S05", "Source_date": "2026-01-23", "Source_type": "earnings release", "Source_locator": "Technology highlights, pp. 18-19", "Source_URL": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-01-23; see chronology for operating dates.", "Context": "Process KOC seismic data.", "Implementation": "Five-year contract includes AI reprocessing.", "Assessment": "Analyst inference: The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.", "Detail_words": 64, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Kuwait Oil Company", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-014", "Initiative_ID": "AI-014", "Title": "AI seismic processing", "Date": "2026-01-23", "Finding": "Five-year contract incorporates AI workflows into seismic interpretation.", "Metric": "Five-year term", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The scope covers multiple seismic types and onshore, offshore and transition environments. AI supports interpretation within the larger processing program.", "Context": "Kuwait Oil Company awarded a five-year seismic processing and reprocessing contract that incorporates AI workflows.", "Assessment": "A contract award establishes commercial scope. The source does not give a separated AI fee, realized interpretation gain or improved discovery outcome. See AI-066 for a distinct KOC program."}], "Latest_disclosure": "2026-01-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-006"], "Research_search": " {\"key\": \"slb-koc-seismic\", \"existing_ids\": [\"AI-014\"], \"related_existing_ids\": [], \"company\": \"Kuwait Oil Company\", \"title\": \"AI-assisted seismic processing contract for Kuwait Oil Company\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Five-year award reported Jan 2026\", \"practical_relevance\": \"The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.\", \"business_problem\": \"Process KOC seismic data.\", \"workflow\": \"Five-year contract includes AI reprocessing.\", \"technology_data\": \"AI workflow; methods absent.\", \"integration_human_role\": \"Not described.\", \"deployment_scope\": \"KOC; coverage absent.\", \"partners\": [\"SLB\"], \"disclosed_results\": [], \"economics\": \"Value/outcomes undisclosed.\", \"timeline\": [{\"date\": \"2025-12-31\", \"event\": \"Five-year seismic AI processing/reprocessing contract awarded (reported in Q4 2025 results)\", \"source_ids\": [\"RS046\"]}], \"limitations\": [\"Vendor-reported contract; no KOC source or specific AI performance metric reviewed.\"], \"assessment\": \"Analyst inference: The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.\", \"sources\": [{\"id\": \"RS046\", \"title\": \"SLB Announces Fourth-Quarter and Full-Year 2025 Results\", \"url\": \"https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae\", \"publisher\": \"SLB\", \"date\": \"2026-01-23\", \"type\": \"earnings release\", \"locator\": \"Technology highlights, pp. 18-19\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile.\"], \"original_id\": \"slb_q4\"}], \"field_sources\": {\"business_problem\": [\"RS046\"], \"workflow\": [\"RS046\"], \"technology_data\": [\"RS046\"], \"integration_human_role\": [\"RS046\"], \"deployment_scope\": [\"RS046\"], \"economics\": [\"RS046\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-006\", \"source_ids\": [\"RS046\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-014\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [{"claim": "AI seismic processing, Jurassic gas AI and G Agent drilling outcomes", "reason": "No official KOC source review completed in this pass to substantiate these entries; the SLB earnings source is vendor disclosure and Kuwait Times is secondary. Retain as reported/unverified pending primary KOC technical or procurement evidence.", "source_url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/"}], "Audit_review": {"id": "AI-014", "company": "Kuwait Oil Company / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release; no KOC-side corroboration.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-015", "Company": "SLB", "Segment": "Oilfield services", "Theme": "AI infrastructure", "Initiative": "Data-center equipment capacity", "Extracted_finding": "Expanded Shreveport manufacturing for AI-related infrastructure demand.", "Stage_at_disclosure": "Expansion", "Reported_metric": "Facility footprint nearly doubled", "Interpretation_limit": "Infrastructure sales, not internal AI use.", "Source_IDs": "S05", "Source_date": "2026-01-23", "Source_type": "Earnings release", "Source_locator": "Digital and New Horizons of Growth sections", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "SLB expanded manufacturing in Shreveport, Louisiana, to meet data-center infrastructure demand.", "Implementation": "The facility footprint nearly doubled. The source describes equipment capacity expansion but does not disclose a plant-specific investment amount or output capacity.", "Assessment": "This is a manufacturing exposure to data-center growth. It should be assessed through demand, utilization and margins, separately from the operating benefits of AI inside oil and gas.", "Detail_words": 61, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-015", "Initiative_ID": "AI-015", "Title": "Data-center equipment capacity", "Date": "2026-01-23", "Finding": "Expanded Shreveport manufacturing for AI-related infrastructure demand.", "Metric": "Facility footprint nearly doubled", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The facility footprint nearly doubled. The source describes equipment capacity expansion but does not disclose a plant-specific investment amount or output capacity.", "Context": "SLB expanded manufacturing in Shreveport, Louisiana, to meet data-center infrastructure demand.", "Assessment": "This is a manufacturing exposure to data-center growth. It should be assessed through demand, utilization and margins, separately from the operating benefits of AI inside oil and gas."}], "Latest_disclosure": "2026-01-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [], "Audit_review": {"id": "AI-015", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [{"severity": "medium", "field": "classification", "issue": "Data-center equipment supply is an AI-demand infrastructure business, not SLB internal AI adoption.", "correction": "Classify as AI-demand infrastructure; results reflect capacity/portfolio rather than operational AI use.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-016", "Company": "Halliburton", "Segment": "Oilfield services", "Theme": "AI products", "Initiative": "LOGIX autonomous drilling", "Extracted_finding": "Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.", "Stage_at_disclosure": "Commercial platform; 2025 case study documents a European operator pilot", "Reported_metric": "Average ROP increase across four sections: 43.6% (Vendor-reported); Well delivery time improvement: >30% (Vendor-reported)", "Interpretation_limit": "Case study operator and field not named. Vendor-reported results; no independent corroboration. The existing 2025 investor presentation is an excerpted product description; separate case study gives additional, vendor-reported workflow details.", "Source_IDs": "S07; RS062; WATCH20UP-BP", "Source_date": "Undated source; see chronology", "Source_type": "investor presentation", "Source_locator": "Slides 11 and 19", "Source_URL": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Improve well placement consistency, penetration rate and well delivery in complex directional drilling.", "Implementation": "Combines automated steering and drilling-parameter recommendations; commands may be sent to the bottom-hole assembly or shown as advice for directional drillers. Remote center and rig crews view drilling data and advisory outputs.\n\nBumerangue: Halliburton’s August contract for bp’s Bumerangue appraisal campaign includes LOGIX automation and remote operations offshore Brazil. The integrated drilling and evaluation scope connects planning, execution and evaluation. The release says LOGIX will be deployed; it does not identify an LLM, individual control permissions or human override arrangements. A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform.", "Assessment": "Analyst inference: Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.", "Detail_words": 61, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Halliburton", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-016", "Initiative_ID": "AI-016", "Title": "LOGIX autonomous drilling", "Date": "2025-Q1", "Finding": "Combines physics models and machine learning in autonomous drilling.", "Metric": "No AI-only financial metric", "Source_IDs": "S07", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Halliburton’s Q1 2025 investor material positions autonomous drilling inside a wider digital well-construction system.", "Context": "LOGIX combines physics-based models and machine learning in a drilling automation offering. Its purpose is consistent execution of the well plan.", "Assessment": "The source establishes a commercial capability, without a named trial in this extract. Well conditions, operator procedures and the automation boundary would matter when assessing performance."}, {"ID": "WATCH20UP-AI-016-BP", "Initiative_ID": "AI-016", "Title": "Bumerangue’s appraisal contract includes drilling automation", "Date": "2026-08-25", "Finding": "Halliburton’s August contract for bp’s Bumerangue appraisal campaign includes LOGIX automation and remote operations offshore Brazil. The integrated drilling and evaluation scope connects planning, execution and evaluation. The release says LOGIX will be deployed; it does not identify an LLM, individual control permissions or human override arrangements.", "Metric": "No number of wells, completed automated wells, operating comparator, quantified benefit, contract value or net ROI disclosed.", "Source_IDs": "WATCH20UP-BP", "Source_URL": "https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil", "Evidence": "Original supplier release read; planned deployment", "Origin": "20 October user-requested source review", "Assessment": "A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform."}], "Latest_disclosure": "2026-08-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-025"], "Research_search": " {\"key\": \"hall-logix\", \"existing_ids\": [\"AI-016\"], \"related_existing_ids\": [], \"company\": \"Halliburton\", \"title\": \"LOGIX autonomous drilling\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Commercial platform; 2025 case study documents a European operator pilot\", \"practical_relevance\": \"Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.\", \"business_problem\": \"Improve well placement consistency, penetration rate and well delivery in complex directional drilling.\", \"workflow\": \"Combines automated steering and drilling-parameter recommendations; commands may be sent to the bottom-hole assembly or shown as advice for directional drillers. Remote center and rig crews view drilling data and advisory outputs.\", \"technology_data\": \"Physics-based models and machine learning on drilling data; LOGIX, iCruise RSS and LWD sensors.\", \"integration_human_role\": \"The case study describes full closed-loop automation, with commands either automated or advisory; engineering/rig personnel retain operational authority as configured.\", \"deployment_scope\": \"Commercial product with a case study of three automated wells over two years for an unnamed European operator.\", \"partners\": [\"Halliburton; unnamed European operator; iCruise\"], \"disclosed_results\": [{\"metric\": \"Average ROP increase across four sections\", \"value\": \"43.6%\", \"period\": \"Three-well pilot over two years\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS062\"]}, {\"metric\": \"Well delivery time improvement\", \"value\": \">30%\", \"period\": \"Three-well pilot over two years\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS062\"]}], \"economics\": \"Case study reports >30% improvement in well delivery time and 43.6% average ROP increase across four well sections; Halliburton-reported, three wells. No AI-only financial return.\", \"timeline\": [{\"date\": \"2025-Q1\", \"event\": \"Halliburton investor presentation includes LOGIX as autonomous drilling product.\", \"source_ids\": [\"S07\"]}, {\"date\": \"Undated; accessed 2026-09-27\", \"event\": \"Halliburton case study reports European operator pilot outcomes.\", \"source_ids\": [\"RS062\"]}], \"limitations\": [\"Case study operator and field not named. Vendor-reported results; no independent corroboration.\", \"The existing 2025 investor presentation is an excerpted product description; separate case study gives additional, vendor-reported workflow details.\"], \"assessment\": \"Analyst inference: Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.\", \"sources\": [{\"id\": \"S07\", \"title\": \"Halliburton Q1 2025 Investor Presentation\", \"url\": \"https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80\", \"publisher\": \"Halliburton\", \"date\": \"2025-Q1\", \"type\": \"investor presentation\", \"locator\": \"Slides 11 and 19\", \"review_status\": \"full text reviewed\", \"supports\": [\"LOGIX is presented as an autonomous drilling platform combining physics-based models and machine learning.\"], \"original_id\": \"hall_logix_presentation\"}, {\"id\": \"RS062\", \"title\": \"Enhanced well delivery with LOGIX automation and remote operations\", \"url\": \"https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop\", \"publisher\": \"Halliburton\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"partner case study\", \"locator\": \"Overview, solution, results\", \"review_status\": \"full text reviewed\", \"supports\": [\"An unnamed European operator piloted closed-loop directional drilling; remote center and rig personnel view data/advisories.\", \"Across three wells, Halliburton reports 43.6% average ROP increase and >30% well-delivery-time improvement.\"], \"original_id\": \"hall_logix_case\"}], \"field_sources\": {\"business_problem\": [\"S07\", \"RS062\"], \"workflow\": [\"S07\", \"RS062\"], \"technology_data\": [\"S07\", \"RS062\"], \"integration_human_role\": [\"S07\", \"RS062\"], \"deployment_scope\": [\"S07\", \"RS062\"], \"economics\": [\"S07\", \"RS062\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-025\", \"source_ids\": [\"S07\", \"RS062\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-016\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient.", "source_url": ""}], "Audit_review": {"id": "AI-016", "company": "Halliburton", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "LOGIX source establishes automation/autonomy but does not consistently identify the operating workflow as AI/ML.", "correction": "Describe the autonomous drilling workflow; retain AI evidence as unproven absent an explicit model disclosure.", "source_url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"}], "residual_gaps": ["The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-20", "Research_origin": ["research/watch-2026-10-20-update/README.md"]}, {"ID": "AI-017", "Company": "Halliburton", "Segment": "Oilfield services", "Theme": "AI products", "Initiative": "DecisionSpace 365 (DS365.ai) cloud E&P software", "Extracted_finding": "This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.", "Stage_at_disclosure": "Commercial product suite presented in 2025; no distinct AI module deployment verified", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Source is investor presentation excerpt and offers little technical detail. Do not infer a particular AI task, closed-loop control, or customer adoption from product-suite naming.", "Source_IDs": "S07", "Source_date": "Undated source; see chronology", "Source_type": "investor presentation", "Source_locator": "Slides 11 and 19", "Source_URL": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Give exploration and production teams cloud software and AI/ML-assisted insight across subsurface, drilling and production decisions.", "Implementation": "Investor material describes DS365.ai within cloud E&P services; it does not enumerate a specific model workflow or customer deployment for the AI module.", "Assessment": "Analyst inference: This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.", "Detail_words": 68, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Halliburton", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-017", "Initiative_ID": "AI-017", "Title": "DS365.ai", "Date": "2025-Q1", "Finding": "AI and ML augment subsurface, drilling and production decisions in cloud E&P services.", "Metric": "Not disclosed", "Source_IDs": "S07", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The broader suite runs on iEnergy and uses an open architecture. The investor material presents subscription services that can connect with other applications.", "Context": "DS365.ai adds AI and machine learning to subsurface, drilling and production workflows within Halliburton’s DecisionSpace 365 cloud services.", "Assessment": "The AI module sits within a larger software product family. The disclosure does not separate its users, revenue, deployment cost or economic return from the wider suite."}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-026"], "Research_search": " {\"key\": \"hall-ds365\", \"existing_ids\": [\"AI-017\"], \"related_existing_ids\": [\"AI-016\"], \"company\": \"Halliburton\", \"title\": \"DecisionSpace 365 (DS365.ai) cloud E&P software\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Commercial product suite presented in 2025; no distinct AI module deployment verified\", \"practical_relevance\": \"This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.\", \"business_problem\": \"Give exploration and production teams cloud software and AI/ML-assisted insight across subsurface, drilling and production decisions.\", \"workflow\": \"Investor material describes DS365.ai within cloud E&P services; it does not enumerate a specific model workflow or customer deployment for the AI module.\", \"technology_data\": \"DecisionSpace 365 subscription services, iEnergy cloud environment, open architecture; AI/ML functionality described at product-family level.\", \"integration_human_role\": \"Not described for the AI module in the reviewed investor presentation.\", \"deployment_scope\": \"Commercial software capability; no named customer or deployment level established.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Not disclosed. Product revenue, AI-specific use and returns not disclosed.\", \"timeline\": [{\"date\": \"2025-Q1\", \"event\": \"Investor presentation identifies DS365.ai within DecisionSpace 365 cloud services.\", \"source_ids\": [\"S07\"]}], \"limitations\": [\"Source is investor presentation excerpt and offers little technical detail.\", \"Do not infer a particular AI task, closed-loop control, or customer adoption from product-suite naming.\"], \"assessment\": \"Analyst inference: This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.\", \"sources\": [{\"id\": \"S07\", \"title\": \"Halliburton Q1 2025 Investor Presentation\", \"url\": \"https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80\", \"publisher\": \"Halliburton\", \"date\": \"2025-Q1\", \"type\": \"investor presentation\", \"locator\": \"Slides 11 and 19\", \"review_status\": \"full text reviewed\", \"supports\": [\"DecisionSpace 365 is described as a cloud E&P services portfolio with AI and machine learning spanning subsurface, drilling and production decisions.\"], \"original_id\": \"hall_ds365\"}, {\"id\": \"S07\", \"title\": \"Halliburton Q1 2025 Investor Presentation\", \"url\": \"https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80\", \"publisher\": \"Halliburton\", \"date\": \"2025-Q1\", \"type\": \"investor presentation\", \"locator\": \"Slides 11 and 19\", \"review_status\": \"full text reviewed\", \"supports\": [\"DecisionSpace 365 is described as cloud E&P services with AI/ML spanning subsurface, drilling and production decisions.\"], \"original_id\": \"hall_logix_presentation\"}], \"field_sources\": {\"business_problem\": [\"S07\"], \"workflow\": [\"S07\"], \"technology_data\": [\"S07\"], \"integration_human_role\": [\"S07\"], \"deployment_scope\": [\"S07\"], \"economics\": [\"S07\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-026\", \"source_ids\": [\"S07\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-017\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-017", "company": "Halliburton", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "Product label and cloud E&P scope do not prove individual customers use an AI feature.", "correction": "Separate DS365.ai product capability from deployment evidence.", "source_url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-018", "Company": "Halliburton / Shape Digital", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "Halliburton–Shape Digital asset performance collaboration", "Extracted_finding": "The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.", "Stage_at_disclosure": "Strategic collaboration announced July 2026; no customer deployment disclosed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Announced collaboration and intended workflow only; no named deployment or metrics. The announcement does not establish AI performance or autonomy.", "Source_IDs": "RS063; S08", "Source_date": "2026-07-21", "Source_type": "company announcement", "Source_locator": "Collaboration scope and connected production planning", "Source_URL": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-07-21; see chronology for operating dates.", "Context": "Connect subsurface and surface information so teams can make coordinated production, reliability, energy-efficiency and safety decisions.", "Implementation": "Landmark Digital Field Solver reservoir, well and production-network models provide context for Shape Lighthouse, Aura and Reef surface AI/engineering tools, aiming at predictive asset-level decisions through the production lifecycle.", "Assessment": "Analyst inference: The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.", "Detail_words": 65, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Halliburton", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-018", "Initiative_ID": "AI-018", "Title": "Asset performance collaboration", "Date": "2026-07-21", "Finding": "Unifies subsurface and surface information for predictive asset decisions.", "Metric": "Not disclosed", "Source_IDs": "S08", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2026 earnings release describes a unified asset view connecting subsurface and surface information, combining domain science with applied AI for predictive decisions over the production lifecycle.", "Context": "Halliburton and Shape Digital announced a strategic collaboration to improve digital asset performance management.", "Assessment": "The announcement explains intended workflow integration. Named customer deployments, implementation milestones and measured reliability or production outcomes are absent from the reviewed disclosure."}], "Latest_disclosure": "2026-07-21", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-027"], "Research_search": " {\"key\": \"hall-shape-collaboration\", \"existing_ids\": [\"AI-018\"], \"related_existing_ids\": [], \"company\": \"Halliburton\", \"title\": \"Halliburton–Shape Digital asset performance collaboration\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Strategic collaboration announced July 2026; no customer deployment disclosed\", \"practical_relevance\": \"The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.\", \"business_problem\": \"Connect subsurface and surface information so teams can make coordinated production, reliability, energy-efficiency and safety decisions.\", \"workflow\": \"Landmark Digital Field Solver reservoir, well and production-network models provide context for Shape Lighthouse, Aura and Reef surface AI/engineering tools, aiming at predictive asset-level decisions through the production lifecycle.\", \"technology_data\": \"Trusted operational data, domain models, Shape Digital applied AI portfolio, integrated asset and production models.\", \"integration_human_role\": \"Decision-support and predictive insights described; autonomous execution or decision authority not specified.\", \"deployment_scope\": \"Collaboration announced; no named customer, asset, deployment schedule, or field scale in source.\", \"partners\": [\"Shape Digital\"], \"disclosed_results\": [], \"economics\": \"No customer outcomes or separate commercial terms disclosed.\", \"timeline\": [{\"date\": \"2026-07-21\", \"event\": \"Strategic collaboration announced in Halliburton Q2 results and release.\", \"source_ids\": [\"RS063\"]}], \"limitations\": [\"Announced collaboration and intended workflow only; no named deployment or metrics.\", \"The announcement does not establish AI performance or autonomy.\"], \"assessment\": \"Analyst inference: The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.\", \"sources\": [{\"id\": \"RS063\", \"title\": \"Halliburton and Shape Digital form strategic collaboration to advance digital asset performance management\", \"url\": \"https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management\", \"publisher\": \"Halliburton\", \"date\": \"2026-07-21\", \"type\": \"company announcement\", \"locator\": \"Collaboration scope and connected production planning\", \"review_status\": \"full text reviewed\", \"supports\": [\"The collaboration combines Landmark Digital Field Solver models with Shape Lighthouse, Aura and Reef to connect subsurface and surface intelligence for predictive asset decisions.\"], \"original_id\": \"hall_shape\"}], \"field_sources\": {\"business_problem\": [\"RS063\"], \"workflow\": [\"RS063\"], \"technology_data\": [\"RS063\"], \"integration_human_role\": [\"RS063\"], \"deployment_scope\": [\"RS063\"], \"economics\": [\"RS063\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-027\", \"source_ids\": [\"RS063\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-018\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-018", "company": "Halliburton / Shape Digital", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management"], "findings": [{"severity": "medium", "field": "stage", "issue": "Strategic collaboration announcement describes planned joint development; no named customer deployment or realized outcome is supplied.", "correction": "Stage as collaboration/product development, not an operating deployment.", "source_url": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-019", "Company": "Baker Hughes", "Segment": "Oilfield services / equipment", "Theme": "AI infrastructure", "Initiative": "Data-center orders", "Extracted_finding": "Raised order target as AI-related data-center demand accelerated.", "Stage_at_disclosure": "Commercial / target", "Reported_metric": "Three-year order target $3bn, previously $1.5bn", "Interpretation_limit": "Target, not AI software revenue or realized orders.", "Source_IDs": "S09", "Source_date": "2026-01-26", "Source_type": "Earnings presentation", "Source_locator": "Slide 6", "Source_URL": "https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Baker Hughes is selling power equipment into data-center demand, including demand associated with AI.", "Implementation": "Its January 2026 presentation reports $1 billion of data-center power orders in 2025 and a $3 billion target for 2026–2028, raised from $1.5 billion. A roughly 1 GW NovaLT slot reservation is a separate item.", "Assessment": "Orders, reservations and future targets have different commitment levels. These amounts concern equipment and power infrastructure, and should not be labeled AI software revenue.", "Detail_words": 73, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Baker Hughes", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-019", "Initiative_ID": "AI-019", "Title": "Data-center orders", "Date": "2026-01-26", "Finding": "Raised order target as AI-related data-center demand accelerated.", "Metric": "Three-year order target $3bn, previously $1.5bn", "Source_IDs": "S09", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Its January 2026 presentation reports $1 billion of data-center power orders in 2025 and a $3 billion target for 2026–2028, raised from $1.5 billion. A roughly 1 GW NovaLT slot reservation is a separate item.", "Context": "Baker Hughes is selling power equipment into data-center demand, including demand associated with AI.", "Assessment": "Orders, reservations and future targets have different commitment levels. These amounts concern equipment and power infrastructure, and should not be labeled AI software revenue."}], "Latest_disclosure": "2026-01-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-019", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "target", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-020", "Company": "Baker Hughes / EPAM", "Segment": "Oilfield services / software", "Theme": "AI products", "Initiative": "Cordant Asset Strategy AI with EPAM", "Extracted_finding": "A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.", "Stage_at_disclosure": "Partner-described case / feature; current product is commercial", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Evidence is partner webinar synopsis, not a detailed customer deployment record. No named operator, outcome metric or model governance details disclosed.", "Source_IDs": "S10; R080", "Source_date": "2025-04-15", "Source_type": "partner webinar description", "Source_locator": "Description and key learnings", "Source_URL": "https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-04-15; see chronology for operating dates.", "Context": "Reduce slow data preparation and specialist effort in asset reliability strategy development, including analysis of engineering documents and unstructured text.", "Implementation": "Partner webinar describes use of generative AI and text processing to support Cordant Asset Strategy and engineering document analysis in asset-performance workflows.", "Assessment": "Analyst inference: A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.", "Detail_words": 68, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Baker Hughes", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-080"], "Related_IDs": [], "Disclosures": [{"ID": "REF-080", "Initiative_ID": "AI-020", "Title": "Cordant built on Azure AI Foundry with Microsoft", "Date": "2025-02", "Finding": "Baker Hughes signed an MOU with Microsoft to speed development and adoption of its Cordant industrial software suite, which is built with Azure AI Foundry. It targets predictive maintenance, operations optimization, safety and energy efficiency, and later releases (25.2.1, 26.1) added AI-driven risk intelligence.", "Metric": "No quantified metric supplied", "Source_IDs": "R080", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Incumbents are standardizing on hyperscaler agent frameworks, so building on the same foundry or model stack makes integration and co-selling easier.", "Partners": "Microsoft Azure AI Foundry, Cordant", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "ORIG-AI-020", "Initiative_ID": "AI-020", "Title": "Cordant asset strategy", "Date": "2025-04-15", "Finding": "Azure AI supports asset strategy, unstructured data processing and engineering document analysis.", "Metric": "Not disclosed", "Source_IDs": "S10", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The partner webinar discusses generative AI, unstructured text processing and Azure services, including Azure Data Factory in SAP S/4HANA environments.", "Context": "Baker Hughes and EPAM describe AI features for Cordant Asset Strategy, addressing slow data preparation and the specialist knowledge required for reliability engineering.", "Assessment": "This is a partner-described implementation case. The public webinar description does not establish deployment scale, measured time savings or the accuracy of generated asset strategies."}], "Latest_disclosure": "2025-04-15", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-028"], "Research_search": " {\"key\": \"baker-cordant-epam\", \"existing_ids\": [\"AI-020\"], \"related_existing_ids\": [], \"company\": \"Baker Hughes\", \"title\": \"Cordant Asset Strategy AI with EPAM\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Partner-described case / feature; current product is commercial\", \"practical_relevance\": \"A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.\", \"business_problem\": \"Reduce slow data preparation and specialist effort in asset reliability strategy development, including analysis of engineering documents and unstructured text.\", \"workflow\": \"Partner webinar describes use of generative AI and text processing to support Cordant Asset Strategy and engineering document analysis in asset-performance workflows.\", \"technology_data\": \"Azure AI services and Azure Data Factory integration within SAP S/4HANA environments, according to partner webinar description.\", \"integration_human_role\": \"Reliability engineers use outputs for strategy development; decision approval and generated-content validation details not published in reviewed source description.\", \"deployment_scope\": \"Partner-described use case; no customer/site or scale identified in the webinar page reviewed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No measured time saved, accuracy, cost or ROI disclosed.\", \"timeline\": [{\"date\": \"2025-04-15\", \"event\": \"EPAM webinar describes Baker Hughes Cordant Asset Strategy and Azure AI case.\", \"source_ids\": [\"S10\"]}], \"limitations\": [\"Evidence is partner webinar synopsis, not a detailed customer deployment record.\", \"No named operator, outcome metric or model governance details disclosed.\"], \"assessment\": \"Analyst inference: A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.\", \"sources\": [{\"id\": \"S10\", \"title\": \"Speed to Value in Asset Performance with Azure AI: a case study on EPAM and Baker Hughes\", \"url\": \"https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes\", \"publisher\": \"EPAM\", \"date\": \"2025-04-15\", \"type\": \"partner webinar description\", \"locator\": \"Description and key learnings\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"EPAM describes generative AI and unstructured text processing for Cordant Asset Strategy, including engineering document analysis and Azure integration.\"], \"original_id\": \"baker_epam\"}], \"field_sources\": {\"business_problem\": [\"S10\"], \"workflow\": [\"S10\"], \"technology_data\": [\"S10\"], \"integration_human_role\": [\"S10\"], \"deployment_scope\": [\"S10\"], \"economics\": [\"S10\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-028\", \"source_ids\": [\"S10\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-020\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-020", "company": "Baker Hughes / EPAM", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Baker Hughes Cordant announcement URL returned 403 during direct reopen.", "correction": "Mark this supplemental URL not reverified; do not treat access failure as contradiction. Preserve any earlier review status separately from this pass.", "source_url": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-021", "Company": "Baker Hughes / C3 AI", "Segment": "Oilfield services / software", "Theme": "AI products", "Initiative": "Baker Hughes–C3 AI enterprise AI joint venture renewal", "Extracted_finding": "Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.", "Stage_at_disclosure": "Multi-year joint venture agreement renewed through June 2028 in May 2025", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Partnership evidence is strong; operational workflow, deployment scale and results are not detailed in renewal announcement. Category reflects strategic partnership rather than a newly verified deployment.", "Source_IDs": "RS064; S11; S12; R079", "Source_date": "2025-05-28", "Source_type": "company announcement", "Source_locator": "Agreement scope", "Source_URL": "https://ir.c3.ai/node/10916/pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-05-28; see chronology for operating dates.", "Context": "Continue joint development, co-selling and scaling of enterprise AI applications across energy and industrial asset operations.", "Implementation": "The renewal extends the existing joint venture and product/commercial collaboration; announcement provides no newly named workflow or operator deployment.", "Assessment": "Analyst inference: Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.", "Detail_words": 67, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Baker Hughes", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-079"], "Related_IDs": [], "Disclosures": [{"ID": "REF-079", "Initiative_ID": "AI-021", "Title": "BakerHughesC3.ai JV renewed through 2028", "Date": "2025-05", "Finding": "C3 AI and Baker Hughes renewed and expanded their joint venture through June 2028 to sell enterprise AI to oil, gas and chemicals, integrated with Cordant. Baker Hughes will also expand internal use of C3 sourcing, inventory and sustainability apps.", "Metric": "JV has generated >$500M revenue from O&G and chemicals since 2019; 5th amendment", "Source_IDs": "R079", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Distribution through an oilfield services company moved more than $500M of AI software, so a channel JV with a service company can be worth more than direct sales.", "Partners": "C3 Agentic AI Platform, Cordant, Shell, Eni, QatarEnergy LNG, Petronas, ExxonMobil, LyondellBasell, Flint Hills", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "ORIG-AI-021", "Initiative_ID": "AI-021", "Title": "Enterprise AI partnership renewal", "Date": "2025-06", "Finding": "Renewed multi-year relationship covering co-selling, joint investment and deployment scaling.", "Metric": "Not disclosed", "Source_IDs": "S11; S12", "Evidence": "Primary filing indexed passage reviewed", "Origin": "Original register", "Implementation": "The filings and earnings material describe co-selling, joint investment and scaling. The commercial relationship should be read against the actual renewal terms.", "Context": "Baker Hughes and C3 AI renewed a multi-year enterprise AI relationship intended to support joint sales and deployment.", "Assessment": "A renewal establishes continued commercial cooperation. It does not, by itself, confirm that earlier exclusive oil-and-gas resale rights remain unchanged or that every deployment generated a return."}], "Latest_disclosure": "2025-05-28", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-029"], "Research_search": " {\"key\": \"baker-c3-partnership\", \"existing_ids\": [\"AI-021\"], \"related_existing_ids\": [], \"company\": \"Baker Hughes\", \"title\": \"Baker Hughes–C3 AI enterprise AI joint venture renewal\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Strategy only\", \"stage\": \"Multi-year joint venture agreement renewed through June 2028 in May 2025\", \"practical_relevance\": \"Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.\", \"business_problem\": \"Continue joint development, co-selling and scaling of enterprise AI applications across energy and industrial asset operations.\", \"workflow\": \"The renewal extends the existing joint venture and product/commercial collaboration; announcement provides no newly named workflow or operator deployment.\", \"technology_data\": \"C3 AI enterprise application software combined with Baker Hughes energy and industrial domain capabilities.\", \"integration_human_role\": \"Customer decision workflows and human controls not specified in the renewal announcement.\", \"deployment_scope\": \"Commercial alliance renewal; no deployment count or operational rollout detail in this disclosure.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Agreement value, revenue allocation and investment split not disclosed.\", \"timeline\": [{\"date\": \"2025-05-28\", \"event\": \"C3 AI and Baker Hughes announced multi-year renewal/expansion through June 2028.\", \"source_ids\": [\"RS064\"]}], \"limitations\": [\"Partnership evidence is strong; operational workflow, deployment scale and results are not detailed in renewal announcement.\", \"Category reflects strategic partnership rather than a newly verified deployment.\"], \"assessment\": \"Analyst inference: Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.\", \"sources\": [{\"id\": \"RS064\", \"title\": \"C3 AI and Baker Hughes Renew and Expand Joint Venture Agreement\", \"url\": \"https://ir.c3.ai/node/10916/pdf\", \"publisher\": \"C3 AI\", \"date\": \"2025-05-28\", \"type\": \"company announcement\", \"locator\": \"Agreement scope\", \"review_status\": \"full text reviewed\", \"supports\": [\"C3 AI and Baker Hughes renewed and expanded their joint venture agreement for multiple years through June 2028.\"], \"original_id\": \"c3_baker_renewal\"}], \"field_sources\": {\"business_problem\": [\"RS064\"], \"workflow\": [\"RS064\"], \"technology_data\": [\"RS064\"], \"integration_human_role\": [\"RS064\"], \"deployment_scope\": [\"RS064\"], \"economics\": [\"RS064\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-029\", \"source_ids\": [\"RS064\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-021\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-021", "company": "Baker Hughes / C3 AI", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.c3.ai/node/10916/pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "C3 AI FY2025 Form 10-K was reopened. It verifies the April 2025 renewal, co-selling/co-investment language, and that Baker Hughes is no longer exclusive reseller; separate May 28 issuer release adds term through June 2028, named deployed customers and internal-use applications. Benefits and customer scope remain vendor-reported/aggregate.", "correction": "", "source_url": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm"}, {"severity": "low", "field": "Supplemental source review", "issue": "The C3 AI FY2026 Form 10-K says Baker Hughes is no longer C3 AI’s exclusive reseller after April 2025 amendment; the filing still describes a strategic relationship.", "correction": "Retain the multi-year relationship/renewal as reported, but update reseller status to non-exclusive from April 2025 onward. Do not describe the relationship as terminated.", "source_url": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-022", "Company": "ExxonMobil", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Permian and operating-site AI", "Extracted_finding": "ExxonMobil CEO separately describes AI deployment. Vantage context: company reports >22,000 U.S. sites monitored, >1m b/d managed/optimized and >29,000 b/d uplift from broad field/remote optimization activities; the Vantage article does not attribute those outcomes to AI.", "Stage_at_disclosure": "Deployed", "Reported_metric": "Vantage context (non-AI-attributed): >22,000 sites monitored; >1m b/d managed/optimized; >29,000 b/d field/remote optimization uplift, period unstated; >20m cubic feet of flaring eliminated and >6,000 driving hours avoided in 2024.", "Interpretation_limit": "Vantage article gives operating-context outcomes but no AI attribution; do not count as AI benefits. Oil-uplift period is not stated.", "Source_IDs": "S13; OCT5-S049", "Source_date": "2025-10-31", "Source_type": "Earnings call transcript", "Source_locator": "Darren Woods response to Paul Sankey", "Source_URL": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "Evidence_status": "Third-party-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "ExxonMobil’s CEO describes AI operating downhole in the Permian and at company sites around the world.", "Implementation": "The Q3 2025 call connects learning from operating data with production optimization. The reproduced transcript gives no named model, vendor, field count or controlled result.", "Assessment": "This provides management confirmation of deployment but little implementation detail. It cannot support a claim that a particular percentage of Permian improvement came from AI.", "Detail_words": 66, "Primary_department": "Production & field operations", "Secondary_departments": ["Reservoir engineering & field development", "Maintenance & reliability"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ExxonMobil", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-022", "Initiative_ID": "AI-022", "Title": "Permian and operating-site AI", "Date": "2025-10-31", "Finding": "CEO describes AI deployed downhole and at global operating sites to learn from data and optimize production.", "Metric": "No standalone metric", "Source_IDs": "S13", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q3 2025 call connects learning from operating data with production optimization. The reproduced transcript gives no named model, vendor, field count or controlled result.", "Context": "ExxonMobil’s CEO describes AI operating downhole in the Permian and at company sites around the world.", "Assessment": "This provides management confirmation of deployment but little implementation detail. It cannot support a claim that a particular percentage of Permian improvement came from AI."}, {"ID": "OCT5-AI-022-1", "Initiative_ID": "AI-022", "Title": "Introducing Vantage: A new lens on upstream operations", "Date": "2025-10-10", "Finding": "ExxonMobil says Vantage monitors >22,000 U.S. sites, manages/optimizes >1m b/d oil, and field/remote operations production optimization realized >29,000 b/d uplift. The article separately reports >20m cubic feet of flaring eliminated in 2024 and >6,000 driving hours avoided in 2024. These describe Vantage/field operations; the source does not attribute the outcomes to AI.", "Metric": "Company-reported operations context: >22,000 sites monitored; >1m b/d managed/optimized; >29,000 b/d uplift from field/remote optimization (period not stated); >20m cubic feet flaring eliminated and >6,000 driving hours avoided in 2024.", "Source_IDs": "OCT5-S049", "Source_URL": "https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Latest_disclosure": "2025-10-31", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete · Supplemented 6 October 2026", "Research_packets": ["oct6-executive-ai-022"], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-022", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Keep Vantage operating-scale/benefit metrics as non-AI contextual figures; do not associate with the separately described AI deployment.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Research_search": "{\"packet_id\": \"oct6-executive-ai-022\", \"review_date\": \"2026-10-06\", \"title\": \"Introducing Vantage: A new lens on upstream operations\", \"publisher\": \"ExxonMobil\", \"source_title\": \"Introducing Vantage: A new lens on upstream operations\", \"source_date\": \"2025-10-10\", \"source_type\": \"Operator article\", \"source_locator\": \"“Big numbers, bigger impact,” lines 54–64\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"Vantage centralizes remote upstream operations. Its disclosed scale and field-optimization results provide operating context; the source does not attribute those results to AI.\", \"technology_data\": \"The issuer article describes remote operations, centralized monitoring and optimization. It does not identify an AI model or separate the effect of AI from other operational changes.\", \"deployment_scope\": \"Company reports centralized real-time upstream monitoring/management; this is context for AI-022.\", \"practical_relevance\": \"Vantage helps explain ExxonMobil’s operating setup. Its broader optimization metrics cannot be used as an AI return without a separate attribution bridge.\", \"business_problem\": \"Monitoring and coordinating a large upstream asset base.\", \"economics\": \"No AI-specific economics; full implementation/operating costs not disclosed.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Do not count metrics as AI-caused benefits; source covers broad operations optimization.\"], \"timeline_event\": \"Vantage article dated 2025-10-10; oil-uplift measurement period not stated.\", \"timeline_date\": \"2025-10-10\", \"source_ids\": [\"OCT5-S049\"], \"stage\": \"Deployed\", \"key\": \"oct6-executive-ai-022\", \"record_ids\": [\"AI-022\"], \"company\": \"ExxonMobil\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S049\"], \"workflow\": [\"OCT5-S049\"], \"technology_data\": [\"OCT5-S049\"], \"deployment_scope\": [\"OCT5-S049\"], \"practical_relevance\": [\"OCT5-S049\"], \"economics\": [\"OCT5-S049\"], \"integration_human_role\": [\"OCT5-S049\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \"Company-reported values are not AI-attributed: 2024 flaring/driving data specified; oil uplift period unstated.\", \"period\": \"2025-10-10\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S049\"]}], \"timeline\": [{\"date\": \"2025-10-10\", \"event\": \"Vantage article dated 2025-10-10; oil-uplift measurement period not stated.\", \"source_ids\": [\"OCT5-S049\"]}], \"sources\": [{\"id\": \"OCT5-S049\", \"title\": \"Introducing Vantage: A new lens on upstream operations\", \"url\": \"https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations\", \"date\": \"2025-10-10\", \"type\": \"Operator article\", \"publisher\": \"ExxonMobil\", \"locator\": \"“Big numbers, bigger impact,” lines 54–64\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}]}", "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Third-party-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-023", "Company": "ExxonMobil", "Segment": "Integrated", "Theme": "AI infrastructure", "Initiative": "Low-carbon data-center opportunity", "Extracted_finding": "Pursuing data-center decarbonization opportunities using capabilities associated with Denbury.", "Stage_at_disclosure": "Development opportunity", "Reported_metric": "Not disclosed", "Interpretation_limit": "Not a final investment decision.", "Source_IDs": "S13", "Source_date": "2025-10-31", "Source_type": "Earnings call transcript", "Source_locator": "Darren Woods response to Paul Sankey", "Source_URL": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "Evidence_status": "Third-party-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "ExxonMobil discussed lower-carbon power opportunities serving data centers, drawing on capabilities associated with the Denbury acquisition.", "Implementation": "The Q3 2025 call frames these as development opportunities. The evidence reviewed here does not identify a completed project-level investment decision.", "Assessment": "The economic exposure would come through power and carbon-management services. The discussion does not establish AI model deployment inside ExxonMobil or contracted returns from a named campus.", "Detail_words": 64, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ExxonMobil", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-023", "Initiative_ID": "AI-023", "Title": "Low-carbon data-center opportunity", "Date": "2025-10-31", "Finding": "Pursuing data-center decarbonization opportunities using capabilities associated with Denbury.", "Metric": "Not disclosed", "Source_IDs": "S13", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q3 2025 call frames these as development opportunities. The evidence reviewed here does not identify a completed project-level investment decision.", "Context": "ExxonMobil discussed lower-carbon power opportunities serving data centers, drawing on capabilities associated with the Denbury acquisition.", "Assessment": "The economic exposure would come through power and carbon-management services. The discussion does not establish AI model deployment inside ExxonMobil or contracted returns from a named campus."}], "Latest_disclosure": "2025-10-31", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [{"claim": "Low-carbon data-center opportunity as internal AI workflow.", "reason": "Data-center power/carbon solution is infrastructure and commercial opportunity in the AI-demand ecosystem.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"}], "Audit_review": {"id": "AI-023", "company": "ExxonMobil", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"], "findings": [{"severity": "medium", "field": "category / classification", "issue": "The disclosed data-center opportunity concerns energy supply for AI demand, not ExxonMobil adopting AI in its own workflow.", "correction": "Classify as AI-demand infrastructure/strategic opportunity and keep distinct from internal AI adoption.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "infrastructure", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Third-party-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-024", "Company": "ExxonMobil", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "ERP and logistics AI", "Extracted_finding": "Standardizing enterprise systems facilitates AI applications such as marine logistics planning.", "Stage_at_disclosure": "Implementation / enabling platform", "Reported_metric": "No AI-only benefit disclosed", "Interpretation_limit": "ERP simplification metrics are not AI savings.", "Source_IDs": "S14; R012", "Source_date": "2026-01-30", "Source_type": "Earnings call transcript", "Source_locator": "Kathryn Mikells response on SAP / ERP transformation", "Source_URL": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "Evidence_status": "Third-party-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "ExxonMobil’s enterprise-system standardization creates a more consistent data and process environment for AI applications.", "Implementation": "The Q4 2025 call gives marine logistics planning as an example of AI-enabled work supported by the ERP transformation.", "Assessment": "The ERP program is an enabling investment with broader benefits. System simplification counts and enterprise savings should not be reclassified as a measured AI return.", "Detail_words": 58, "Primary_department": "Supply chain & logistics", "Secondary_departments": ["Finance, investment & strategy", "IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ExxonMobil", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-012"], "Related_IDs": [], "Disclosures": [{"ID": "REF-012", "Initiative_ID": "AI-024", "Title": "AI plus process mining behind $15B cost-cut push", "Date": "2024-11", "Finding": "ExxonMobil tied its $15B operating-cost reduction target for 2027 to AI combined with process intelligence run through its Global Business Solutions group. It uses Celonis process mining across procurement, supply chain, logistics, maintenance and AP/AR, and the CFO acknowledged that historically siloed, inconsistent processes had held back automation.", "Metric": "$15B cost-savings target by 2027", "Source_IDs": "R012", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardize and mine processes before layering on AI agents; back-office AI ROI depends on clean end-to-end process data.", "Partners": "Celonis, AWS, Snowflake, SAP, Salesforce", "Original_date": "2024-11", "Related_IDs": []}, {"ID": "ORIG-AI-024", "Initiative_ID": "AI-024", "Title": "ERP and logistics AI", "Date": "2026-01-30", "Finding": "Standardizing enterprise systems facilitates AI applications such as marine logistics planning.", "Metric": "No AI-only benefit disclosed", "Source_IDs": "S14", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q4 2025 call gives marine logistics planning as an example of AI-enabled work supported by the ERP transformation.", "Context": "ExxonMobil’s enterprise-system standardization creates a more consistent data and process environment for AI applications.", "Assessment": "The ERP program is an enabling investment with broader benefits. System simplification counts and enterprise savings should not be reclassified as a measured AI return."}], "Latest_disclosure": "2026-01-30", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "ERP simplification and marine logistics planning AI.", "reason": "Original reference is a third-party transcript; no Exxon primary source for a named AI workflow was located in this bounded pass. ERP simplification is not itself AI. This is not a disproval; it remains unverified.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-024", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Third-party-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-025", "Company": "Shell", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Technology and AI cost reduction", "Extracted_finding": "AI is one component of structural simplification and cost reduction.", "Stage_at_disclosure": "Program / target", "Reported_metric": "Total structural reduction target $5–7bn, 2022–2028", "Interpretation_limit": "Target includes portfolio and non-AI actions.", "Source_IDs": "S15", "Source_date": "2025-03", "Source_type": "Investor presentation", "Source_locator": "Slides 15 and 46", "Source_URL": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Shell includes technology and AI within its structural simplification and cost-reduction plans.", "Implementation": "Capital Markets Day 2025 presents a company-wide structural reduction target of $5–7 billion over 2022–2028. Portfolio actions and other non-AI measures also contribute.", "Assessment": "This is evidence that AI participates in a cost program. The disclosure provides no separate AI savings bridge, making the full corporate target unsuitable as an AI benefit estimate.", "Detail_words": 64, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Shell", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-025", "Initiative_ID": "AI-025", "Title": "Technology and AI cost reduction", "Date": "2025-03", "Finding": "AI is one component of structural simplification and cost reduction.", "Metric": "Total structural reduction target $5–7bn, 2022–2028", "Source_IDs": "S15", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Capital Markets Day 2025 presents a company-wide structural reduction target of $5–7 billion over 2022–2028. Portfolio actions and other non-AI measures also contribute.", "Context": "Shell includes technology and AI within its structural simplification and cost-reduction plans.", "Assessment": "This is evidence that AI participates in a cost program. The disclosure provides no separate AI savings bridge, making the full corporate target unsuitable as an AI benefit estimate."}], "Latest_disclosure": "2025-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Shell $5-7bn target as AI-generated savings.", "reason": "Shell lists AI among examples within a multicomponent $5-7bn structural-cost-reduction target from 2022–2028; no AI-only attribution.", "source_url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-025", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Company-wide structural cost-reduction target; AI is one of several levers and no AI share is disclosed.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-026", "Company": "Shell", "Segment": "Integrated", "Theme": "AI infrastructure", "Initiative": "Immersion cooling fluids", "Extracted_finding": "GTL-derived fluids support AI data-center cooling.", "Stage_at_disclosure": "Commercial product", "Reported_metric": "Up to 33% claimed improvement; footnote describes cost savings", "Interpretation_limit": "Product claim; efficiency and cost language differ within deck.", "Source_IDs": "S15", "Source_date": "2025-03", "Source_type": "Investor presentation", "Source_locator": "Slides 15 and 46", "Source_URL": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "Shell markets gas-to-liquids-derived immersion cooling fluids for data-center applications, including AI computing.", "Implementation": "Its investor presentation cites an improvement of up to 33%, while the associated footnote describes cost savings. The wording does not support treating that figure as a universal efficiency gain.", "Assessment": "The commercial route is specialty fluids. Assess the exact test conditions and cost denominator before using the percentage in a customer business case.", "Detail_words": 65, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Research & innovation"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Shell", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-026", "Initiative_ID": "AI-026", "Title": "Immersion cooling fluids", "Date": "2025-03", "Finding": "GTL-derived fluids support AI data-center cooling.", "Metric": "Up to 33% claimed improvement; footnote describes cost savings", "Source_IDs": "S15", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Its investor presentation cites an improvement of up to 33%, while the associated footnote describes cost savings. The wording does not support treating that figure as a universal efficiency gain.", "Context": "Shell markets gas-to-liquids-derived immersion cooling fluids for data-center applications, including AI computing.", "Assessment": "The commercial route is specialty fluids. Assess the exact test conditions and cost denominator before using the percentage in a customer business case."}], "Latest_disclosure": "2025-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [{"claim": "Immersion cooling fluids as oil-and-gas AI adoption.", "reason": "AI data center cooling product / infrastructure market, not AI use by Shell operating functions.", "source_url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-026", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-027", "Company": "Eni", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "AI agents, digital plant and drilling automation portfolio", "Extracted_finding": "Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.", "Stage_at_disclosure": "2025 outcomes and targets reported in Eni for 2025", "Reported_metric": "Use cases: ~300 (+40% vs. 2024) (Company-reported); Digital plant target: 2–3% production uplift and emissions reduction (Company-reported); Drilling automation target: 35% less drilling time (Company-reported)", "Interpretation_limit": "Targets can be misread as actual results; metrics are company-reported targets. 23 October scope limit: Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications.", "Source_IDs": "RS009; S16; R026; WATCH23AM-SLB-Q3", "Source_date": "2026-04", "Source_type": "annual report", "Source_locator": "Data and artificial intelligence", "Source_URL": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-04; see chronology for operating dates.", "Context": "Industrial plants need improved uptime/energy efficiency and drilling must reduce cycle time while maintaining safety.", "Implementation": "Eni reports AI applications for digital plant optimization and drilling automation; AI agents are developing across portfolio.\n\n23 October supplier update: SLB says Eni expanded petrophysics and geomechanics AI workflows using an on-premise Lumi platform.", "Assessment": "Analyst inference: Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.", "Detail_words": 71, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Eni", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-026"], "Related_IDs": [], "Disclosures": [{"ID": "REF-026", "Initiative_ID": "AI-027", "Title": "HPC6 supercomputer and company-wide responsible-AI training", "Date": "2025-06", "Finding": "Eni runs HPC6, launched November 2024 and ranked in the global TOP10, for AI-driven geoscience, safety (the 'Safety Pre-sense' tool) and other uses, and has opened HPC6 access to startups through Call4Innovators. From 2025 it rolled out a company-wide training program on responsible AI, regulation, ethics and data literacy, with advanced modules.", "Metric": "HPC6 in global TOP10; 10 Call4Innovators projects selected (Dec 2025)", "Source_IDs": "R026", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Owning compute is a differentiator for subsurface AI; pairing it with mandatory AI literacy helps meet EU AI Act obligations.", "Partners": "HPE, AMD, CINECA, Plug and Play", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "ORIG-AI-027", "Initiative_ID": "AI-027", "Title": "ENI AI portfolio", "Date": "2026", "Finding": "Disclosed expanding use-case portfolio and AI agents.", "Metric": "Around 300 use cases; +40% vs 2024; about 20 agents", "Source_IDs": "S16", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The presentation places AI alongside computing and industrial digitalization initiatives. It does not provide a complete application-level deployment list.", "Context": "Eni’s 2026 Capital Markets Update describes an expanding portfolio of approximately 300 AI use cases, up about 40% from 2024, plus around 20 agents.", "Assessment": "Use-case counts measure portfolio activity. Nearby digital-plant and drilling-automation figures cannot automatically be attributed to all AI use cases, and agent count does not measure usage or value."}, {"ID": "WATCH23AM-AI-027", "Initiative_ID": "AI-027", "Title": "Eni adds a concrete on-premise subsurface workflow", "Date": "2026-10-23", "Finding": "SLB says Eni expanded petrophysics and geomechanics AI workflows using an on-premise Lumi platform.", "Metric": "No task-level benefit, asset population or full cost quantified in this update. Eni’s existing portfolio metrics and targets remain separately sourced; none is allocated to Lumi.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications."}], "Latest_disclosure": "2026-10-23", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-014"], "Research_search": " {\"key\": \"eni-ai-portfolio\", \"existing_ids\": [\"AI-027\"], \"company\": \"Eni\", \"title\": \"AI agents, digital plant and drilling automation portfolio\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"2025 outcomes and targets reported in Eni for 2025\", \"practical_relevance\": \"Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.\", \"business_problem\": \"Industrial plants need improved uptime/energy efficiency and drilling must reduce cycle time while maintaining safety.\", \"workflow\": \"Eni reports AI applications for digital plant optimization and drilling automation; AI agents are developing across portfolio.\", \"technology_data\": \"About 300 AI use cases, about 20 AI agents; specific models/data and production coverage absent.\", \"integration_human_role\": \"Human role, automation scope and workflow approval details not described in summary disclosure.\", \"deployment_scope\": \"Company-wide, with digital plant and drilling applications; specific sites not named.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Use cases\", \"value\": \"~300 (+40% vs. 2024)\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS009\"]}, {\"metric\": \"Digital plant target\", \"value\": \"2–3% production uplift and emissions reduction\", \"period\": \"target\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS009\"]}, {\"metric\": \"Drilling automation target\", \"value\": \"35% less drilling time\", \"period\": \"target\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS009\"]}], \"economics\": \"Targets: digital plant 2–3% production enhancement and 2–3% emissions reduction; drilling automation 35% reduction in drilling time. These are targets, not verified realized results.\", \"timeline\": [{\"date\": \"2026-04-22\", \"event\": \"Eni for 2025 reports use cases, agents and expected impacts.\", \"source_ids\": [\"RS009\"]}], \"limitations\": [\"Targets can be misread as actual results; metrics are company-reported targets.\"], \"assessment\": \"Analyst inference: Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.\", \"sources\": [{\"id\": \"RS009\", \"title\": \"Eni for 2025 — Just Transition\", \"url\": \"https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf\", \"publisher\": \"Eni\", \"date\": \"2026-04\", \"type\": \"annual report\", \"locator\": \"Data and artificial intelligence\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Eni reports around 300 AI use cases (+40% vs. 2024), around 20 AI agents; digital plant target ranges of 2–3% production enhancement and emissions reduction and drilling automation target of 35% less drilling time.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-014\"}], \"field_sources\": {\"business_problem\": [\"RS009\"], \"workflow\": [\"RS009\"], \"technology_data\": [\"RS009\"], \"integration_human_role\": [\"RS009\"], \"deployment_scope\": [\"RS009\"], \"economics\": [\"RS009\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-014\", \"source_ids\": [\"RS009\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-027\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Broad Eni AI portfolio/value claims from Capital Markets Update.", "reason": "Existing official PDF source remains not re-opened in this cohort pass; no contradiction found, but portfolio-level claims not re-verified against its source text.", "source_url": "https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf"}], "Audit_review": {"id": "AI-027", "company": "Eni", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Targets can be misread as actual results; metrics are company-reported targets. 23 October scope limit: Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-23"}, {"ID": "AI-028", "Company": "Eni / Khazna", "Segment": "Integrated / infrastructure", "Theme": "AI infrastructure", "Initiative": "Ferrera Erbognone campus", "Extracted_finding": "Heads of terms for a joint venture developing an AI data-center campus.", "Stage_at_disclosure": "Heads of terms / development", "Reported_metric": "500 MW IT capacity", "Interpretation_limit": "Not operating capacity; distinct from gas-generation MW.", "Source_IDs": "S17", "Source_date": "2025-07-11", "Source_type": "Company announcement", "Source_locator": "Opening paragraphs", "Source_URL": "https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Within confirmed 24-month window", "Context": "Eni and Khazna agreed heads of terms for a joint venture to develop a 500 MW IT-capacity campus at Ferrera Erbognone, near Milan.", "Implementation": "The July 2025 announcement combines Khazna’s data-center expertise with Eni’s energy capabilities. A proposed gas-fired Blue Power plant, designed for CO₂ capture, would support the campus.", "Assessment": "IT capacity measures computing load, which differs from generation capacity. Heads of terms set development principles. Budget, equity shares, commissioning date and FID are not established by this source.", "Detail_words": 78, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Eni", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-028", "Initiative_ID": "AI-028", "Title": "Ferrera Erbognone campus", "Date": "2025-07-11", "Finding": "Heads of terms for a joint venture developing an AI data-center campus.", "Metric": "500 MW IT capacity", "Source_IDs": "S17", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2025 announcement combines Khazna’s data-center expertise with Eni’s energy capabilities. A proposed gas-fired Blue Power plant, designed for CO₂ capture, would support the campus.", "Context": "Eni and Khazna agreed heads of terms for a joint venture to develop a 500 MW IT-capacity campus at Ferrera Erbognone, near Milan.", "Assessment": "IT capacity measures computing load, which differs from generation capacity. Heads of terms set development principles. Budget, equity shares, commissioning date and FID are not established by this source."}], "Latest_disclosure": "2025-07-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "AI-demand infrastructure", "Audit_flags": [], "Audit_review": {"id": "AI-028", "company": "Eni / Khazna", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "infrastructure", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-029", "Company": "TotalEnergies / Mistral AI", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Joint AI innovation lab for researcher and industrial decision support", "Extracted_finding": "The announced lab will design and test AI workflows for researchers, industrial asset decision support, customers and renewable production; production performance and financial returns are not yet disclosed.", "Stage_at_disclosure": "Joint AI innovation lab announced 2025-06-12; named uses planned for design/testing", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "", "Source_IDs": "S18; R022", "Source_date": "2025-06-12", "Source_type": "company announcement", "Source_locator": "A shared AI innovation lab", "Source_URL": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-06-12; see chronology for operating dates.", "Context": "Researchers, operations teams and customers need AI support tailored to energy and geoscience knowledge.", "Implementation": "Initial use cases: assistant for 1,000 researchers; industrial decision-support for asset performance/CO2; customer energy-saving support.", "Assessment": "Analyst inference: Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.", "Detail_words": 66, "Primary_department": "Research & innovation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "TotalEnergies", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-022"], "Related_IDs": ["AI-030"], "Disclosures": [{"ID": "REF-022", "Initiative_ID": "AI-029", "Title": "Mistral joint lab: assistant for 1,000 researchers", "Date": "2025-06", "Finding": "TotalEnergies and Mistral AI set up a joint innovation lab. First use cases include an AI assistant for TotalEnergies' 1,000 researchers, decision support for industrial asset performance and CO2 reduction, customer energy-saving tools, and specialized assistants for support functions.", "Metric": "1,000 researchers; 300-person Digital Factory; 100+ digital solutions, 60 using ML/GenAI", "Source_IDs": "R022", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Start with a bounded expert population (R&D) for a domain assistant, then extend to support functions.", "Partners": "Mistral AI", "Original_date": "2025-06", "Related_IDs": ["AI-030"]}, {"ID": "ORIG-AI-029", "Initiative_ID": "AI-029", "Title": "Research assistant and innovation lab", "Date": "2025-06-12", "Finding": "Joint lab to build researcher assistance and industrial decision-support tools.", "Metric": "Assistant intended for 1,000 researchers", "Source_IDs": "S18", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Planned work includes an assistant for 1,000 researchers, industrial decision support, emissions and energy-efficiency applications, and customer energy advice.", "Context": "TotalEnergies and Mistral AI announced a joint innovation laboratory in June 2025 to build AI tools with company-specific expertise.", "Assessment": "The researcher count describes intended reach. The source gives no achieved adoption rate or assistant-level economics. The lab’s planned tools should be distinguished from TotalEnergies’ existing AI estate."}], "Latest_disclosure": "2025-06-12", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-005"], "Research_search": " {\"key\": \"total-mistral-lab\", \"existing_ids\": [\"AI-029\"], \"company\": \"TotalEnergies\", \"title\": \"Joint AI innovation lab for researcher and industrial decision support\", \"department\": \"Research & innovation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Joint AI innovation lab announced 2025-06-12; named uses planned for design/testing\", \"practical_relevance\": \"The announced lab will design and test AI workflows for researchers, industrial asset decision support, customers and renewable production; production performance and financial returns are not yet disclosed.\", \"business_problem\": \"Researchers, operations teams and customers need AI support tailored to energy and geoscience knowledge.\", \"workflow\": \"Initial use cases: assistant for 1,000 researchers; industrial decision-support for asset performance/CO2; customer energy-saving support.\", \"technology_data\": \"Mistral generative AI; detailed architecture, model validation and data governance not disclosed.\", \"integration_human_role\": \"AI outputs are intended to support employees and customers; expert review details are not given.\", \"deployment_scope\": \"Innovation lab; enterprise scope and proof-of-concept status unspecified.\", \"partners\": [\"Mistral AI\"], \"disclosed_results\": [], \"economics\": \"Not disclosed for the Mistral lab; the 2026 €100m frontier-model investment is a distinct AI-077 initiative and is not attributed to this lab.\", \"timeline\": [{\"date\": \"2025-06-12\", \"event\": \"Joint innovation lab announced with researcher, industrial and customer use cases.\", \"source_ids\": [\"S18\"]}], \"limitations\": [], \"assessment\": \"Analyst inference: Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.\", \"sources\": [{\"id\": \"S18\", \"title\": \"TotalEnergies to Collaborate with Mistral AI to Increase the Application of Artificial Intelligence\", \"url\": \"https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng\", \"publisher\": \"TotalEnergies\", \"date\": \"2025-06-12\", \"type\": \"company announcement\", \"locator\": \"A shared AI innovation lab\", \"review_status\": \"full text reviewed\", \"supports\": [\"Joint lab use cases include an assistant for 1,000 researchers, decision-support for industrial performance and CO2 emissions, and customer energy-saving support.\", \"TotalEnergies reported Digital Factory had developed over 100 solutions, 60 using ML through generative AI; these are aggregate portfolio counts, not outcome metrics.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-007\"}], \"field_sources\": {\"business_problem\": [\"S18\"], \"workflow\": [\"S18\"], \"technology_data\": [\"S18\"], \"integration_human_role\": [\"S18\"], \"deployment_scope\": [\"S18\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Separate 2025 lab use cases from 2026 frontier-model investment; do not collapse into one deployment.\", \"related_existing_ids\": [\"AI-030\"], \"cohort\": \"majors\", \"packet_id\": \"majors-005\", \"source_ids\": [\"S18\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-029\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-029", "company": "TotalEnergies / Mistral AI", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-030", "Company": "TotalEnergies", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Digital Factory portfolio of AI and digital solutions", "Extracted_finding": "Useful company-level production of use cases; reported solution count is not an operational results metric.", "Stage_at_disclosure": "2025-06-12 disclosure, with five-year factory anniversary", "Reported_metric": "Digital Factory team: 300 developers, data scientists and digital experts (Company-reported); Factory solutions: 100+ total; 60 use ML-to-GenAI methods (Company-reported)", "Interpretation_limit": "Portfolio counts do not establish active use, adoption or business value per solution. Specific Mistral joint lab is separate AI-029 and launch describes tests/plans, not production.", "Source_IDs": "RS115; S18", "Source_date": "2025-06-12", "Source_type": "company announcement", "Source_locator": "Digital Factory baseline and joint AI lab", "Source_URL": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-06-12; see chronology for operating dates.", "Context": "Business units need centrally developed solutions for earth sciences, machinery, industrial asset performance, renewable production and customer energy use.", "Implementation": "Digital Factory’s 300 developers/data/digital experts built over 100 solutions, 60 employing technologies from machine learning to generative AI. Mistral collaboration spawned separate planned joint lab use cases; those should remain AI-029, not be collapsed into this portfolio profile.", "Assessment": "Useful company-level production of use cases; reported solution count is not an operational results metric.", "Detail_words": 63, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "TotalEnergies", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-030", "Initiative_ID": "AI-030", "Title": "Digital Factory AI portfolio", "Date": "2025-06-12", "Finding": "Company reports AI in earth sciences, equipment diagnostics and predictive maintenance.", "Metric": "60 of >100 digital solutions use ML through generative AI", "Source_IDs": "S18", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The applications span machine learning through generative AI, including earth sciences, equipment diagnostics and predictive maintenance.", "Context": "TotalEnergies describes an existing Digital Factory with 300 experts and more than 100 digital solutions, of which 60 use AI.", "Assessment": "The portfolio predates the Mistral lab. Counting all 60 as Mistral deployments would overstate the partnership’s scope, and the disclosure does not isolate the portfolio’s financial return."}], "Latest_disclosure": "2025-06-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-013"], "Research_search": " {\"key\": \"total-digital-factory\", \"existing_ids\": [\"AI-030\"], \"related_existing_ids\": [\"AI-029\"], \"company\": \"TotalEnergies\", \"title\": \"Digital Factory portfolio of AI and digital solutions\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"2025-06-12 disclosure, with five-year factory anniversary\", \"practical_relevance\": \"Useful company-level production of use cases; reported solution count is not an operational results metric.\", \"business_problem\": \"Business units need centrally developed solutions for earth sciences, machinery, industrial asset performance, renewable production and customer energy use.\", \"workflow\": \"Digital Factory’s 300 developers/data/digital experts built over 100 solutions, 60 employing technologies from machine learning to generative AI. Mistral collaboration spawned separate planned joint lab use cases; those should remain AI-029, not be collapsed into this portfolio profile.\", \"technology_data\": \"Solution-level architectures are heterogeneous; source mentions ML-to-GenAI range but no platform, models or data integration details.\", \"integration_human_role\": \"Central digital teams develop solutions for business users; customer/industrial operating decisions remain unspecified.\", \"deployment_scope\": \"Company-wide factory; counts from June 2025 company announcement.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Digital Factory team\", \"value\": \"300 developers, data scientists and digital experts\", \"period\": \"2025-06-12\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS115\"]}, {\"metric\": \"Factory solutions\", \"value\": \"100+ total; 60 use ML-to-GenAI methods\", \"period\": \"2025-06-12\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS115\"]}], \"economics\": \"No AI-only economic value quantified.\", \"timeline\": [{\"date\": \"2025-06-12\", \"event\": \"Digital Factory portfolio of AI and digital solutions documented or disclosed.\", \"source_ids\": [\"RS115\"]}], \"limitations\": [\"Portfolio counts do not establish active use, adoption or business value per solution.\", \"Specific Mistral joint lab is separate AI-029 and launch describes tests/plans, not production.\"], \"assessment\": \"Useful company-level production of use cases; reported solution count is not an operational results metric.\", \"sources\": [{\"id\": \"RS115\", \"title\": \"TotalEnergies to collaborate with Mistral AI\", \"url\": \"https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf\", \"publisher\": \"TotalEnergies\", \"date\": \"2025-06-12\", \"type\": \"company announcement\", \"locator\": \"Digital Factory baseline and joint AI lab\", \"review_status\": \"full text reviewed\", \"supports\": [\"TotalEnergies says its Digital Factory had 300 digital experts and developed >100 solutions, 60 using technologies from ML to GenAI.\", \"The announcement creates a distinct Mistral joint innovation lab for researcher assistant, industrial asset decision support, customer and renewable use cases; it says those use cases will be designed/tested, not deployed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-016\"}], \"field_sources\": {\"business_problem\": [\"RS115\"], \"workflow\": [\"RS115\"], \"technology_data\": [\"RS115\"], \"integration_human_role\": [\"RS115\"], \"deployment_scope\": [\"RS115\"], \"economics\": [\"RS115\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-013\", \"source_ids\": [\"RS115\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-030\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-030", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-031", "Company": "Repsol / Accenture", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Multi-agent workflow platform for employee productivity", "Extracted_finding": "Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.", "Stage_at_disclosure": "Pilot. Repsol’s 2026 report describes a multi-agent employee-workflow platform piloted from 2025 for over four months and updated to 34 agents across three workflows. This remains explicitly pilot-stage despite the breadth of agents/workflows.", "Reported_metric": "Agents in multi-agent workflows: 34 across 3 workflows (Company-reported)", "Interpretation_limit": "Agent counts vary by disclosure date; not a contradiction. Annual report window within research range.", "Source_IDs": "S19; RS007; R029", "Source_date": "2026-02-19", "Source_type": "company announcement", "Source_locator": "Pilot project / GenAI competence center", "Source_URL": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2026-02-19; see chronology for operating dates.", "Context": "Employees need to coordinate multi-step information work and apply internal knowledge across business processes.", "Implementation": "Platform configures agents into workflows; pilot teams supervised agents on departmental processes; separate upstream Automated Well Data Ingestion automates daily drilling/geology reports.", "Assessment": "Analyst inference: Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.", "Detail_words": 72, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Organizational operating model", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Repsol", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-028", "REF-029"], "Related_IDs": [], "Disclosures": [{"ID": "REF-029", "Initiative_ID": "AI-031", "Title": "Accenture and NVIDIA to scale AI agents", "Date": "2025-02", "Finding": "Repsol expanded its Digital Program by deploying AI agent systems with Accenture's AI Refinery and the NVIDIA AI platform, targeting process optimization, planning, forecasting, maintenance and customer personalization. It also plans to explore NVIDIA Omniverse digital twins.", "Metric": "650+ digital initiatives; 1,200+ people on digital projects; 50 tech partners", "Source_IDs": "R029", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "SI-led agent factories are a common way integrateds get to scale fast; independent tools need to plug into SI delivery models.", "Partners": "Accenture AI Refinery, NVIDIA AI Enterprise, NVIDIA Omniverse", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-028", "Initiative_ID": "AI-031", "Title": "Multi-agent platform: 22 agents supervised by 50 staff", "Date": "2025-10", "Finding": "Repsol launched a multi-agent platform built with Accenture in which AI agents self-organize and coordinate tasks within workflows. The pilot ran three teams with 22 agents in the 700-person Digitalization and Services department, supervised by 50+ employees for over four months.", "Metric": "22 agents; 3 agent teams; 50+ supervisors; 4+ months", "Source_IDs": "S19", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Pilot agents inside your own IT/digital function first, with explicit human supervisors, before pushing to operations.", "Partners": "Accenture", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "ORIG-AI-031", "Initiative_ID": "AI-031", "Title": "Multi-agent platform", "Date": "2025-10-15", "Finding": "Piloting a platform that enables teams to construct coordinated AI agents.", "Metric": "22 agents across three pilot teams; >50 people supervising", "Source_IDs": "S19", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "In Digitalization and Services, three teams used 22 agents for more than four months. More than 50 people subsequently worked together supervising agent teams.", "Context": "Repsol and Accenture developed a pilot platform for coordinated AI agents, supported by dedicated knowledge bases and a professional collaboration interface.", "Assessment": "The pilot demonstrates a defined organizational setting. Its 22 agents are distinct from the 22 scaled GenAI use cases in AI-032. No platform-specific financial return is disclosed."}], "Latest_disclosure": "2026-02-19", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-007"], "Research_search": " {\"key\": \"repsol-multiagent\", \"existing_ids\": [\"AI-031\"], \"company\": \"Repsol\", \"title\": \"Multi-agent workflow platform for employee productivity\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Pilot deployed from 2025 for over four months; 2026 annual report snapshot updated to 34 agents / 3 workflows\", \"practical_relevance\": \"Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.\", \"business_problem\": \"Employees need to coordinate multi-step information work and apply internal knowledge across business processes.\", \"workflow\": \"Platform configures agents into workflows; pilot teams supervised agents on departmental processes; separate upstream Automated Well Data Ingestion automates daily drilling/geology reports.\", \"technology_data\": \"Repsol-built agent platform with Accenture; custom knowledge bases, professional interface and language models. Annual report lists 34 agents/3 workflows, while Oct 2025 snapshot reported 22 agents/3 teams.\", \"integration_human_role\": \"Over 50 people supervised pilot teams; employee review role is explicit.\", \"deployment_scope\": \"Pilot in Digitalization and Services department (>700 employees) plus reported downstream workflows and an upstream automated-report workflow.\", \"partners\": [\"Accenture\"], \"disclosed_results\": [{\"metric\": \"Agents in multi-agent workflows\", \"value\": \"34 across 3 workflows\", \"period\": \"2025 annual report\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS007\"]}], \"economics\": \"No quantified efficiency metric disclosed; Repsol claims numerous efficiencies and improvements.\", \"timeline\": [{\"date\": \"2025-10-15\", \"event\": \"Multi-agent pilot announced; 22 agents across three teams.\", \"source_ids\": [\"S19\"]}, {\"date\": \"2026-02\", \"event\": \"Annual report reports 34 agents in three workflows and Automated Well Data Ingestion.\", \"source_ids\": [\"RS007\"]}], \"limitations\": [\"Agent counts vary by disclosure date; not a contradiction. Annual report window within research range.\"], \"assessment\": \"Analyst inference: Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.\", \"sources\": [{\"id\": \"S19\", \"title\": \"Repsol launches intelligent multi-agent platform\", \"url\": \"https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml\", \"publisher\": \"Repsol\", \"date\": \"2025-10-15\", \"type\": \"company announcement\", \"locator\": \"Pilot project / GenAI competence center\", \"review_status\": \"full text reviewed\", \"supports\": [\"Accenture collaboration produced an agent-construction platform and pilot in Digitalization and Services department.\", \"Three teams with 22 agents supported processes for over four months; over 50 people supervised the agent teams.\", \"Repsol reports 60 production GenAI use cases and 22 scaled as of Oct 2025; these are company-reported portfolio measures.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-010\"}, {\"id\": \"RS007\", \"title\": \"Consolidated Management Report 2025 — Repsol\", \"url\": \"https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf\", \"publisher\": \"Repsol\", \"date\": \"2026-02-19\", \"type\": \"annual report\", \"locator\": \"5.5 Technology and digitalization\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Repsol reports Automated Well Data Ingestion automates daily drilling and geology reports and integrates data into corporate systems to speed exploration decisions.\", \"Annual report later lists 34 agents in three workflows, distinct from the Oct 2025 pilot snapshot of 22 agents.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-011\"}], \"field_sources\": {\"business_problem\": [\"S19\"], \"workflow\": [\"S19\", \"RS007\"], \"technology_data\": [\"S19\", \"RS007\"], \"integration_human_role\": [\"S19\"], \"deployment_scope\": [\"S19\", \"RS007\"], \"economics\": [\"S19\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [\"AI-032\"], \"cohort\": \"majors\", \"packet_id\": \"majors-007\", \"source_ids\": [\"S19\", \"RS007\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-031\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one supporting long document was reviewed at relevant excerpts only.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-031", "company": "Repsol / Accenture", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Normalized_stage": "Pilot", "Stage_review": {"id": "AI-031", "stage": "Pilot", "stage_basis": "Repsol’s 2026 report describes a multi-agent employee-workflow platform piloted from 2025 for over four months and updated to 34 agents across three workflows. This remains explicitly pilot-stage despite the breadth of agents/workflows.", "source_url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "source_date": "2026-02-19 (update; pilot began during 2025)", "evidence_origin": "Operator", "limits": "Agent count is a dated snapshot; no mature production availability or outcome attribution beyond reported pilot results.", "source_id": "S19"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-032", "Company": "Repsol", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Generative AI Competence Center and employee scaling", "Extracted_finding": "Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.", "Stage_at_disclosure": "In place >2 years by October 2025; report updated with 2025 counts", "Reported_metric": "GenAI use cases in production: 60+; 22 scaled (Oct 2025 release) (Company-reported); Staff trained: 5,000+ (Company-reported)", "Interpretation_limit": "Case counts vary by source/date and counting definition; do not conflate competence-center cases with 22-agent pilot or later 34-agent scale. No independent outcomes or value quantification.", "Source_IDs": "S19; RS007; R027", "Source_date": "2026-02", "Source_type": "company announcement", "Source_locator": "Competence Center, platform pilot and user adoption", "Source_URL": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2026-02; see chronology for operating dates.", "Context": "Repsol needs a repeatable way to identify and govern genAI cases, train staff and move selected solutions into production across a large multi-energy enterprise.", "Implementation": "Competence Center develops use cases and employee skills; by 2025 company reported 60+ use cases in production, 22 scaled, and 5,000+ trained. A later consolidated report says >75 initiatives, while the October press release reports platform pilot with 22 agents/3 teams; those different counts should not be merged.", "Assessment": "Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.", "Detail_words": 66, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Repsol", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-027"], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-032", "Initiative_ID": "AI-032", "Title": "GenAI competence center", "Date": "2025-10-15", "Finding": "Scaling use cases and employee capability across the company.", "Metric": ">60 production use cases; 22 scaled; >5,000 employees trained", "Source_IDs": "S19", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The October 2025 release reports more than 60 use cases in production, 22 scaled and more than 5,000 employees trained in GenAI or advanced prompting.", "Context": "Repsol’s Generative AI Competence Center supports production use cases and workforce capability across the company.", "Assessment": "These are different measures of maturity. Training does not establish active use, while the release does not enumerate all production cases or provide comparable application-level economics."}, {"ID": "REF-027", "Initiative_ID": "AI-032", "Title": "Copilot to 25,000 staff; pilot saved 121 min/week", "Date": "2026-02", "Finding": "Repsol extended Copilot to about 25,000 employees after a four-month M365 Copilot pilot with 550 users showed 121 minutes saved per person per week and 16% higher work quality. It runs 60+ GenAI use cases and 34 agents, and has trained 5,000+ employees plus all senior management in prompting.", "Metric": "~25,000 employees; 121 min/week saved; +16% quality; 60+ use cases; 34 agents; 5,000+ trained; 200k+ lines via GitHub Copilot", "Source_IDs": "R027", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Run a measured pilot (time saved, quality, would-not-give-up rate) before mass licensing; it gives the board-ready business case.", "Partners": "Microsoft 365 Copilot, GitHub Copilot", "Original_date": "2026-02", "Related_IDs": []}], "Latest_disclosure": "2026-02", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-004"], "Research_search": " {\"key\": \"repsol-genai-center\", \"existing_ids\": [\"AI-032\"], \"related_existing_ids\": [\"AI-031\"], \"company\": \"Repsol\", \"title\": \"Generative AI Competence Center and employee scaling\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"In place >2 years by October 2025; report updated with 2025 counts\", \"practical_relevance\": \"Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.\", \"business_problem\": \"Repsol needs a repeatable way to identify and govern genAI cases, train staff and move selected solutions into production across a large multi-energy enterprise.\", \"workflow\": \"Competence Center develops use cases and employee skills; by 2025 company reported 60+ use cases in production, 22 scaled, and 5,000+ trained. A later consolidated report says >75 initiatives, while the October press release reports platform pilot with 22 agents/3 teams; those different counts should not be merged.\", \"technology_data\": \"Competence-center operating model and responsible AI; multi-agent platform itself is a distinct AI-031 profile. Report identifies automated daily well-data ingestion as a separate workflow detail related to the center.\", \"integration_human_role\": \"Human-in-the-loop design and supervision emphasized; 50+ people reportedly supervised agent teams in pilot; employees trained on GenAI/prompting.\", \"deployment_scope\": \"Enterprise-wide program; case counts and training numbers are reported by Repsol; pilot specific to Digitalization and Services team of 700+.\", \"partners\": [\"Accenture\", \"NVIDIA\"], \"disclosed_results\": [{\"metric\": \"GenAI use cases in production\", \"value\": \"60+; 22 scaled (Oct 2025 release)\", \"period\": \"2025-10-15\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S19\"]}, {\"metric\": \"Staff trained\", \"value\": \"5,000+\", \"period\": \"2025-10-15 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S19\"]}], \"economics\": \"No dollar economics; efficiency described qualitatively.\", \"timeline\": [{\"date\": \"2025-10-15\", \"event\": \"Generative AI Competence Center and employee scaling documented or disclosed.\", \"source_ids\": [\"S19\", \"RS007\"]}], \"limitations\": [\"Case counts vary by source/date and counting definition; do not conflate competence-center cases with 22-agent pilot or later 34-agent scale.\", \"No independent outcomes or value quantification.\"], \"assessment\": \"Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.\", \"sources\": [{\"id\": \"S19\", \"title\": \"Repsol launches intelligent multi-agent platform\", \"url\": \"https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml\", \"publisher\": \"Repsol\", \"date\": \"2025-10-15\", \"type\": \"company announcement\", \"locator\": \"Competence Center, platform pilot and user adoption\", \"review_status\": \"full text reviewed\", \"supports\": [\"Repsol says its GenAI Competence Center had produced 60+ use cases in production and 22 scaled; 5,000+ employees trained.\", \"Platform was a pilot in Digitalization and Services with 3 teams and 22 agents supporting processes for more than four months; 50+ people supervised agents.\", \"A separate 2025 consolidated management report later says 34 agents across three workflows; dates/scope differ, so counts should not be blended.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-005\"}, {\"id\": \"RS007\", \"title\": \"2025 Consolidated Management Report\", \"url\": \"https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf\", \"publisher\": \"Repsol\", \"date\": \"2026-02\", \"type\": \"annual report\", \"locator\": \"Section 5.5 Technology and digitalization\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"2025 report says Competence Center promoted >75 initiatives after two years, and the multi-agent platform had 34 agents in three workflows.\", \"Report names automated well-data ingestion for daily drilling/geology reports, integration into corporate systems and faster exploration decision-making.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-006\"}], \"field_sources\": {\"business_problem\": [\"S19\"], \"workflow\": [\"S19\", \"RS007\"], \"technology_data\": [\"S19\", \"RS007\"], \"integration_human_role\": [\"S19\"], \"deployment_scope\": [\"S19\", \"RS007\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-004\", \"source_ids\": [\"S19\", \"RS007\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-032\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one supporting long document was reviewed at relevant excerpts only.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-032", "company": "Repsol", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Repsol’s page says Copilot for Web reached ~25,000 employees while M365 Copilot licenses are progressively scaling; a 550-person, four-month pilot reported 121 minutes/week per participant. The page describes >60 use cases in process, not 60+ in production.", "correction": "Keep product reach distinct from M365 licensing and pilot results. Do not attribute the 121 minutes to reduced labor cost or use this page to support a “60+ in production/22 scaled” count.", "source_url": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-033", "Company": "Equinor", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "AI-enabled maintenance, field planning and seismic portfolio", "Extracted_finding": "Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.", "Stage_at_disclosure": "Scaled. Equinor’s 2026-01-07 operator update reports multiple implemented AI applications across maintenance, field planning and seismic, with company-reported outcomes covering 2025 and cumulative history since 2020.", "Reported_metric": "AI-attributed value: USD 130 million (Company-reported); Predictive maintenance value: USD 120 million (Company-reported); Johan Sverdrup phase 3 planning saving: USD 12 million (Company-reported partner saving); Seismic interpretation area: 2 million km² (Company-reported)", "Interpretation_limit": "Parent total 2025 value and subcase numbers use different periods; don’t sum or treat as additive. No independent value audit or model architecture.", "Source_IDs": "S20", "Source_date": "2026-01-07", "Source_type": "company announcement", "Source_locator": "Implemented applications and value disclosure", "Source_URL": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2026-01-07; see chronology for operating dates.", "Context": "Equipment failures, planning search spaces and seismic interpretation volume exceed what teams can efficiently examine manually.", "Implementation": "Predictive monitoring covers rotating equipment; AI generates and ranks well/development alternatives for expert review; seismic tooling accelerates interpretation. These are three distinct implementations reported together, not a single model.", "Assessment": "Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.", "Detail_words": 66, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Equinor", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-025"], "Related_IDs": ["AI-034", "AI-035", "AI-036"], "Disclosures": [{"ID": "REF-025", "Initiative_ID": "AI-033", "Title": "AI saved Equinor $130M in 2025", "Date": "2026-01", "Finding": "Equinor said AI created $130M in value and savings in 2025 and more than $330M since 2020. Predictive maintenance on 700+ rotating machines with 24,000 sensors accounts for $120M since 2020; AI interpreted 2 million km2 of seismic in 2025 at 10x capacity; and a well-planning tool saved $12M on Johan Sverdrup Phase 3.", "Metric": "$130M in 2025; $330M+ since 2020; 700+ machines, 24,000 sensors; 10x seismic interpretation; $12M Sverdrup saving", "Source_IDs": "S20", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Publish value by use case; a few industrial ML workhorses (rotating equipment, seismic) still deliver most of the dollars while GenAI tools spread to employees.", "Partners": "", "Original_date": "2026-01", "Related_IDs": ["AI-034", "AI-035", "AI-036"]}, {"ID": "ORIG-AI-033", "Initiative_ID": "AI-033", "Title": "Industrial AI portfolio value", "Date": "2026-01-07", "Finding": "Reports value creation and savings for Equinor and partners.", "Metric": "$130m in 2025; >$330m since 2020", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Management says traditional machine learning on operating data provides much of the industrial value. AI-034 to AI-036 describe contributing maintenance, planning and seismic examples.", "Context": "Equinor reports $130 million of AI-related value and savings for itself and partners in 2025, and more than $330 million since 2020.", "Assessment": "These are company-defined benefits, including partners. The component examples overlap the portfolio total, so adding them again would double-count value."}], "Latest_disclosure": "2026-01-07", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-005"], "Research_search": " {\"key\": \"equinor-ai-portfolio\", \"existing_ids\": [\"AI-033\"], \"related_existing_ids\": [\"AI-034\", \"AI-035\", \"AI-036\"], \"company\": \"Equinor\", \"title\": \"AI-enabled maintenance, field planning and seismic portfolio\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Implemented cases disclosed 2026-01-07; results refer to calendar 2025 and cumulative since 2020\", \"practical_relevance\": \"Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.\", \"business_problem\": \"Equipment failures, planning search spaces and seismic interpretation volume exceed what teams can efficiently examine manually.\", \"workflow\": \"Predictive monitoring covers rotating equipment; AI generates and ranks well/development alternatives for expert review; seismic tooling accelerates interpretation. These are three distinct implementations reported together, not a single model.\", \"technology_data\": \"Company says traditional ML on operational data; predictive system spans >700 rotating machines/24,000 sensors; seismic tool processed 2m km² in 2025. Other algorithms/integration not disclosed.\", \"integration_human_role\": \"Experts focus on ranked planning options; maintenance tool flags failure/maintenance needs; geoscientists use expanded seismic interpretation. Human control and specific automation boundaries not disclosed.\", \"deployment_scope\": \"Equinor offshore platforms and land facilities; rotating asset coverage across all facilities, planning includes Johan Sverdrup phase 3, seismic use 2m km² in 2025.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"AI-attributed value\", \"value\": \"USD 130 million\", \"period\": \"Calendar 2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S20\"]}, {\"metric\": \"Predictive maintenance value\", \"value\": \"USD 120 million\", \"period\": \"Cumulative since 2020\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S20\"]}, {\"metric\": \"Johan Sverdrup phase 3 planning saving\", \"value\": \"USD 12 million\", \"period\": \"Case disclosed 2026-01-07\", \"attribution\": \"Company-reported partner saving\", \"source_ids\": [\"S20\"]}, {\"metric\": \"Seismic interpretation area\", \"value\": \"2 million km²\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S20\"]}], \"economics\": \"Equinor reported $130m value in 2025 and >$330m cumulative since 2020. $120m maintenance is cumulative since 2020; $12m for Johan Sverdrup is partner saving. Prevent double-counting these within 2025 total.\", \"timeline\": [{\"date\": \"2026-01-07\", \"event\": \"AI-enabled maintenance, field planning and seismic portfolio documented or disclosed.\", \"source_ids\": [\"S20\"]}], \"limitations\": [\"Parent total 2025 value and subcase numbers use different periods; don’t sum or treat as additive.\", \"No independent value audit or model architecture.\"], \"assessment\": \"Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.\", \"sources\": [{\"id\": \"S20\", \"title\": \"Equinor: AI delivered USD 130m value in 2025\", \"url\": \"https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million\", \"publisher\": \"Equinor\", \"date\": \"2026-01-07\", \"type\": \"company announcement\", \"locator\": \"Implemented applications and value disclosure\", \"review_status\": \"full text reviewed\", \"supports\": [\"Company says AI generated USD 130m value for Equinor and partners in calendar 2025, with >USD 330m cumulative since 2020.\", \"Three applications: 700+ rotating machines/24,000 sensors predictive maintenance (USD 120m since 2020); well and field planning, including USD12m Johan Sverdrup phase-3 saving; seismic interpretation of 2m km2 in 2025 with tenfold capacity.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-007\"}], \"field_sources\": {\"business_problem\": [\"S20\"], \"workflow\": [\"S20\"], \"technology_data\": [\"S20\"], \"integration_human_role\": [\"S20\"], \"deployment_scope\": [\"S20\"], \"economics\": [\"S20\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-005\", \"source_ids\": [\"S20\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-033\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-033", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-033", "stage": "Scaled", "stage_basis": "Equinor’s 2026-01-07 operator update reports multiple implemented AI applications across maintenance, field planning and seismic, with company-reported outcomes covering 2025 and cumulative history since 2020.", "source_url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "source_date": "2026-01-07 (outcomes: 2025 and cumulative since 2020)", "evidence_origin": "Operator", "limits": "Parent total and subcase values have differing periods and must not be summed. Company values are not independently audited.", "source_id": "S20"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Portfolio roll-up. The USD 120m predictive-maintenance and USD 12m Johan Sverdrup figures are also profiled as AI-034 and AI-035 and use different periods; they are not additive to the USD 130m. Stats count AI-033's realized outcome once, through its components, and count only the USD 130m in dollar claims.", "Outcome_counted": false, "Outcome_count_note": "Portfolio aggregate overlaps other outcome profiles; excluded from headline and scorecard result counts.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-034", "Company": "Equinor", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Predictive maintenance", "Extracted_finding": "Sensor models identify equipment failure and maintenance needs.", "Stage_at_disclosure": "Deployed", "Reported_metric": ">700 machines; $120m value since 2020", "Interpretation_limit": "Cumulative; overlaps overall value.", "Source_IDs": "S20", "Source_date": "2026-01-07", "Source_type": "Company announcement", "Source_locator": "Three implemented contributors and Hege Skryseth comments", "Source_URL": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Within confirmed 24-month window", "Context": "Predictive maintenance monitors more than 700 rotating machines using 24,000 sensors across Equinor facilities.", "Implementation": "Models predict failures and maintenance needs. Equinor attributes $120 million of cumulative value to the application since 2020.", "Assessment": "Earlier detection may avoid shutdowns, but this figure is cumulative and belongs within the broader AI value total. It is not annual recurring savings or solely Equinor net income.", "Detail_words": 61, "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Production & field operations", "Refining, LNG & process operations"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Equinor", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-034", "Initiative_ID": "AI-034", "Title": "Predictive maintenance", "Date": "2026-01-07", "Finding": "Sensor models identify equipment failure and maintenance needs.", "Metric": ">700 machines; $120m value since 2020", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Models predict failures and maintenance needs. Equinor attributes $120 million of cumulative value to the application since 2020.", "Context": "Predictive maintenance monitors more than 700 rotating machines using 24,000 sensors across Equinor facilities.", "Assessment": "Earlier detection may avoid shutdowns, but this figure is cumulative and belongs within the broader AI value total. It is not annual recurring savings or solely Equinor net income."}], "Latest_disclosure": "2026-01-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-034", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "700+ machines", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "USD 120m is cumulative since 2020 and overlaps the AI-033 portfolio figure; not additive.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-035", "Company": "Equinor", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Well and field planning", "Extracted_finding": "AI generated an alternative Johan Sverdrup phase 3 development solution.", "Stage_at_disclosure": "Applied", "Reported_metric": "$12m partnership savings", "Interpretation_limit": "Included case study, not incremental to portfolio total.", "Source_IDs": "S20", "Source_date": "2026-01-07", "Source_type": "Company announcement", "Source_locator": "Three implemented contributors and Hege Skryseth comments", "Source_URL": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Within confirmed 24-month window", "Context": "Equinor uses AI to generate alternative well and field-development plans for experts to assess.", "Implementation": "For Johan Sverdrup phase 3, management says an AI-generated alternative saved the partnership $12 million.", "Assessment": "The example shows value from a changed development choice. It does not establish that all planning decisions are automated, or that the saving repeats annually across the portfolio.", "Detail_words": 57, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["Drilling & completions", "Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Equinor", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-035", "Initiative_ID": "AI-035", "Title": "Well and field planning", "Date": "2026-01-07", "Finding": "AI generated an alternative Johan Sverdrup phase 3 development solution.", "Metric": "$12m partnership savings", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "For Johan Sverdrup phase 3, management says an AI-generated alternative saved the partnership $12 million.", "Context": "Equinor uses AI to generate alternative well and field-development plans for experts to assess.", "Assessment": "The example shows value from a changed development choice. It does not establish that all planning decisions are automated, or that the saving repeats annually across the portfolio."}], "Latest_disclosure": "2026-01-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-035", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "USD 12m is a case within the AI-033 portfolio figure; not additive.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-036", "Company": "Equinor", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "Seismic interpretation", "Extracted_finding": "AI expands seismic interpretation capacity.", "Stage_at_disclosure": "Deployed", "Reported_metric": "10x capacity; 2m km² interpreted in 2025", "Interpretation_limit": "Capacity claim, not discovery success rate.", "Source_IDs": "S20", "Source_date": "2026-01-07", "Source_type": "Company announcement", "Source_locator": "Three implemented contributors and Hege Skryseth comments", "Source_URL": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "Evidence_status": "Primary source text reviewed", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Within confirmed 24-month window", "Context": "Equinor’s AI interpretation tools increase the volume of seismic information that geoscientists can evaluate.", "Implementation": "The company reports tenfold interpretation capacity and two million square kilometres interpreted in 2025.", "Assessment": "Interpretation throughput is an intermediate operating measure. Discovery success and commercial reserves depend on additional geological and economic factors that the cited metric does not capture.", "Detail_words": 54, "Primary_department": "Exploration & geoscience", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Equinor", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-036", "Initiative_ID": "AI-036", "Title": "Seismic interpretation", "Date": "2026-01-07", "Finding": "AI expands seismic interpretation capacity.", "Metric": "10x capacity; 2m km² interpreted in 2025", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The company reports tenfold interpretation capacity and two million square kilometres interpreted in 2025.", "Context": "Equinor’s AI interpretation tools increase the volume of seismic information that geoscientists can evaluate.", "Assessment": "Interpretation throughput is an intermediate operating measure. Discovery success and commercial reserves depend on additional geological and economic factors that the cited metric does not capture."}], "Latest_disclosure": "2026-01-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-036", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-037", "Company": "Aramco", "Segment": "NOC / integrated", "Theme": "Operational AI", "Initiative": "AI contribution to FY2025 Technology Realized Value", "Extracted_finding": "Aramco reports SAR 19.9bn ($5.3bn) FY2025 TRV, including SAR 9.8bn ($2.6bn) from AI-driven solutions and SAR 10.1bn ($2.7bn) from non-AI technologies. This replaces the imprecise “over half” statement.", "Stage_at_disclosure": "FY2025 aggregate technology value reported for deployed technologies; AI subtotal is portfolio-level.", "Reported_metric": "FY2025: $5.3bn total TRV; $2.6bn AI-driven and $2.7bn non-AI-driven. Separately, official FY2025 results materials allocate portfolio-wide TRV to $3.9bn capex avoidance, $1.0bn opex reduction and $0.4bn revenue enhancement; AI-only allocation is not given.", "Interpretation_limit": "Aramco describes its aggregate TRV metric as third-party verified and incremental to a business baseline. No independent assurance opinion was reviewed here. The $2.6bn AI subtotal is company-reported; AI-only allocation by project or benefit type and full costs are not disclosed. TRV includes capital or operating expenditure avoided and revenue added. Capital avoidance is not incremental cash income, and aggregate TRV is not project-level net cash ROI.", "Source_IDs": "S21; RS109; R133; OCT5-S043; OCT5-S044", "Source_date": "2026-03-10", "Source_type": "Operator annual report", "Source_locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38, Technology Realized Value definition and AI/non-AI split", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Evidence_status": "Relevant original annual-report pages reviewed on 2026-10-06.", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2026-03-10; see chronology for operating dates.", "Context": "Assessing value across a broad technology portfolio.", "Implementation": "The earnings materials aggregate capital avoided, operating costs reduced and revenue enhanced.", "Assessment": "Useful evidence of management measurement discipline. A reader should request the counterfactual baseline, validation methodology and allocation by project before using TRV in an investment model.", "Detail_words": 69, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-000", "REF-133"], "Related_IDs": ["AI-038"], "Disclosures": [{"ID": "REF-133", "Initiative_ID": "AI-037", "Title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "Date": "2025-01", "Finding": "Aramco's CEO said technology delivered $4B of value in 2024, double the 2023 figure, with AI embedded across operations. The company has trained hundreds of AI specialists and more than 6,000 AI-enabled professionals, and it has developed 300+ AI use cases. It is also building the METABRAIN industrial LLM on 90 years of company data. Separately, in late 2025 Aramco agreed to take a significant minority stake in HUMAIN, the PIF-backed national AI company.", "Metric": "$4B tech value 2024 (vs $2B 2023); 6,000+ AI-enabled professionals; 300+ use cases; METABRAIN LLM", "Source_IDs": "R133", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "National programs link NOC workforce training to sovereign AI infrastructure; expect the same pattern in other Gulf NOCs.", "Partners": "Aramco, Aramco Digital, HUMAIN, PIF", "Original_date": "2025-01", "Related_IDs": ["AI-038"]}, {"ID": "REF-000", "Initiative_ID": "AI-037", "Title": "AI drives over half of $5.3B technology value", "Date": "2026-03", "Finding": "Aramco reported $5.3B of Technology Realized Value (TRV) in 2025 from AI, digital and other technology, with CEO Amin Nasser saying over half was enabled by AI. Cumulative TRV since 2023 is $11.3B, and the company guided to an additional $3-5B in 2026.", "Metric": "$5.3B TRV 2025 (>50% AI-enabled); $11.3B cumulative since 2023; $3-5B additional TRV expected 2026", "Source_IDs": "S21", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Publishing a single audited-style 'realized value' KPI for AI/tech turns AI from a cost center into a board-level P&L story; builders should instrument value capture from day one.", "Partners": "", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "ORIG-AI-037", "Initiative_ID": "AI-037", "Title": "AI-enabled technology value", "Date": "2026-03-10", "Finding": "CEO attributes more than half of reported technology value to AI.", "Metric": ">$2.65bn implied AI-enabled share of $5.3bn total", "Source_IDs": "S21", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Multiplying the disclosed total by 50% implies an AI-enabled share above $2.65 billion. That threshold is calculated here, rather than quoted as a separate company figure.", "Context": "Aramco’s CEO said more than half of $5.3 billion in 2025 Technology Realized Value came from AI.", "Assessment": "TRV is a company-defined technology measure. It should not be treated as audited incremental net income, and comparability with another operator’s AI savings requires matching definitions."}, {"ID": "OCT5-AI-037-1", "Initiative_ID": "AI-037", "Title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "Date": "2026-03-10", "Finding": "Aramco reports SAR 19.9bn ($5.3bn) FY2025 Technology Realized Value: SAR 9.8bn ($2.6bn) attributed to AI-driven solutions and SAR 10.1bn ($2.7bn) to non-AI technologies. The report defines TRV as incremental financial benefit versus baseline from capital or operating expenditure avoided or revenue increase. This replaces the less precise CEO-call description “over half.”", "Metric": "Company-reported FY2025 aggregate: $5.3bn total TRV, including $2.6bn AI-driven and $2.7bn non-AI-driven. The separate FY2025 results presentation supplies the total-portfolio benefit mix; no AI-only allocation by benefit type is given.", "Source_IDs": "OCT5-S043", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-037-2", "Initiative_ID": "AI-037", "Title": "Saudi Aramco FY2025 results presentation — TRV value composition", "Date": "2026-03-10", "Finding": "The FY2025 results material breaks the $5.3bn company-wide TRV into $3.9bn capex avoidance, $1.0bn opex reduction and $0.4bn revenue enhancement. Its “around half from AI solutions” statement is portfolio-level; it does not allocate the category composition to the $2.6bn AI subtotal.", "Metric": "Portfolio-wide FY2025 TRV mix: $3.9bn capex avoidance + $1.0bn opex reduction + $0.4bn revenue enhancement = $5.3bn total. No AI-only value-type allocation.", "Source_IDs": "OCT5-S044", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "Evidence": "Official operator FY2025 results presentation slide 10 independently reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Latest_disclosure": "2026-03-10", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai-037", "root_backfill-002"], "Research_search": "{\"packet_id\": \"oct6-executive-ai-037\", \"review_date\": \"2026-10-06\", \"title\": \"Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value\", \"publisher\": \"Aramco\", \"source_title\": \"Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value\", \"source_date\": \"2026-03-10\", \"source_type\": \"Operator annual report\", \"source_locator\": \"Printed pp. 64–65 / PDF pp. 37–38\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"Aramco reports the contribution of deployed AI solutions to its FY2025 technology portfolio. Individual projects are not separately valued here.\", \"technology_data\": \"TRV measures capital or operating expenditure avoided, or added revenue, versus a baseline. The annual report splits AI-driven and non-AI value; it does not disclose the calculation for each project.\", \"deployment_scope\": \"FY2025 aggregate; AI decomposition across value types and projects not disclosed.\", \"practical_relevance\": \"The AI subtotal gives a portfolio figure to investigate. Ask which projects produced it and how benefits compare with the full cost of implementation and ongoing operations.\", \"business_problem\": \"Understanding how much of a technology portfolio’s reported benefit comes from AI.\", \"economics\": \"TRV is not net cash ROI; no project-level bridge or verifier opinion reviewed.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Portfolio aggregate; issuer says third-party verified but verification scope not reviewed. No project-level denominator or AI-only benefit composition. Do not allocate total capex avoidance to AI or add distinct-year records.\"], \"timeline_event\": \"FY2025 TRV and AI/non-AI splits reported in annual report; deployment dates not supplied.\", \"timeline_date\": \"2026-03-10\", \"source_ids\": [\"OCT5-S043\", \"RS109\"], \"supplementary_sources\": [{\"title\": \"Saudi Aramco FY 2025 Results presentation\", \"url\": \"https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf\", \"date\": \"2026-03-10\", \"publisher\": \"Aramco\", \"locator\": \"Slide 10 / PDF p. 10: $3.9bn capex avoidance, $1.0bn opex reduction, $0.4bn revenue enhancement; mix is total TRV only.\"}], \"stage\": \"FY2025 aggregate technology value reported for deployed technologies; AI subtotal is portfolio-level.\", \"key\": \"oct6-executive-ai-037\", \"record_ids\": [\"AI-037\"], \"company\": \"Aramco\", \"category\": \"Internal adoption\", \"department\": \"Finance, investment & strategy\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S043\"], \"workflow\": [\"OCT5-S043\"], \"technology_data\": [\"OCT5-S043\"], \"deployment_scope\": [\"OCT5-S043\"], \"practical_relevance\": [\"OCT5-S043\"], \"economics\": [\"OCT5-S043\"], \"integration_human_role\": [\"OCT5-S043\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \"Issuer-reported $2.6bn AI-attributed TRV; $3.9bn capex avoidance is for total portfolio and must not be allocated to AI.\", \"period\": \"2026-03-10\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S043\"]}], \"timeline\": [{\"date\": \"2026-03-10\", \"event\": \"FY2025 TRV and AI/non-AI splits reported in annual report; deployment dates not supplied.\", \"source_ids\": [\"OCT5-S043\"]}], \"sources\": [{\"id\": \"OCT5-S043\", \"title\": \"Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value\", \"url\": \"https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf\", \"date\": \"2026-03-10\", \"type\": \"Operator annual report\", \"publisher\": \"Aramco\", \"locator\": \"Printed pp. 64–65 / PDF pp. 37–38\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}, {\"id\": \"RS109\", \"title\": \"Saudi Aramco FY 2025 Results presentation\", \"url\": \"https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf\", \"date\": \"2026-03-10\", \"type\": \"Investor presentation\", \"publisher\": \"Aramco\", \"locator\": \"Slide 10 / PDF p. 10: Technology Realized Value, total-portfolio mix\", \"review_status\": \"Full primary text reviewed\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "No initiative-level allocation, causal methodology or independent validation is disclosed; the AI share is high-level management attribution.", "source_url": ""}], "Audit_review": {"id": "AI-037", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks"], "findings": [{"severity": "medium", "field": "metrics / economics", "issue": "Total 2025 TRV is $5.3bn and combines technology value, but the CEO media-call source explicitly says over half was enabled by AI. Earlier audit wording that no AI allocation was disclosed was too strong.", "correction": "Retain the source’s company-reported “over half enabled by AI” attribution; keep total TRV and component values clearly marked as aggregate, not project-level or independently validated.", "source_url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks"}], "residual_gaps": ["No initiative-level allocation, causal methodology or independent validation is disclosed; the AI share is high-level management attribution."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Aramco describes its aggregate TRV metric as third-party verified and incremental to a business baseline. No independent assurance opinion was reviewed here. The $2.6bn AI subtotal is company-reported; AI-only allocation by project or benefit type and full costs are not disclosed. TRV includes capital or operating expenditure avoided and revenue added. Capital avoidance is not incremental cash income, and aggregate TRV is not project-level net cash ROI.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Relevant original annual-report pages reviewed on 2026-10-06.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-038", "Company": "Aramco", "Segment": "NOC / integrated", "Theme": "Operational AI", "Initiative": "Industrial AI deployment portfolio", "Extracted_finding": "Portfolio context for assessing scale; individual workflow outcomes must be evaluated separately.", "Stage_at_disclosure": "Company-reported deployed portfolio; additional development pipeline", "Reported_metric": "AI-driven Technology Realized Value: $1.8 billion (Aramco-reported portfolio measure)", "Interpretation_limit": "No public workflow-level cost or benefit bridge in this announcement. Do not add the portfolio value to the benefits of constituent initiatives.", "Source_IDs": "S22; R004; OCT5-S023", "Source_date": "2025-11-24", "Source_type": "Company announcement / remarks", "Source_locator": "AI deployment and NVIDIA sections", "Source_URL": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-11-24; see chronology for operating dates.", "Context": "Downtime, maintenance failures and production efficiency.", "Implementation": "AI models support facility monitoring, maintenance prediction and production optimization.", "Assessment": "Broad organizational ambition. These counts are not equivalent to unique economic projects, and do not establish deployment depth at every facility.", "Detail_words": 67, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Enterprise deployment scale", "Scope_type": "Company AI portfolio or program", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-004"], "Related_IDs": [], "Disclosures": [{"ID": "REF-004", "Initiative_ID": "AI-038", "Title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "Date": "2025-06", "Finding": "Aramco says it processes ~10 billion data points daily and is using AI to reprocess old seismic data, optimize autonomous drilling and detect pipeline corrosion. It is training 6,000+ AI developers with Imperial College, Caltech and KAUST, and partnered with Groq on a regional AI inferencing center.", "Metric": "10B data points/day; 1,500 PB storage; 6,000+ AI developers being trained; 5 WEF Lighthouse awards", "Source_IDs": "R004", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Pair large-scale technical upskilling with high-value 'old data, new answers' use cases like seismic reprocessing to show quick ROI.", "Partners": "Groq, Qualcomm, Imperial College, Caltech, KAUST", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "ORIG-AI-038", "Initiative_ID": "AI-038", "Title": "AI deployment portfolio", "Date": "2025-11-24", "Finding": "Models address downtime, maintenance and production optimization.", "Metric": "442 identified use cases; >200 solutions deployed; >100 in development", "Source_IDs": "S22", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The company identified 442 use cases, with more than 200 solutions deployed and over 100 in development. It also described GPU computing and AI workforce training.", "Context": "Aramco’s late-2025 portfolio spans facility downtime, maintenance and production optimization, using large volumes of operating data.", "Assessment": "Identified, developing and deployed counts refer to different stages. They are a dated portfolio snapshot, rather than an independently verified count of economically successful applications."}, {"ID": "OCT5-AI-038-1", "Initiative_ID": "AI-038", "Title": "Qualcomm MoU for on-device edge AI over 450 MHz network", "Date": "2025-05-18", "Finding": "Aramco Digital and Qualcomm signed an MoU to explore edge AI on rugged devices, robots, drones, cameras and sensors connected through Aramco Digital's 450 MHz 5G industrial network. The register's AI-038 already lists Qualcomm as a partner; this adds the dated agreement source.", "Metric": "", "Source_IDs": "OCT5-S023", "Source_URL": "https://jpt.spe.org/aramco-signs-mous-worth-up-to-90-billion-for-lng-ai-emissions-reduction-and-asset-management-initiatives", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review"}], "Latest_disclosure": "2025-11-24", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["root_backfill-001"], "Research_search": " {\"key\": \"aramco-portfolio\", \"existing_ids\": [\"AI-038\"], \"company\": \"Aramco\", \"title\": \"Industrial AI deployment portfolio\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"stage\": \"Company-reported deployed portfolio; additional development pipeline\", \"verification\": \"verified primary\", \"partners\": [], \"sources\": [{\"id\": \"S22\", \"title\": \"Aramco deploys AI and supercomputing to drive digital transformation\", \"url\": \"https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation\", \"date\": \"2025-11-24\", \"locator\": \"AI deployment and NVIDIA sections\", \"publisher\": \"Aramco\", \"type\": \"Company announcement / remarks\", \"review_status\": \"Full primary text reviewed\", \"supports\": [], \"accessed\": \"2026-09-27\", \"original_id\": \"AR1\"}], \"field_sources\": {\"business_problem\": [\"S22\"], \"workflow\": [\"S22\"], \"technology_data\": [\"S22\"], \"integration_human_role\": [\"S22\"], \"deployment_scope\": [\"S22\"], \"economics\": [\"S22\"]}, \"disclosed_results\": [{\"metric\": \"AI-driven Technology Realized Value\", \"value\": \"$1.8 billion\", \"period\": \"2024\", \"attribution\": \"Aramco-reported portfolio measure\", \"source_ids\": [\"S22\"]}], \"timeline\": [{\"date\": \"2025-11-24\", \"event\": \"Portfolio scale disclosed.\", \"source_ids\": [\"S22\"]}], \"limitations\": [\"No public workflow-level cost or benefit bridge in this announcement.\", \"Do not add the portfolio value to the benefits of constituent initiatives.\"], \"economics\": \"Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.\", \"practical_relevance\": \"Portfolio context for assessing scale; individual workflow outcomes must be evaluated separately.\", \"business_problem\": \"Downtime, maintenance failures and production efficiency.\", \"workflow\": \"AI models support facility monitoring, maintenance prediction and production optimization.\", \"technology_data\": \"More than 10 billion data points daily; NVIDIA GPU computing supports drilling plans and well placement.\", \"integration_human_role\": \"A commitment to train more than 6,000 AI developers supports adoption. Application-level approval controls are not detailed.\", \"deployment_scope\": \"Late-2025 company portfolio: 442 identified use cases, over 200 deployed solutions and more than 100 in development.\", \"assessment\": \"Broad organizational ambition. These counts are not equivalent to unique economic projects, and do not establish deployment depth at every facility.\", \"corrections_to_existing\": \"Preserves the portfolio separately from the Fadhili process-control workflow.\", \"cohort\": \"root_backfill\", \"packet_id\": \"root_backfill-001\", \"source_ids\": [\"S22\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-038\"]}", "Research_status": "Expanded source review · Supplemented 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-038", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation"], "findings": [{"severity": "medium", "field": "workflow", "issue": "The source describes a portfolio spanning multiple AI applications rather than one bounded initiative.", "correction": "Retain as a broad portfolio record; avoid attributing specific deployment metrics without program-specific support.", "source_url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Basis: USD 1.8bn AI-driven Technology Realized Value reported Nov 2025, an earlier and differently defined measure than AI-037 (over half of FY2025 TRV of USD 5.3bn). Overlapping portfolio measures; not additive.", "Outcome_counted": false, "Outcome_count_note": "Portfolio aggregate overlaps other outcome profiles; excluded from headline and scorecard result counts.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-039", "Company": "Aramco / PIF / HUMAIN", "Segment": "NOC / sovereign capital", "Theme": "AI infrastructure", "Initiative": "Proposed minority investment in HUMAIN", "Extracted_finding": "Strategic AI ecosystem investment, separate from field adoption.", "Stage_at_disclosure": "Non-binding term sheet announced; completion subject to definitive agreement and approvals.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No verified closing date or operating outcome in the reviewed release.", "Source_IDs": "RS110; S22; R002", "Source_date": "2026-03-10", "Source_type": "Company announcement / remarks", "Source_locator": "HUMAIN investment plans; TRV highlights", "Source_URL": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2026-03-10; see chronology for operating dates.", "Context": "Expanding access to AI capabilities through ownership.", "Implementation": "FY 2025 results describe plans progressing toward a significant minority interest in HUMAIN.", "Assessment": "Potentially ambitious ownership strategy. The announcement provides no basis for treating transaction value as oilfield AI deployment spend.", "Detail_words": 57, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-002"], "Related_IDs": [], "Disclosures": [{"ID": "REF-002", "Initiative_ID": "AI-039", "Title": "Aramco takes significant minority stake in HUMAIN", "Date": "2025-10", "Finding": "PIF and Aramco signed a non-binding term sheet for Aramco to acquire a significant minority stake in HUMAIN, combining PIF and Aramco AI programs and assets under the PIF company. Nasser said it would strengthen Aramco's leadership in industrial AI applications.", "Metric": "Non-binding term sheet; PIF retains majority", "Source_IDs": "R002", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Majors are consolidating AI compute, models and talent into shared vehicles rather than building everything in-house; expect sovereign AI stacks to become preferred suppliers in the Gulf.", "Partners": "HUMAIN, PIF, ALLaM", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "ORIG-AI-039", "Initiative_ID": "AI-039", "Title": "HUMAIN minority investment plan", "Date": "2025-11-24", "Finding": "Proposed stake combines AI capabilities, data centers, models and cloud infrastructure.", "Metric": "Significant minority stake; no value given", "Source_IDs": "S22", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "HUMAIN’s intended scope includes data centers, cloud infrastructure, advanced models such as ALLAM, and AI applications.", "Context": "Aramco proposed a significant minority investment in PIF-backed HUMAIN to combine AI assets, capabilities and talent.", "Assessment": "The reviewed source establishes a proposed sovereign-capital transaction. It does not confirm closing, valuation or ownership percentage, and the activity differs from conventional buyout-fund investing."}], "Latest_disclosure": "2026-03-10", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["root_backfill-003"], "Research_search": " {\"key\": \"aramco-humain\", \"existing_ids\": [\"AI-039\"], \"company\": \"Aramco\", \"title\": \"Proposed minority investment in HUMAIN\", \"category\": \"Investment or acquisition\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"stage\": \"Planned investment; no closing established in reviewed evidence\", \"verification\": \"verified primary\", \"partners\": [], \"sources\": [{\"id\": \"RS110\", \"title\": \"Aramco announces fourth quarter and full-year 2025 results\", \"url\": \"https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release\", \"date\": \"2026-03-10\", \"locator\": \"HUMAIN investment plans; TRV highlights\", \"publisher\": \"Aramco\", \"type\": \"Company announcement / remarks\", \"review_status\": \"Full primary text reviewed\", \"supports\": [], \"accessed\": \"2026-09-27\", \"original_id\": \"AR6\"}], \"field_sources\": {\"business_problem\": [\"RS110\"], \"workflow\": [\"RS110\"], \"technology_data\": [\"RS110\"], \"integration_human_role\": [\"RS110\"], \"deployment_scope\": [\"RS110\"], \"economics\": [\"RS110\"]}, \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-03-10\", \"event\": \"Investment plans reiterated in annual results.\", \"source_ids\": [\"RS110\"]}], \"limitations\": [\"No verified closing date or operating outcome in the reviewed release.\"], \"economics\": \"Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.\", \"practical_relevance\": \"Strategic AI ecosystem investment, separate from field adoption.\", \"business_problem\": \"Expanding access to AI capabilities through ownership.\", \"workflow\": \"FY 2025 results describe plans progressing toward a significant minority interest in HUMAIN.\", \"technology_data\": \"The release does not disclose application architecture.\", \"integration_human_role\": \"Post-investment operating responsibilities are not established.\", \"deployment_scope\": \"Proposed corporate investment; transaction terms and ownership percentage are not established.\", \"assessment\": \"Potentially ambitious ownership strategy. The announcement provides no basis for treating transaction value as oilfield AI deployment spend.\", \"corrections_to_existing\": \"Retain planned status; the cited FY 2025 update does not establish closing.\", \"cohort\": \"root_backfill\", \"packet_id\": \"root_backfill-003\", \"source_ids\": [\"RS110\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-039\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-039", "company": "Aramco / PIF / HUMAIN", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release"], "findings": [{"severity": "high", "field": "stage", "issue": "Primary HUMAIN source describes a non-binding term sheet, subject to definitive agreements/approvals.", "correction": "Stage as proposed investment/agreement; do not present as completed acquisition or capital deployed.", "source_url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "investment", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-040", "Company": "ADNOC / AIQ", "Segment": "NOC / technology", "Theme": "Operational AI", "Initiative": "ENERGYai agentic subsurface workflows", "Extracted_finding": "Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.", "Stage_at_disclosure": "90-day PoC completed Jan 2025; planned operational version/test deployment was forward-looking", "Reported_metric": "accuracy improvement in major seismic interpretation aspects: 70% (Company-reported)", "Interpretation_limit": "Jan 2025 source verifies a PoC and forward plan, not the later completion or broad deployment claims in existing records.", "Source_IDs": "RS016; S23; R006; R007; R008; R009", "Source_date": "2025-01-16", "Source_type": "company announcement", "Source_locator": "Trial findings and next-phase paragraphs", "Source_URL": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-01-16; see chronology for operating dates.", "Context": "Accelerate seismic interpretation and improve reservoir monitoring/production insights using large volumes of company subsurface and operational data.", "Implementation": "Agents were evaluated for seismic survey analysis, reservoir-performance insight and anomaly detection; users could query results in natural language. Trial findings included improved data quality through error detection, standardization and dataset enrichment.", "Assessment": "Analyst inference: Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.", "Detail_words": 80, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Operational AI ambition", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ADNOC", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-006", "REF-007", "REF-008", "REF-009"], "Related_IDs": [], "Disclosures": [{"ID": "REF-007", "Initiative_ID": "AI-040", "Title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "Date": "2025-01", "Finding": "ADNOC and AIQ completed a 90-day trial of ENERGYai, a 70B-parameter LLM tuned on 50+ years of ADNOC data and drawing on over 15% of its onshore and offshore wells. The trial showed a 70% improvement in accuracy on major aspects of seismic interpretation plus better anomaly detection in reservoir monitoring.", "Metric": "70B-parameter LLM; 90-day PoC; >15% of wells; 70% seismic interpretation accuracy gain", "Source_IDs": "R007", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-boxed PoCs with a hard accuracy metric on a core technical workflow make the case for scaling agentic tools.", "Partners": "AIQ", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "ORIG-AI-040", "Initiative_ID": "AI-040", "Title": "ENERGYai", "Date": "2025-01-16", "Finding": "Agentic AI combines company data and an LLM for seismic interpretation, reservoir monitoring and production insights.", "Metric": "70% accuracy improvement in selected seismic aspects; 70bn-parameter LLM", "Source_IDs": "S23", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "A 90-day proof of concept used data from over 15% of ADNOC wells. The January 2025 report cites a 70% accuracy improvement in selected seismic aspects and planned five-agent test deployment in H1 2025.", "Context": "ENERGYai combines a 70-billion-parameter language model with ADNOC knowledge and proprietary data to support upstream technical work.", "Assessment": "The accuracy claim concerns selected trial tasks. The next-stage date was a target at that disclosure, and the source does not prove completion or a general 70% productivity gain."}, {"ID": "REF-006", "Initiative_ID": "AI-040", "Title": "ENERGYai deployment contract", "Date": "2025-03-12", "Finding": "AIQ announced a three-year, $340 million contract covering ENERGYai and associated AI solutions for ADNOC. The plan covers more than 28 producing fields, with an initial five-agent version expected in mid-2025 before wider scaling.", "Metric": "$340m / three years / planned coverage of 28+ producing fields", "Source_IDs": "R006", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Buying agentic AI as a multi-year outcome contract with a captive AI vendor is an alternative to piecemeal pilots.", "Partners": "AIQ, G42, Microsoft, OpenAI, OSDU", "Original_date": "2025-03", "Related_IDs": []}, {"ID": "REF-009", "Initiative_ID": "AI-040", "Title": "AIQ and SLB co-develop ENERGYai for subsurface", "Date": "2025-05", "Finding": "AIQ and SLB agreed to advance development and deployment of the ENERGYai agentic solution across ADNOC's subsurface operations, spanning geology, seismic and reservoir workflows. SLB brings its digital and agentic AI portfolio, including its Tela assistant.", "Metric": "No quantified metric supplied", "Source_IDs": "R009", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "OFS majors are becoming the channel for agentic subsurface AI; enablement builders should plan to integrate with SLB and Halliburton platforms.", "Partners": "AIQ, SLB (Tela)", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "REF-008", "Initiative_ID": "AI-040", "Title": "ADNOC moves AI from pilots to enterprise-wide operations", "Date": "2026-08", "Finding": "ADNOC is embedding AI across exploration, drilling, production, maintenance and decision-making, aiming to become 'the world's most AI-enabled energy company.' Tools include ENERGYai agents, the Neuron 5 asset-performance platform for predicting equipment failure, and an internal AI Lab that cuts proof-of-concept-to-deployment time 3x.", "Metric": "3x faster PoC-to-deployment; months-to-days on technical tasks", "Source_IDs": "R008", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A dedicated internal AI lab with a mandate to shorten PoC-to-production is what turns pilots into operating change.", "Partners": "AIQ, Microsoft, G42", "Original_date": "2026-08", "Related_IDs": []}], "Latest_disclosure": "2025-01-16", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-002"], "Research_search": " {\"key\": \"adnoc-energyai\", \"existing_ids\": [\"AI-040\"], \"company\": \"ADNOC\", \"title\": \"ENERGYai agentic subsurface workflows\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"90-day PoC completed Jan 2025; planned operational version/test deployment was forward-looking\", \"practical_relevance\": \"Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.\", \"business_problem\": \"Accelerate seismic interpretation and improve reservoir monitoring/production insights using large volumes of company subsurface and operational data.\", \"workflow\": \"Agents were evaluated for seismic survey analysis, reservoir-performance insight and anomaly detection; users could query results in natural language. Trial findings included improved data quality through error detection, standardization and dataset enrichment.\", \"technology_data\": \"70B-parameter LLM, 50+ years ADNOC knowledge, petabytes of internal data; trial used data from >15% of onshore/offshore wells. Detailed models, data controls and integrations not disclosed.\", \"integration_human_role\": \"ADNOC subject-matter experts participated; 100+ specialists and secure AI infrastructure supported the trial. Engineers interact with natural-language outputs.\", \"deployment_scope\": \"Proof-of-concept scope; next version was described as five agents with field tests and intended scale to thousands of wells, but completion evidence not included here.\", \"partners\": [\"AIQ\", \"G42\", \"Microsoft\"], \"disclosed_results\": [{\"metric\": \"accuracy improvement in major seismic interpretation aspects\", \"value\": \"70%\", \"period\": \"90-day PoC concluded Jan 2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS016\"]}], \"economics\": \"No financial return or operating impact disclosed; company reports 70% improvement in accuracy in major seismic interpretation aspects.\", \"timeline\": [{\"date\": \"2024-11-04\", \"event\": \"ENERGYai launched/development announced (outside primary source window but confirmed retrospectively in Jan 2025 release)\", \"source_ids\": [\"RS016\"]}, {\"date\": \"2025-01-16\", \"event\": \"90-day PoC completion announced; next operating version expected H1 2025, per plan\", \"source_ids\": [\"RS016\"]}], \"limitations\": [\"Jan 2025 source verifies a PoC and forward plan, not the later completion or broad deployment claims in existing records.\"], \"assessment\": \"Analyst inference: Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.\", \"sources\": [{\"id\": \"RS016\", \"title\": \"ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution\", \"url\": \"https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution\", \"publisher\": \"ADNOC\", \"date\": \"2025-01-16\", \"type\": \"company announcement\", \"locator\": \"Trial findings and next-phase paragraphs\", \"review_status\": \"full text reviewed\", \"supports\": [\"90-day ENERGYai proof-of-concept used a 70-billion-parameter LLM with ADNOC knowledge and proprietary operational data.\", \"Trial supported seismic interpretation, reservoir monitoring and anomaly detection; ADNOC reports 70% better accuracy in major seismic interpretation aspects.\", \"Initial planned operational version had five subsurface agents, asset test deployment and intention to scale across thousands of wells.\"], \"original_id\": \"adnoc_energyai\"}], \"field_sources\": {\"business_problem\": [\"RS016\"], \"workflow\": [\"RS016\"], \"technology_data\": [\"RS016\"], \"integration_human_role\": [\"RS016\"], \"deployment_scope\": [\"RS016\"], \"economics\": [\"RS016\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Exact-match audit: ENERGYai maps only to AI-040. AI-071 is employee capability-building, not the same initiative.\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-002\", \"source_ids\": [\"RS016\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-040\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-040", "company": "ADNOC / AIQ", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Result from a 90-day proof of concept.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-041", "Company": "Woodside Energy", "Segment": "E&P / LNG", "Theme": "Enterprise AI", "Initiative": "Lumina enterprise agent platform and adoption governance", "Extracted_finding": "Practical enterprise platform with governance, but agent count alone does not establish operating value.", "Stage_at_disclosure": "Company page last updated 2026-05-27; exact Lumina release date not stated", "Reported_metric": "Lumina agent count: Over 50 (Company-reported)", "Interpretation_limit": "Page is current and undated except last-updated stamp; use that separately from launch chronology. Maint Intel and Startup Advisor are separate existing exact initiatives AI-042 and AI-043 and not duplicated.", "Source_IDs": "S24", "Source_date": "Undated source; see chronology", "Source_type": "company webpage", "Source_locator": "Lumina, Copilot, ChatGPT Enterprise and governance", "Source_URL": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Staff need integrated everyday AI access while enterprise use requires cybersecurity, risk controls and consistent deployment.", "Implementation": "Lumina supports >50 agents for business-function tasks; Copilot handles Office workflows; ChatGPT Enterprise is sandbox for analytics/agent experimentation before agents move to Lumina. AI Centre of Excellence provides platform/tools; AI Council oversees risk, governance and priority.", "Assessment": "Practical enterprise platform with governance, but agent count alone does not establish operating value.", "Detail_words": 61, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Woodside", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-035"], "Related_IDs": [], "Disclosures": [{"ID": "REF-035", "Initiative_ID": "AI-041", "Title": "Lumina enterprise AI platform runs 50+ agents", "Date": "2025-12", "Finding": "Woodside's enterprise AI platform Lumina supports more than 50 AI agents handling day-to-day tasks across business functions, overseen by an AI Centre of Excellence and a senior-leader AI Council. Microsoft Copilot covers personal productivity and ChatGPT Enterprise is used as a sandbox for testing custom agents before they move into Lumina (date approximate; undated company page).", "Metric": "50+ AI agents in production", "Source_IDs": "S24", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A layered stack (Copilot for everyone, ChatGPT Enterprise as sandbox, an in-house agent platform for production) plus a governance council is a template worth copying.", "Partners": "Microsoft Copilot, ChatGPT Enterprise", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "ORIG-AI-041", "Initiative_ID": "AI-041", "Title": "Lumina, Copilot and ChatGPT Enterprise", "Date": "2026-05-27", "Finding": "Enterprise agents and employee tools support business workflows and prototype testing.", "Metric": ">50 Lumina agents", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Copilot serves personal productivity in Office 365. ChatGPT Enterprise provides an analytics and prototype environment before custom agents enter Lumina.", "Context": "Woodside’s Lumina platform supports more than 50 agents across business functions, alongside Microsoft Copilot and ChatGPT Enterprise.", "Assessment": "The architecture distinguishes employee assistance from integrated workflows. Agent count alone does not reveal frequency of use, human review effort or realized financial benefit."}], "Latest_disclosure": "", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-006"], "Research_search": " {\"key\": \"woodside-lumina\", \"existing_ids\": [\"AI-041\"], \"related_existing_ids\": [], \"company\": \"Woodside\", \"title\": \"Lumina enterprise agent platform and adoption governance\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company page last updated 2026-05-27; exact Lumina release date not stated\", \"practical_relevance\": \"Practical enterprise platform with governance, but agent count alone does not establish operating value.\", \"business_problem\": \"Staff need integrated everyday AI access while enterprise use requires cybersecurity, risk controls and consistent deployment.\", \"workflow\": \"Lumina supports >50 agents for business-function tasks; Copilot handles Office workflows; ChatGPT Enterprise is sandbox for analytics/agent experimentation before agents move to Lumina. AI Centre of Excellence provides platform/tools; AI Council oversees risk, governance and priority.\", \"technology_data\": \"Enterprise agent platform; Copilot/M365 and ChatGPT Enterprise. Model/data connections for individual Lumina agents not disclosed.\", \"integration_human_role\": \"Responsible AI assessment for every use case and senior cross-functional council; platform supports staff tasks rather than autonomous operating decisions.\", \"deployment_scope\": \"Woodside enterprise-wide adoption claim; >50 agents, but no department, user count or asset-specific production scale.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Lumina agent count\", \"value\": \"Over 50\", \"period\": \"Page updated 2026-05-27\", \"attribution\": \"Company-reported\", \"source_ids\": [\"S24\"]}], \"economics\": \"No quantitative value or cost reduction.\", \"timeline\": [{\"date\": \"updated 2026-05-27; page undated\", \"event\": \"Lumina enterprise agent platform and adoption governance documented or disclosed.\", \"source_ids\": [\"S24\"]}], \"limitations\": [\"Page is current and undated except last-updated stamp; use that separately from launch chronology.\", \"Maint Intel and Startup Advisor are separate existing exact initiatives AI-042 and AI-043 and not duplicated.\"], \"assessment\": \"Practical enterprise platform with governance, but agent count alone does not establish operating value.\", \"sources\": [{\"id\": \"S24\", \"title\": \"Woodside Artificial Intelligence\", \"url\": \"https://www.woodside.com/what-we-do/innovation/artificial-intelligence\", \"publisher\": \"Woodside\", \"date\": \"updated 2026-05-27; page undated\", \"type\": \"company webpage\", \"locator\": \"Lumina, Copilot, ChatGPT Enterprise and governance\", \"review_status\": \"full text reviewed\", \"supports\": [\"Woodside says Lumina supports >50 agents for business tasks; Microsoft Copilot supports Office productivity; ChatGPT Enterprise is an AI sandbox before custom agents integrate with Lumina.\", \"Responsible AI assessment, Centre of Excellence and executive AI Council described.\", \"Same page includes Maint Intel and Startup Advisor, already treated as separate exact profiles AI-042/AI-043; page does not state initial deployment dates.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-008\"}], \"field_sources\": {\"business_problem\": [\"S24\"], \"workflow\": [\"S24\"], \"technology_data\": [\"S24\"], \"integration_human_role\": [\"S24\"], \"deployment_scope\": [\"S24\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-006\", \"source_ids\": [\"S24\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-041\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date.", "source_url": ""}], "Audit_review": {"id": "AI-041", "company": "Woodside Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/what-we-do/innovation/artificial-intelligence"], "findings": [{"severity": "medium", "field": "dates", "issue": "Woodside’s current page lacks a dated first deployment for Lumina; last-update/access date is not the launch date.", "correction": "Keep timeline undated/current-page only and do not claim first deployment occurred in-window.", "source_url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence"}], "residual_gaps": ["Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-042", "Company": "Woodside Energy", "Segment": "E&P / LNG", "Theme": "Operational AI", "Initiative": "Maint Intel AI maintenance interval recommendations", "Extracted_finding": "Woodside’s AI page supports the previously unresolved 15% maintenance-hours estimate at Angel.", "Stage_at_disclosure": "Operating. Woodside says AI maintenance-interval tool Maint Intel was tested on Angel and launched at North West Shelf; global rollout was scheduled during 2026. Existing operation on NWS is established, broader rollout remains planned.", "Reported_metric": "Company-reported model processing time: five days to under two hours. Separately, the AI page identifies potential 15% lower annual maintenance hours at Angel over five years; not an achieved reduction.", "Interpretation_limit": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.", "Source_IDs": "RS006; S24", "Source_date": "2026-01-05", "Source_type": "company announcement", "Source_locator": "Maint Intel", "Source_URL": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Tested at Angel and launched at North West Shelf. January article: 5 January 2026. AI page last updated 27 May 2026; original publication date and estimate period unknown.", "Context": "Maintenance intervals must preserve reliability while avoiding unnecessary preventive work and model-analysis delays.", "Implementation": "Analyzes maintenance records, identifies equipment failure modes and feeds statistical models that recommend optimal intervals against reliability targets.", "Assessment": "The maintenance-hours estimate now has a traceable source. Whether it becomes a saving depends on adoption, reliability and costs.", "Detail_words": 67, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Woodside", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-042", "Initiative_ID": "AI-042", "Title": "Maint Intel", "Date": "2026-05-27", "Finding": "Maintenance scheduling model compares historical performance with reliability targets.", "Metric": "Potential 15% annual maintenance-hour reduction at Angel over five years", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Woodside launched the tool at the North West Shelf Project. At Angel, it identified potential to reduce annual maintenance hours by 15% over the next five years.", "Context": "Maint Intel recommends maintenance schedules by comparing historical equipment performance with reliability targets.", "Assessment": "This is identified potential. The source does not report those hours as already avoided. Reliability outcomes and implementation costs would determine the value of changing maintenance intervals."}], "Latest_disclosure": "2026-01-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-006"], "Research_search": "{\"key\": \"woodside-maint-intel\", \"existing_ids\": [\"AI-042\"], \"company\": \"Woodside\", \"title\": \"Woodside’s maintenance estimate has a traceable source\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Operating at North West Shelf; potential maintenance-hours reduction\", \"practical_relevance\": \"Were the recommended intervals adopted, and did reliability hold while maintenance hours fell?\", \"business_problem\": \"Maintenance intervals must preserve reliability while avoiding unnecessary preventive work and model-analysis delays.\", \"workflow\": \"Analyzes maintenance records, identifies equipment failure modes and feeds statistical models that recommend optimal intervals against reliability targets.\", \"technology_data\": \"Maint Intel compares maintenance history with reliability targets to recommend intervals. A separate January article identifies AWS support and the North West Shelf launch.\", \"integration_human_role\": \"Maintenance teams use recommended intervals; automatic execution is not established.\", \"deployment_scope\": \"Tested at Angel and launched at North West Shelf. January article: 5 January 2026. AI page last updated 27 May 2026; original publication date and estimate period unknown.\", \"partners\": [\"AWS\"], \"disclosed_results\": [{\"metric\": \"Model processing time\", \"value\": \"5 days to under 2 hours\", \"period\": \"Maint Intel project\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS006\"]}], \"economics\": \"Company-reported model processing time: five days to under two hours. Separately, the AI page identifies potential 15% lower annual maintenance hours at Angel over five years; not an achieved reduction.\", \"timeline\": [{\"date\": \"2026-01-05\", \"event\": \"Maint Intel launched at North West Shelf following Angel testing.\", \"source_ids\": [\"RS006\"]}, {\"date\": \"Updated 2026-05-27; publication day unknown\", \"review_on\": \"2026-10-18\", \"event\": \"Woodside’s AI page supports the previously unresolved 15% maintenance-hours estimate at Angel. The AI page describes potential 15% lower annual maintenance hours over five years, not hours already avoided. The January article reports model processing cut from five days to under two hours; that is not a field-labor saving.\", \"source_ids\": [\"S24\", \"RS006\"]}], \"limitations\": [\"The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.\"], \"assessment\": \"The maintenance-hours estimate now has a traceable source. Whether it becomes a saving depends on adoption, reliability and costs.\", \"sources\": [{\"id\": \"RS006\", \"title\": \"Right fix at right time – AI delivers for Woodside\", \"url\": \"https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside\", \"date\": \"2026-01-05\", \"publisher\": \"Woodside\", \"locator\": \"Angel pilot / North West Shelf launch, AWS collaboration and model processing-time paragraphs\", \"excerpt\": \"five days to under two hours\", \"access\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\", \"review_status\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\"}, {\"id\": \"S24\", \"title\": \"Artificial Intelligence at Woodside\", \"url\": \"https://www.woodside.com/what-we-do/innovation/artificial-intelligence\", \"date\": \"Updated 2026-05-27; publication day unknown\", \"publisher\": \"Woodside\", \"locator\": \"Maintenance section; last-updated stamp at foot of page\", \"excerpt\": \"a potential 15% reduction of maintenance hours\", \"access\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\", \"review_status\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"RS006\", \"S24\"], \"workflow\": [\"RS006\", \"S24\"], \"technology_data\": [\"RS006\", \"S24\"], \"deployment_scope\": [\"RS006\", \"S24\"], \"economics\": [\"RS006\", \"S24\"], \"integration_human_role\": [\"RS006\", \"S24\"], \"practical_relevance\": [\"RS006\", \"S24\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-006\", \"source_ids\": [\"RS006\", \"S24\"], \"report_date\": \"2026-10-18\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-042\"], \"earlier_review\": {\"report_date\": \"2026-09-27\", \"assessment\": \"Analyst inference: Named workflow, partner, pilot and reported computational improvement; potential maintenance-hour savings are prospective and not independently verified.\", \"limitations\": [\"The reopened source does not substantiate the 15% annual maintenance-hours estimate present in the prior packet; that figure should be omitted pending a traceable source.\", \"Further global rollout was scheduled during 2026; completion was not confirmed in this source.\"], \"note\": \"Superseded by targeted 18 October 2026 source check.\"}, \"last_source_check\": \"2026-10-18\"}", "Research_status": "Targeted source detail checked 18 October 2026 evening", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-042", "company": "Woodside Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside"], "findings": [{"severity": "high", "field": "metrics / economics", "issue": "Prior packet included 15% yearly maintenance-hours reduction for five years, but the reopened January 2026 Woodside article reports only potential savings generically and does not provide that figure.", "correction": "Remove the 15% metric and related economics sentence unless supported by another traceable source; retain the five-day-to-under-two-hour model-processing-time result.", "source_url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside"}], "residual_gaps": ["Achieved annual maintenance-hours reduction, reliability effects, costs and global rollout completion remain unverified."], "partition": 2, "reviewer_outcome": "corrected", "source_recheck_on": "2026-10-18", "resolution": "S24 maintenance section now traced. Earlier audit correctly found no 15% figure in the January article alone; current source-supported estimate remains prospective."}, "Audit_outcome": "corrected", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-042", "stage": "Operating", "stage_basis": "Woodside says AI maintenance-interval tool Maint Intel was tested on Angel and launched at North West Shelf; global rollout was scheduled during 2026. Existing operation on NWS is established, broader rollout remains planned.", "source_url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "source_date": "2026-01-05", "evidence_origin": "Operator", "limits": "Global rollout completion unverified. The separate AI page supports a potential 15% maintenance-hours estimate, not an achieved saving.", "source_id": "RS006"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.", "Outcome_counted": false, "Outcome_count_note": "Potential maintenance-hours reduction remains prospective. Processing time is a separate computational measure, not maintenance labor avoided; no new counted result.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Operating at North West Shelf; potential maintenance-hours reduction", "independent_validation": "Not established by this collection"}, "Earlier_source_review": {"as_of": "Before 18 October evening source check", "finding": "Named maintenance recommendation workflow, stated processing-time improvement and staged implementation; annual maintenance savings were not substantiated in the reopened source.", "limit": "The reopened source does not substantiate the 15% annual maintenance-hours estimate present in the prior packet; that figure should be omitted pending a traceable source. Further global rollout was scheduled during 2026; completion was not confirmed in this source.", "source_date": "2026-01-05", "note": "Earlier review retained for provenance; current fields below supersede it."}, "Current_source_check": "2026-10-18", "Research_origin": ["research/watch-2026-10-18-pm/README.md"]}, {"ID": "AI-043", "Company": "Woodside Energy", "Segment": "E&P / LNG", "Theme": "Operational AI", "Initiative": "Startup Advisor for LNG plant commissioning/startup", "Extracted_finding": "The Startup Advisor combines AI, automation and digital procedures to guide LNG plant start-up operators step by step and compare live performance with prior proven standards; Woodside reports it has been used in live operations and critical situations, without a dated first deployment or quantified safety/production result.", "Stage_at_disclosure": "Proven in live operations; date of first deployment not disclosed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Official page does not date the use case, so deployment cannot be confirmed to have started in research window.", "Source_IDs": "S24", "Source_date": "Undated source; see chronology", "Source_type": "company announcement", "Source_locator": "Startup Advisor", "Source_URL": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "LNG plant startups are complex, safety-sensitive and depend on experienced operator knowledge.", "Implementation": "Guides operators through digital procedures and compares live startup conditions with established performance standards; it can review past startups and inform real-time decisions in critical situations.", "Assessment": "Analyst inference: practical decision-support workflow is concrete, but the deployment date is undated and no performance measure is disclosed.", "Detail_words": 61, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Woodside", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-043", "Initiative_ID": "AI-043", "Title": "Startup Advisor", "Date": "2026-05-27", "Finding": "AI and digital procedures support LNG startup decisions using historical and live performance.", "Metric": "Not disclosed", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "It combines AI, automation and digital procedures, comparing real-time plant performance with historical startup experience. Woodside says operators have used it in live operations.", "Context": "Woodside’s Startup Advisor supports operators during complex LNG plant startups.", "Assessment": "The disclosure supports operational use. It does not quantify startup duration, failure reduction or cash savings, and it describes support for operators rather than unattended plant control."}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-020"], "Research_search": " {\"key\": \"woodside-startup-advisor\", \"existing_ids\": [\"AI-043\"], \"company\": \"Woodside\", \"title\": \"Startup Advisor for LNG plant commissioning/startup\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Proven in live operations; date of first deployment not disclosed\", \"practical_relevance\": \"The Startup Advisor combines AI, automation and digital procedures to guide LNG plant start-up operators step by step and compare live performance with prior proven standards; Woodside reports it has been used in live operations and critical situations, without a dated first deployment or quantified safety/production result.\", \"business_problem\": \"LNG plant startups are complex, safety-sensitive and depend on experienced operator knowledge.\", \"workflow\": \"Guides operators through digital procedures and compares live startup conditions with established performance standards; it can review past startups and inform real-time decisions in critical situations.\", \"technology_data\": \"AI, automation and digital procedures; data inputs and model architecture are not disclosed.\", \"integration_human_role\": \"Control-room leads and panel operators use and endorse the tool; tool supports rather than replaces operational judgment.\", \"deployment_scope\": \"Woodside LNG operations; specific sites and deployment dates not disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No quantified time, safety or production benefit disclosed.\", \"timeline\": [{\"date\": \"undated\", \"event\": \"Woodside describes live-operational use in official AI page; reviewed 2026-09-27.\", \"source_ids\": [\"S24\"]}], \"limitations\": [\"Official page does not date the use case, so deployment cannot be confirmed to have started in research window.\"], \"assessment\": \"Analyst inference: practical decision-support workflow is concrete, but the deployment date is undated and no performance measure is disclosed.\", \"sources\": [{\"id\": \"S24\", \"title\": \"Woodside Artificial Intelligence — Startup Advisor\", \"url\": \"https://www.woodside.com/what-we-do/innovation/artificial-intelligence\", \"publisher\": \"Woodside\", \"date\": null, \"accessed\": \"2026-09-27\", \"type\": \"company announcement\", \"locator\": \"Startup Advisor\", \"review_status\": \"full text reviewed\", \"supports\": [\"Startup Advisor uses AI, automation and digital procedures to guide LNG plant operators through startup steps and compare real-time performance to a standard.\", \"Woodside says it has been proven in live operations and used during critical situations to support decisions; control-room leads and panel operators endorsed the tool.\"], \"original_id\": \"majors-021\"}], \"field_sources\": {\"business_problem\": [\"S24\"], \"workflow\": [\"S24\"], \"technology_data\": [\"S24\"], \"integration_human_role\": [\"S24\"], \"deployment_scope\": [\"S24\"], \"economics\": [\"S24\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-020\", \"source_ids\": [\"S24\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-043\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date.", "source_url": ""}], "Audit_review": {"id": "AI-043", "company": "Woodside Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/what-we-do/innovation/artificial-intelligence"], "findings": [{"severity": "medium", "field": "dates", "issue": "Startup Advisor page gives no publication or launch date.", "correction": "Treat as a current capability with unverified in-window timing, not a dated deployment.", "source_url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence"}], "residual_gaps": ["Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-044", "Company": "Occidental", "Segment": "E&P", "Theme": "Operational AI", "Initiative": "AI-assisted subsurface work and Gulf of America operational use cases", "Extracted_finding": "Oxy management described current AI use in Gulf of America supply chain, asset integrity and reservoir characterization, and AI assistance in Permian/onshore subsurface work. A separate complex Gulf subsurface project was still under development; no AI-specific outcome was reported.", "Stage_at_disclosure": "Management-reported current use in selected Gulf and onshore work; complex Gulf subsurface project in development in 2025, with execution expected in 1–2 years.", "Reported_metric": "No AI-specific quantified outcome disclosed", "Interpretation_limit": "Evidence is management statements in third-party earnings-call transcripts. Do not merge current onshore/Gulf uses with the future complex Gulf project or attribute operating reliability/results solely to AI.", "Source_IDs": "R050; R051; S25", "Source_date": "2025-02-19; 2025-08-07; 2026-05-06 (management statements)", "Source_type": "Earnings-call transcripts (third-party transcriptions of management remarks)", "Source_locator": "Q4 2024 call: AI CoE and Gulf use cases; Q2 2025 call: onshore AI assistance and future Gulf project; Q1 2026 call: broader company AI role", "Source_URL": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "Evidence_status": "Management-reported transcript evidence reviewed; operational details and outcomes not independently validated", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Q4 2024 and Q2 2025 operating claims; Q1 2026 strategic update. The planned Gulf project is not reported as deployed.", "Context": "Oxy described AI use in selected operating functions and subsurface work, with a coordinated AI center of excellence; complex Gulf subsurface application remained a forward-looking project.", "Implementation": "Management named AI in Gulf of America supply-chain management, asset integrity and reservoir characterization, and said AI was already helping subsurface work in the Permian and other onshore areas. The CEO described the complex Gulf subsurface project as under development, with execution expected in the following one to two years. Model, data, human-review and implementation details were not disclosed.", "Assessment": "The transcripts support current management-reported use cases plus a separate planned Gulf subsurface project. They do not provide implementation detail, deployment scale, or AI-specific performance/economics.", "Detail_words": 63, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Occidental", "Review_tier": "Reviewed transcripts; secondary renderings", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-050", "REF-051"], "Related_IDs": [], "Disclosures": [{"ID": "REF-050", "Initiative_ID": "AI-044", "Title": "AI center of excellence and Gulf of America AI", "Date": "2025-02", "Finding": "Oxy said it created an AI center of excellence to align all intercompany AI initiatives. Its Gulf of America operations use AI for supply chain management, asset integrity and reservoir characterization. Oxy also cited large new seismic datasets, including Algeria's largest-ever seismic acquisition, as fuel for AI applications.", "Metric": "No quantified metric supplied", "Source_IDs": "R050", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A central AI CoE that aligns business-unit initiatives, applied first to a complex asset, is a common starting operating model.", "Partners": "", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-051", "Initiative_ID": "AI-044", "Title": "AI subsurface project targets complex Gulf reservoirs", "Date": "2025-08", "Finding": "CEO Vicki Hollub said Oxy is building internal AI capabilities, with a project aimed at the highly complex Gulf of America subsurface, and hoped to begin acting on results within 1-2 years. She said AI is already helping with Permian and other onshore subsurface work. Oxy also credited teams 'using AI' in its operations center with record artificial-lift reliability and workover performance.", "Metric": "No quantified metric supplied", "Source_IDs": "R051", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Point AI at the hardest subsurface problems where incremental recovery is worth the most, and report ops-center reliability wins alongside.", "Partners": "", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "ORIG-AI-044", "Initiative_ID": "AI-044", "Title": "AI-enhanced subsurface development", "Date": "2026-05-06", "Finding": "Plans combine technical data with AI to improve development and performance.", "Metric": "No AI-only metric", "Source_IDs": "S25", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The Q1 2026 earnings call presents an expansion of AI-enabled technical work, without isolating a named deployment’s economics.", "Context": "Occidental describes plans to combine technical subsurface data with AI to improve development decisions and operating performance.", "Assessment": "Historical well results and reserves replacement provide business context, but the call does not attribute them solely to AI. The disclosed ambition exceeds the implementation detail available here."}], "Latest_disclosure": "2026-05-06", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["north_america-020"], "Research_search": " {\"key\": \"oxy-subsurface-ai-claim\", \"existing_ids\": [\"AI-044\"], \"company\": \"Occidental\", \"title\": \"AI-enhanced subsurface development\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Management-reported current AI use in selected Gulf and Permian/onshore work; complex Gulf subsurface project in development in 2025 with execution expected in 1–2 years.\", \"practical_relevance\": \"Management named current AI use in operational functions and onshore subsurface work; a more complex Gulf subsurface project remained prospective. No AI-specific results were quantified.\", \"business_problem\": \"Apply AI to subsurface characterization/development and operational processes across complex assets.\", \"workflow\": \"Q4 2024 remarks named AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization. Q2 2025 remarks said AI already helped subsurface work in the Permian and other onshore areas; separately, a complex Gulf subsurface project was under development with execution expected in one to two years.\", \"technology_data\": \"Model, data inputs, software and validation are not disclosed in the reviewed transcript remarks.\", \"integration_human_role\": \"Management described an AI center of excellence coordinating initiatives; workflow-level human review and controls are not specified.\", \"deployment_scope\": \"Company remarks identify Gulf of America operational areas and Permian/other onshore subsurface work, but provide no asset counts, users or deployment scale. The complex Gulf subsurface project was prospective.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific economics or outcome metric disclosed. Do not attribute broader production or reliability results to AI.\", \"timeline\": [{\"date\": \"2025-02-19\", \"event\": \"Q4 2024 call described AI use in Gulf of America supply chain, asset integrity and reservoir characterization, and an AI center of excellence; Algeria seismic was described as future-use data.\", \"source_ids\": [\"R050\"]}, {\"date\": \"2025-08-07\", \"event\": \"Q2 2025 call described AI assistance in Permian/other onshore subsurface work and a separate complex Gulf project in development, with execution expected in 1–2 years.\", \"source_ids\": [\"R051\"]}, {\"date\": \"2026-05-06\", \"event\": \"Q1 2026 call described AI as taking a larger role across disciplines; no additional named subsurface workflow.\", \"source_ids\": [\"S25\"]}], \"limitations\": [\"Source record is based on third-party earnings-call transcripts of management remarks.\", \"Gulf and onshore use cases are management-reported; technical implementation and scale are not disclosed.\", \"The complex Gulf subsurface project was forward-looking in Aug. 2025 and must not be described as deployed.\", \"No AI-attributed result or economics disclosed.\"], \"assessment\": \"Management remarks support several named uses and existing onshore assistance, alongside a distinct planned Gulf subsurface project. They do not quantify implementation scale or results.\", \"sources\": [{\"id\": \"R050\", \"title\": \"Occidental Petroleum Q4 2024 Earnings Call Transcript\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/\", \"publisher\": \"The Motley Fool (transcription of Occidental management remarks)\", \"date\": \"2025-02-19\", \"type\": \"earnings-call transcript\", \"locator\": \"CEO remarks on Gulf of America and AI center of excellence\", \"review_status\": \"full text reviewed\", \"supports\": [\"Management named current AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization and an AI center of excellence.\", \"Algeria seismic acquisition was described as a dataset for future AI use; no outcome or implementation detail was disclosed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"R050\"}, {\"id\": \"R051\", \"title\": \"Occidental (OXY) Q2 2025 Earnings Call Transcript\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/\", \"publisher\": \"The Motley Fool (transcription of Occidental management remarks)\", \"date\": \"2025-08-07\", \"type\": \"earnings-call transcript\", \"locator\": \"CEO remarks on subsurface AI\", \"review_status\": \"full text reviewed\", \"supports\": [\"CEO said AI was already helping subsurface work in Permian and other onshore areas.\", \"Complex Gulf subsurface AI project remained under development; execution was expected in the following one to two years.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"R051\"}, {\"id\": \"S25\", \"title\": \"Occidental Q1 2026 Earnings Call Transcript\", \"url\": \"https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf\", \"publisher\": \"Occidental\", \"date\": \"2026-05-06\", \"type\": \"issuer-hosted earnings-call transcript\", \"locator\": \"CEO remarks on AI across disciplines\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Management said AI was taking a larger role across disciplines; this is broad context and does not add a named subsurface workflow.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-025\"}], \"field_sources\": {\"business_problem\": [\"R050\", \"R051\"], \"workflow\": [\"R050\", \"R051\"], \"technology_data\": [], \"integration_human_role\": [\"R050\"], \"deployment_scope\": [\"R050\", \"R051\"], \"economics\": [\"R050\", \"R051\"]}, \"verification\": \"management-reported transcript evidence reviewed; not independently validated\", \"corrections_to_existing\": \"Q4 2024 and Q2 2025 calls support current named uses and onshore assistance; the complex Gulf subsurface project was still prospective. This corrects the earlier Q1-2026-only review framing.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-020\", \"source_ids\": [\"R050\", \"R051\", \"S25\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-044\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "management-reported transcript evidence reviewed; not independently validated", "Audit_flags": [{"claim": "Second-pass review", "reason": "Transcript sources are third-party renderings of management remarks; exact implementation, scale and AI-attributed results are not disclosed.", "source_url": ""}], "Audit_review": {"id": "AI-044", "company": "Occidental", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "stage / scope", "issue": "Oxy management described current AI use in Gulf of America supply chain, asset integrity and reservoir characterization in the Q4 2024 transcript, and AI assistance in Permian/onshore subsurface work in the Q2 2025 call. The complex Gulf subsurface project was still being developed, with execution expected in one to two years.", "correction": "Replace the Q1-2026-only “no specific workflow” reading: keep current onshore/Gulf operational use distinct from the forward-looking complex Gulf project; attribute transcript claims to management and note no AI-specific outcome was reported.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}, {"severity": "low", "field": "Supplemental source review", "issue": "The Q4 2024 call names current Gulf of America AI use in supply-chain management, asset integrity, and reservoir characterization and an enterprise AI center of excellence. Keep Algeria seismic dataset as future-use potential; no results or implementation detail were disclosed.", "correction": "", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/"}, {"severity": "low", "field": "Supplemental source review", "issue": "Oxy management describes AI already helping subsurface work in Permian/onshore, while the complex Gulf of America project is being developed with execution expected in the following one to two years.", "correction": "Separate existing onshore assistance from the forward-looking Gulf project; do not describe Gulf AI project as live/deployed. Retain source as secondary transcript.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}], "residual_gaps": ["Transcript sources are third-party renderings of management remarks; exact implementation, scale and AI-attributed results are not disclosed."], "partition": 3, "reviewer_outcome": "partially supported; current onshore use and planned Gulf work are distinct"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Reviewed third-party transcript", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": " D identifies the third-party transcript basis; it does not mean the management passages were unread.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "management-reported transcript evidence reviewed; not independently validated", "claim_status": "Management-reported transcript evidence reviewed; operational details and outcomes not independently validated", "independent_validation": "Not established by this collection"}}, {"ID": "AI-045", "Company": "Diamondback Energy", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "Machine learning / AI for artificial-lift choice and late-life well optimization", "Extracted_finding": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "Stage_at_disclosure": "Issuer letters describe ML/AI and process-automation work in November 2025, then broader scaling of automation and predictive maintenance in May 2026. An August 2026 third-party-hosted call transcript describes artificial-lift optimization in day-to-day use as incredibly early. Pilot is an analyst normalization of early use, not an explicit issuer project-stage label; project-specific scale remains undisclosed.", "Reported_metric": "No result specifically attributed to this AI workflow. Nov 2025 letter states the intended gains; May 2026 letter says expected future uptime and LOE improvements. Reported corporate production and LOE are not causally linked to AI.", "Interpretation_limit": "Qualitative issuer disclosure only; implementation level and actual performance unquantified. May 2026 release explicitly frames uptime/LOE as expected in coming years.", "Source_IDs": "RS040; S26; R047; SW7-011; SW7-012", "Source_date": "2025-11-03", "Source_type": "Issuer disclosure", "Source_locator": "PDF p. 1, lines 61-64; full 3-page PDF read", "Source_URL": "https://ir.diamondbackenergy.com/node/17156/pdf", "Evidence_status": "Original source opened; relevant passage read.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives.", "Context": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "Implementation": "Issuer letter confirms ML and AI and the target workflow, but does not specify data inputs, models, automation/control mode, vendor, well count or roll-out boundaries.", "Assessment": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "Detail_words": 65, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Diamondback Energy", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-047", "REF-048"], "Related_IDs": [], "Disclosures": [{"ID": "REF-047", "Initiative_ID": "AI-045", "Title": "ML and AI in field ops cut downtime post-Endeavor", "Date": "2026-05", "Finding": "Diamondback management said that, after the Endeavor merger, layering machine learning onto its data streams and putting AI into field operations was reducing downtime. They cited this as a big part of the Q1 2026 production beat, together with completion-design and workover optimization.", "Metric": "No quantified metric supplied", "Source_IDs": "R047", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Combining merged teams' playbooks with ML on unified data streams can produce a production beat within a few quarters.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-048", "Initiative_ID": "AI-045", "Title": "Moving to 'pump by exception' with AI and automation", "Date": "2026-08", "Finding": "CFO Jere Thompson said AI 'is helping Diamondback in the office today' and that AI and automation will be very big drivers of a shallower or cheaper-to-maintain production base as the company moves to a pump-by-exception model. Operations leadership said AI and automation are applied first to day-to-day artificial-lift optimization and downtime management. CEO Kaes Van't Hof called it the 'first inning' and said even a 1% improvement in decline rate matters.", "Metric": "No quantified metric supplied", "Source_IDs": "S26", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Framing AI value as lower base decline and sustaining capex makes the investor case concrete, even when adoption is early.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "ORIG-AI-045", "Initiative_ID": "AI-045", "Title": "Lift optimization and downtime", "Date": "2026-08-10", "Finding": "Early AI and automation deployment targets daily lift optimization and reduced downtime.", "Metric": "No realized financial metric", "Source_IDs": "S26", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The August 2026 call explores the importance of improving base production and decline performance. It does not report a measured AI-specific decline-rate improvement.", "Context": "Diamondback discusses early AI and automation applications for daily artificial lift optimization and reduced downtime.", "Assessment": "A 1% decline-rate discussion is illustrative. Using it as an achieved AI result would convert a scenario into evidence. Named field scope and realized savings remain undisclosed."}, {"ID": "SW7-D-AI-045", "Initiative_ID": "AI-045", "Title": "Machine learning / AI for artificial-lift choice and late-life well optimization", "Date": "2025-11-03", "Finding": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "Metric": "No result specifically attributed to this AI workflow. Nov 2025 letter states the intended gains; May 2026 letter says expected future uptime and LOE improvements. Reported corporate production and LOE are not causally linked to AI.", "Source_IDs": "SW7-011; SW7-012", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2026-08-10", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-045", "north_america-014"], "Research_search": "{\"key\": \"swarm7-ai-045\", \"packet_id\": \"swarm7-ai-045\", \"record_ids\": [\"AI-045\"], \"existing_ids\": [\"AI-045\"], \"related_existing_ids\": [], \"company\": \"Diamondback Energy\", \"title\": \"Machine learning / AI for artificial-lift choice and late-life well optimization\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives. Late-life producing wells and artificial-lift processes; scale not disclosed.\", \"practical_relevance\": \"Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.\", \"business_problem\": \"\", \"workflow\": \"Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.\", \"technology_data\": \"Issuer letter confirms ML and AI and the target workflow, but does not specify data inputs, models, automation/control mode, vendor, well count or roll-out boundaries.\", \"integration_human_role\": \"Production operations teams are described as driving the automation; no approval/override mechanism disclosed.\", \"deployment_scope\": \"Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives. Late-life producing wells and artificial-lift processes; scale not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No result specifically attributed to this AI workflow. Nov 2025 letter states the intended gains; May 2026 letter says expected future uptime and LOE improvements. Reported corporate production and LOE are not causally linked to AI.\", \"period\": \"Period as described in source; metric measurement period not separately disclosed.\", \"attribution\": \"Issuer-reported; no independent validation.\", \"source_ids\": [\"SW7-011\", \"SW7-012\"]}], \"economics\": \"Intended to optimize lift cost and ultimately lower LOE; no AI-specific amount, achieved savings or payback reported.\", \"timeline\": [{\"date\": \"2025-11-03\", \"event\": \"Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives.\", \"source_ids\": [\"SW7-011\", \"SW7-012\"]}], \"limitations\": [\"Qualitative issuer disclosure only; implementation level and actual performance unquantified. May 2026 release explicitly frames uptime/LOE as expected in coming years.\"], \"assessment\": \"Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.\", \"sources\": [{\"title\": \"Diamondback Energy Letter to Stockholders (Q3 2025)\", \"date\": \"2025-11-03\", \"url\": \"https://ir.diamondbackenergy.com/node/17156/pdf\", \"locator\": \"PDF p. 1, lines 61-64; full 3-page PDF read\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"SW7-011\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Diamondback Energy Letter to Stockholders (Q1 2026)\", \"date\": \"2026-05-04\", \"url\": \"https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11\", \"locator\": \"lines 72-77, especially line 74; full release read\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"SW7-012\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW7-011\", \"SW7-012\"], \"field_sources\": {\"workflow\": [\"SW7-011\", \"SW7-012\"], \"technology_data\": [\"SW7-011\", \"SW7-012\"], \"deployment_scope\": [\"SW7-011\", \"SW7-012\"], \"practical_relevance\": [\"SW7-011\", \"SW7-012\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Issuer-reported; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [{"claim": "AI-driven lift optimization produced 1% decline improvement.", "reason": "Existing source is secondary; decline-rate reference was illustrative and no measured AI-specific result was verified.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/"}, {"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-045", "company": "Diamondback Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Asset_location": "Not disclosed", "Followup_review": {"date": "2026-10-01", "decision": "Enriched", "note": "Primary evidence dated November 2025 and May 2026; prior August secondary disclosure remains historical."}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Original source opened; relevant passage read.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-046", "Company": "SLB / Qualcomm", "Segment": "Oilfield services / technology", "Theme": "AI products", "Initiative": "Agora edge AI and Qualcomm collaboration", "Extracted_finding": "Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.", "Stage_at_disclosure": "MoU announced June 2026; product capability, no named deployment tied to the MoU", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "MoU and intended application; not evidence of deployment or customer results. Do not conflate SLB historical Agora deployments with this collaboration.", "Source_IDs": "RS065; S27", "Source_date": "2026-06-09", "Source_type": "company announcement", "Source_locator": "Collaboration scope", "Source_URL": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-06-09; see chronology for operating dates.", "Context": "Support real-time decisions in remote production environments where latency, constrained connectivity, power and continuity matter.", "Implementation": "Combine low-power edge AI processing with SLB Agora edge AI and IIoT solutions for applications at wells, facilities and production systems.", "Assessment": "Analyst inference: Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.", "Detail_words": 57, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-046", "Initiative_ID": "AI-046", "Title": "Agora edge AI", "Date": "2026-07-24", "Finding": "MoU combines low-power compute and SLB edge solutions for well, facility and production decisions.", "Metric": "Not disclosed", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The proposal pairs Qualcomm’s low-power processing with Agora edge AI and industrial connectivity.", "Context": "SLB and Qualcomm signed an MoU focused on AI at wells and production facilities.", "Assessment": "Edge processing places computation near equipment. The agreement does not establish completed joint deployments, latency improvements or autonomous control authority. Its operating value depends on a subsequent implemented use case."}], "Latest_disclosure": "2026-06-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-030"], "Research_search": " {\"key\": \"slb-agora-qualcomm\", \"existing_ids\": [\"AI-046\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"Agora edge AI and Qualcomm collaboration\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"MoU announced June 2026; product capability, no named deployment tied to the MoU\", \"practical_relevance\": \"Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.\", \"business_problem\": \"Support real-time decisions in remote production environments where latency, constrained connectivity, power and continuity matter.\", \"workflow\": \"Combine low-power edge AI processing with SLB Agora edge AI and IIoT solutions for applications at wells, facilities and production systems.\", \"technology_data\": \"Qualcomm low-power compute and AI; SLB Agora edge AI/IoT and production solutions.\", \"integration_human_role\": \"MoU does not specify the degree of automation, supervision or safety interlocks for future applications.\", \"deployment_scope\": \"Collaboration aims at energy operations; no customer or installed scope disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No contract value, deployment costs or quantified economic result.\", \"timeline\": [{\"date\": \"2026-06-09\", \"event\": \"SLB and Qualcomm announce MoU for energy edge AI solutions.\", \"source_ids\": [\"RS065\"]}], \"limitations\": [\"MoU and intended application; not evidence of deployment or customer results.\", \"Do not conflate SLB historical Agora deployments with this collaboration.\"], \"assessment\": \"Analyst inference: Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.\", \"sources\": [{\"id\": \"RS065\", \"title\": \"SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations\", \"url\": \"https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm\", \"publisher\": \"SLB\", \"date\": \"2026-06-09\", \"type\": \"company announcement\", \"locator\": \"Collaboration scope\", \"review_status\": \"full text reviewed\", \"supports\": [\"The MoU combines Qualcomm low-power edge computing/AI with SLB Agora for real-time decisions across wells, facilities and production systems.\"], \"original_id\": \"slb_qualcomm\"}], \"field_sources\": {\"business_problem\": [\"RS065\"], \"workflow\": [\"RS065\"], \"technology_data\": [\"RS065\"], \"integration_human_role\": [\"RS065\"], \"deployment_scope\": [\"RS065\"], \"economics\": [\"RS065\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-030\", \"source_ids\": [\"RS065\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-046\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-046", "company": "SLB / Qualcomm", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm"], "findings": [{"severity": "medium", "field": "stage", "issue": "Agora/Qualcomm announcement describes a collaboration and edge-AI solution scope, not a named operating customer.", "correction": "Classify as commercial collaboration/product roadmap until a field deployment is evidenced.", "source_url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-047", "Company": "SLB", "Segment": "Oilfield services", "Theme": "AI products", "Initiative": "Digital Marketplace for energy AI agents and applications", "Extracted_finding": "A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.", "Stage_at_disclosure": "Launched June 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company announcement reports marketplace inventory, not paid adoption or customer operating results.", "Source_IDs": "RS043; S27; R072", "Source_date": "2026-06-15", "Source_type": "company announcement", "Source_locator": "Marketplace offerings and governance", "Source_URL": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-06-15; see chronology for operating dates.", "Context": "Help energy companies discover, evaluate and deploy specialized digital/AI applications in existing digital environments.", "Implementation": "Governed store lists SLB and partner agents, models, skills, connectors and apps spanning Tela, Lumi and Delfi. Offers are vetted for security, interoperability and compatibility.", "Assessment": "Analyst inference: A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.", "Detail_words": 59, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Industry software platform", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-072"], "Related_IDs": [], "Disclosures": [{"ID": "REF-072", "Initiative_ID": "AI-047", "Title": "SLB Digital Marketplace lists ~200 AI and digital products", "Date": "2026-06", "Finding": "SLB launched a curated marketplace where customers can find and deploy AI agents, domain models, skills, data connectors and apps, including Tela AI skills and agents. Every listing is certified against SLB standards for security and interoperability, and ISVs can apply to publish through the SLB partner program.", "Metric": "~200 products at launch; 30+ partners", "Source_IDs": "R072", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "This is effectively an app store for O&G agents, so startups can use it as a distribution channel, but SLB controls certification and the customer relationship.", "Partners": "Tela, Lumi, Delfi, Ocean store, ISVs", "Original_date": "2026-06", "Related_IDs": []}, {"ID": "ORIG-AI-047", "Initiative_ID": "AI-047", "Title": "Digital Marketplace for Tela", "Date": "2026-07-24", "Finding": "Governed marketplace distributes AI capabilities from SLB and ecosystem developers.", "Metric": "Not disclosed", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Offerings connect with Tela and require security, compatibility and interoperability certification before listing.", "Context": "SLB’s Digital Marketplace distributes agents, domain models and supporting tools through a governed channel.", "Assessment": "A common distribution channel may reduce integration effort. Availability alone gives no evidence of paid adoption, deployed agents or incremental customer value. This is a product layer related to AI-008 and AI-009."}], "Latest_disclosure": "2026-06-15", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-002"], "Research_search": " {\"key\": \"slb-marketplace\", \"existing_ids\": [\"AI-047\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"Digital Marketplace for energy AI agents and applications\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched June 2026\", \"practical_relevance\": \"A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.\", \"business_problem\": \"Help energy companies discover, evaluate and deploy specialized digital/AI applications in existing digital environments.\", \"workflow\": \"Governed store lists SLB and partner agents, models, skills, connectors and apps spanning Tela, Lumi and Delfi. Offers are vetted for security, interoperability and compatibility.\", \"technology_data\": \"Marketplace links into SLB Lumi/Delfi platforms and Tela AI assistant; detailed technical interface architecture not disclosed.\", \"integration_human_role\": \"Users evaluate/access tools; publication and certification controls provide pre-listing governance.\", \"deployment_scope\": \"Launch inventory ~200 products, including >30 partners; listing does not confirm downstream deployment.\", \"partners\": [\"Over 30 partners and ISVs\"], \"disclosed_results\": [], \"economics\": \"Marketplace monetization model not quantified; SLB describes sales and consumption-scale ecosystem.\", \"timeline\": [{\"date\": \"2026-06-15\", \"event\": \"Marketplace launched\", \"source_ids\": [\"RS043\"]}], \"limitations\": [\"Company announcement reports marketplace inventory, not paid adoption or customer operating results.\"], \"assessment\": \"Analyst inference: A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.\", \"sources\": [{\"id\": \"RS043\", \"title\": \"SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy\", \"url\": \"https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy\", \"publisher\": \"SLB\", \"date\": \"2026-06-15\", \"type\": \"company announcement\", \"locator\": \"Marketplace offerings and governance\", \"review_status\": \"full text reviewed\", \"supports\": [\"Marketplace launched as channel for SLB and third parties to offer agents, domain models, tools, connectors and applications within existing environments.\", \"Approximately 200 products from SLB and 30+ partners; offers are certified for security, interoperability and compatibility before listing.\", \"Availability of products does not establish end-customer installation or operational results.\"], \"original_id\": \"slb_market\"}], \"field_sources\": {\"business_problem\": [\"RS043\"], \"workflow\": [\"RS043\"], \"technology_data\": [\"RS043\"], \"integration_human_role\": [\"RS043\"], \"deployment_scope\": [\"RS043\"], \"economics\": [\"RS043\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-002\", \"source_ids\": [\"RS043\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-047\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-047", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy"], "findings": [{"severity": "medium", "field": "stage", "issue": "Marketplace launch and listed products establish an offering, not resulting operator adoption.", "correction": "Describe marketplace launch; no customer deployment or outcome inferred.", "source_url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-048", "Company": "SLB", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "Intelligent Production Studio", "Extracted_finding": "This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.", "Stage_at_disclosure": "Launched at SLB Production and Recovery Summit, Q2 2026; demonstrator and training facility", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Studio demonstrations must not be presented as production deployment at customer assets. No public independent validation of any production uplift attributable to Studio.", "Source_IDs": "RS066; RS067; S27", "Source_date": "2026-06-18", "Source_type": "company feature", "Source_locator": "Autonomous pad and visitor details", "Source_URL": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-06-18; see chronology for operating dates.", "Context": "Let operators test integrated production systems and assess how digital workflows behave before committing to field deployment.", "Implementation": "Physical autonomous well pad runs real three-phase oil, water and gas flows across multiple streams; demonstrations include surveillance/anomaly detection, robotic inspection, autonomous chemical injection, digital twins and closed-loop optimization.", "Assessment": "Analyst inference: This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.", "Detail_words": 56, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-048", "Initiative_ID": "AI-048", "Title": "Intelligent Production Studio", "Date": "2026-07-24", "Finding": "Demonstration environment combines AI, edge intelligence, twins and autonomous well-pad workflows.", "Metric": "Autonomous pad with real three-phase flow", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB describes an autonomous well pad operating with real three-phase flow in a demonstration environment.", "Context": "Intelligent Production Studio demonstrates production workflows using AI, edge intelligence and digital twins.", "Assessment": "A physical demonstration is stronger than a software concept but differs from a customer field rollout. Commercial reliability, intervention rates and economics require evidence beyond the demonstration setting."}], "Latest_disclosure": "2026-06-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-031"], "Research_search": " {\"key\": \"slb-production-studio\", \"existing_ids\": [\"AI-048\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"Intelligent Production Studio\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched at SLB Production and Recovery Summit, Q2 2026; demonstrator and training facility\", \"practical_relevance\": \"This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.\", \"business_problem\": \"Let operators test integrated production systems and assess how digital workflows behave before committing to field deployment.\", \"workflow\": \"Physical autonomous well pad runs real three-phase oil, water and gas flows across multiple streams; demonstrations include surveillance/anomaly detection, robotic inspection, autonomous chemical injection, digital twins and closed-loop optimization.\", \"technology_data\": \"AI, edge intelligence, connected equipment, digital twins, robotic inspection, sensors and control systems integrated on operating infrastructure.\", \"integration_human_role\": \"Visitors evaluate and plan potential configuration; source describes a demonstrator rather than customer-operated autonomous assets.\", \"deployment_scope\": \"Houston demonstration/training environment; no external field deployment implied.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No standalone economics or measured customer outcomes disclosed.\", \"timeline\": [{\"date\": \"2026-Q2\", \"event\": \"Studio launched; more than 100 customers/partners visited by June 2026.\", \"source_ids\": [\"RS066\"]}], \"limitations\": [\"Studio demonstrations must not be presented as production deployment at customer assets.\", \"No public independent validation of any production uplift attributable to Studio.\"], \"assessment\": \"Analyst inference: This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.\", \"sources\": [{\"id\": \"RS066\", \"title\": \"Inside the Future of Production Operations\", \"url\": \"https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio\", \"publisher\": \"SLB\", \"date\": \"2026-06-18\", \"type\": \"company feature\", \"locator\": \"Autonomous pad and visitor details\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB says the Houston studio combines AI, automation and connected systems around an operating autonomous pad managing three-phase flow; over 100 customers and partners visited.\"], \"original_id\": \"slb_studio_launch\"}, {\"id\": \"RS067\", \"title\": \"Intelligent Production Studio\", \"url\": \"https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio\", \"publisher\": \"SLB\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company product documentation\", \"locator\": \"Inside the studio and FAQ\", \"review_status\": \"full text reviewed\", \"supports\": [\"Studio demonstrates AI surveillance, digital twins, robotic inspection, autonomous chemical injection and closed-loop workflows; this is demonstration scope.\"], \"original_id\": \"slb_studio_product\"}], \"field_sources\": {\"business_problem\": [\"RS066\", \"RS067\"], \"workflow\": [\"RS066\", \"RS067\"], \"technology_data\": [\"RS066\", \"RS067\"], \"integration_human_role\": [\"RS066\", \"RS067\"], \"deployment_scope\": [\"RS066\", \"RS067\"], \"economics\": [\"RS066\", \"RS067\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-031\", \"source_ids\": [\"RS066\", \"RS067\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-048\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-048", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Production Studio page is product material; results in related SLB digital-solution case studies should not be attributed to IPS without explicit linkage.", "correction": "Separate product capability from customer case metrics.", "source_url": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-049", "Company": "Aker BP / Cognite", "Segment": "E&P / software", "Theme": "Operational AI", "Initiative": "Atlas AI agents for root-cause analysis and E&P workflows", "Extracted_finding": "Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.", "Stage_at_disclosure": "Expanded partnership disclosed 2025-09-16", "Reported_metric": "Engineer time for root-cause analysis: Reduced by over 70% (partner-reported) (Vendor-reported customer statement)", "Interpretation_limit": "Cognite is source publisher; productivity metric lacks sample/baseline. Other workflows are described as ongoing initiatives, not necessarily all live.", "Source_IDs": "S28; R092", "Source_date": "2025-09-16", "Source_type": "partner case study", "Source_locator": "RCA agents and additional workflows", "Source_URL": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-09-16; see chronology for operating dates.", "Context": "Engineers investigate operational issues by manually gathering context; other processes rely on document-heavy supply-chain and equipment data workflows.", "Implementation": "RCA agents identify relevant information, construct cause maps and suggest next analysis steps. Related work automates document processing for supply chain and barrier management, registers equipment, prepares predictive-maintenance data foundations and optimizes production chemistry.", "Assessment": "Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.", "Detail_words": 75, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Aker BP", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-092"], "Related_IDs": [], "Disclosures": [{"ID": "REF-092", "Initiative_ID": "AI-049", "Title": "Aker BP expands AI-first strategy with Cognite Atlas AI", "Date": "2025-09", "Finding": "Aker BP deepened its Cognite partnership to deploy agent-based applications built with Cognite Atlas AI on Cognite Data Fusion. Use cases include root-cause analysis, document automation, barrier management, equipment registration, predictive maintenance and production chemistry.", "Metric": "Root-cause-analysis agents cut engineering time >70%", "Source_IDs": "R092", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A contextualized knowledge graph underneath is what makes agents credible in operations, and time saved per engineering task is the proof metric buyers respond to.", "Partners": "Aker BP, Cognite Atlas AI, Cognite Data Fusion", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "ORIG-AI-049", "Initiative_ID": "AI-049", "Title": "Atlas AI partnership expansion", "Date": "2025-09-16", "Finding": "Expanded agent adoption aims to improve efficiency and workforce productivity.", "Metric": "Partner reports >70% less engineering time for root-cause analysis", "Source_IDs": "S28", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The September 2025 announcement names document processing, equipment registration, barrier management and production chemistry. Root-cause agents retrieve information, build cause maps and suggest analysis steps.", "Context": "Aker BP expanded Cognite Atlas AI adoption, using pre-built agents and a low-code workbench connected to Cognite Data Fusion.", "Assessment": "Cognite reports over 70% less engineering time for root-cause analysis. This is a partner-reported task result. The release does not disclose sample size, total deployment cost or a company-wide cash benefit."}], "Latest_disclosure": "2025-09-16", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-007"], "Research_search": " {\"key\": \"aker-cognite-rca\", \"existing_ids\": [\"AI-049\"], \"related_existing_ids\": [], \"company\": \"Aker BP / Cognite\", \"title\": \"Atlas AI agents for root-cause analysis and E&P workflows\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Expanded partnership disclosed 2025-09-16\", \"practical_relevance\": \"Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.\", \"business_problem\": \"Engineers investigate operational issues by manually gathering context; other processes rely on document-heavy supply-chain and equipment data workflows.\", \"workflow\": \"RCA agents identify relevant information, construct cause maps and suggest next analysis steps. Related work automates document processing for supply chain and barrier management, registers equipment, prepares predictive-maintenance data foundations and optimizes production chemistry.\", \"technology_data\": \"Atlas AI agents and low-code workbench connected to Cognite Data Fusion with live OT, IT and engineering data; exact agent models/permission setup not disclosed.\", \"integration_human_role\": \"Engineers evaluate RCA agent suggestions; partner says human insight and autonomous agents work together.\", \"deployment_scope\": \"Aker BP partnership across exploration/production; no agent/site count reported in release.\", \"partners\": [\"Cognite\"], \"disclosed_results\": [{\"metric\": \"Engineer time for root-cause analysis\", \"value\": \"Reduced by over 70% (partner-reported)\", \"period\": \"As reported 2025-09-16\", \"attribution\": \"Vendor-reported customer statement\", \"source_ids\": [\"S28\"]}], \"economics\": \"Aker BP/Cognite report >70% reduction in RCA analysis time; no sample size, baseline, duration or dollar conversion. Other benefits qualitative.\", \"timeline\": [{\"date\": \"2025-09-16\", \"event\": \"Atlas AI agents for root-cause analysis and E&P workflows documented or disclosed.\", \"source_ids\": [\"S28\"]}], \"limitations\": [\"Cognite is source publisher; productivity metric lacks sample/baseline.\", \"Other workflows are described as ongoing initiatives, not necessarily all live.\"], \"assessment\": \"Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.\", \"sources\": [{\"id\": \"S28\", \"title\": \"Aker BP leverages Cognite Atlas AI to pioneer AI-first future\", \"url\": \"https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and\", \"publisher\": \"Cognite\", \"date\": \"2025-09-16\", \"type\": \"partner case study\", \"locator\": \"RCA agents and additional workflows\", \"review_status\": \"full text reviewed\", \"supports\": [\"Aker BP expanded Atlas AI adoption. Described workflows include document processing for supply chain, barrier management, automatic equipment registration, data foundation for predictive maintenance and production chemistry.\", \"Root-cause agents identify relevant info, build cause maps and suggest steps; Cognite/Aker BP report >70% reduction in engineer RCA time.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-009\"}], \"field_sources\": {\"business_problem\": [\"S28\"], \"workflow\": [\"S28\"], \"technology_data\": [\"S28\"], \"integration_human_role\": [\"S28\"], \"deployment_scope\": [\"S28\"], \"economics\": [\"S28\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-007\", \"source_ids\": [\"S28\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-049\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-049", "company": "Aker BP / Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "World Oil is secondary coverage and repeats Aker BP/Cognite claims, including the >70% RCA-time reduction; it is not independent validation.", "correction": "Keep the result labeled partner/company-reported and unbenchmarked; the existing Cognite primary partner release remains the preferred supporting citation.", "source_url": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-050", "Company": "PETRONAS", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "TIRAI GenAI assistant for myPROdata subsurface data", "Extracted_finding": "A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.", "Stage_at_disclosure": "Operating. PETRONAS’s 2025-11-03 release says its TIRAI assistant went live on 2025-10-18 for MBR2025 block pages in myPROdata.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No subscriber adoption or impact metrics disclosed.", "Source_IDs": "S29", "Source_date": "2025-11-03", "Source_type": "company announcement", "Source_locator": "TIRAI release", "Source_URL": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-11-03; see chronology for operating dates.", "Context": "Reduce friction in exploring Malaysia petroleum acreage data and finding relevant technical documents.", "Implementation": "Chatbot assists users on SmartViewer, SmartData Catalog, MBR and MBR+ pages and links to ElasticDocs reports, maps, images, seismic data and well markers.", "Assessment": "Analyst inference: A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.", "Detail_words": 64, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "PETRONAS", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-050", "Initiative_ID": "AI-050", "Title": "TIRAI subsurface assistant", "Date": "2025-11-03", "Finding": "Chatbot supports subsurface exploration and retrieval of reports, maps, seismic and well data.", "Metric": "Released October 18, 2025", "Source_IDs": "S29", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Released on 18 October 2025, the GenAI chatbot appears in SmartViewer, SmartData Catalog and MBR/MBR+ pages for MBR2025 blocks. ElasticDocs links users to reports, maps, seismic and well markers.", "Context": "PETRONAS launched TIRAI to help myPROdata users explore subsurface information during block evaluation.", "Assessment": "The disclosed function is data discovery and retrieval. The source does not establish autonomous interpretation, faster drilling decisions or measured user productivity."}], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-006"], "Research_search": " {\"key\": \"petronas-tirai\", \"existing_ids\": [\"AI-050\"], \"company\": \"PETRONAS\", \"title\": \"TIRAI GenAI assistant for myPROdata subsurface data\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Live from 18 Oct 2025 for MBR2025 block pages\", \"practical_relevance\": \"A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.\", \"business_problem\": \"Reduce friction in exploring Malaysia petroleum acreage data and finding relevant technical documents.\", \"workflow\": \"Chatbot assists users on SmartViewer, SmartData Catalog, MBR and MBR+ pages and links to ElasticDocs reports, maps, images, seismic data and well markers.\", \"technology_data\": \"Generative AI chatbot connected to PETRONAS myPROdata; underlying model, retrieval/indexing and safeguards not disclosed.\", \"integration_human_role\": \"Supports subscribers/investors in data exploration; human interpretation remains necessary for subsurface evaluation.\", \"deployment_scope\": \"myPROdata pages for MBR2025 blocks.\", \"partners\": [\"PETRONAS myPROdata\"], \"disclosed_results\": [], \"economics\": \"Commercial value and user outcomes not disclosed.\", \"timeline\": [{\"date\": \"2025-10-18\", \"event\": \"TIRAI went live on selected MBR2025 pages\", \"source_ids\": [\"S29\"]}], \"limitations\": [\"No subscriber adoption or impact metrics disclosed.\"], \"assessment\": \"Analyst inference: A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.\", \"sources\": [{\"id\": \"S29\", \"title\": \"TIRAI - Your New PETRONAS myPROdata Intelligence Assistance is Live!\", \"url\": \"https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live\", \"publisher\": \"PETRONAS myPROdata\", \"date\": \"2025-11-03\", \"type\": \"company announcement\", \"locator\": \"TIRAI release\", \"review_status\": \"full text reviewed\", \"supports\": [\"TIRAI GenAI chatbot went live on 18 October 2025 on SmartViewer, SmartData Catalog, MBR and MBR+ pages for MBR2025 blocks.\", \"Assistant links users to ElasticDocs resources including reports, maps, images, seismic data and well markers for subsurface data exploration.\"], \"original_id\": \"petronas_myprodata\"}], \"field_sources\": {\"business_problem\": [\"S29\"], \"workflow\": [\"S29\"], \"technology_data\": [\"S29\"], \"integration_human_role\": [\"S29\"], \"deployment_scope\": [\"S29\"], \"economics\": [\"S29\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-006\", \"source_ids\": [\"S29\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-050\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-050", "company": "PETRONAS", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-050", "stage": "Operating", "stage_basis": "PETRONAS’s 2025-11-03 release says its TIRAI assistant went live on 2025-10-18 for MBR2025 block pages in myPROdata.", "source_url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "source_date": "2025-11-03 (go-live 2025-10-18)", "evidence_origin": "Operator", "limits": "Deployment is scoped to specified block pages; no user adoption, answer-quality or time-saved metric reported.", "source_id": "S29"}, "Record_type": "vendor_product", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-051", "Company": "Blackstone / Enverus", "Segment": "Energy software / PE", "Theme": "AI products", "Initiative": "Enverus acquisition and reported capital deployment", "Extracted_finding": "Ownership transaction provides context to the energy software provider; it is not itself an AI implementation.", "Stage_at_disclosure": "Agreement announced August 2025; Enverus included in Q1 2026 capital deployed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The earnings disclosure supports capital deployment, but does not supply the exact legal closing date or a standalone Enverus transaction value. Investment ownership does not establish operating benefits from AI.", "Source_IDs": "S30; S47", "Source_date": "2026-04-23", "Source_type": "company announcement", "Source_locator": "Transaction announcement", "Source_URL": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "Blackstone funds; sellers Hellman & Friedman and Genstar", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-04-23; see chronology for operating dates.", "Context": "Blackstone announced acquisition of energy data and analytics company Enverus from Hellman & Friedman and Genstar.", "Implementation": "Investment/acquisition transaction only; no operating AI workflow is evidenced by the deal announcement.", "Assessment": "Classify as capital/ownership context, separate from Enverus ONE product evidence.", "Detail_words": 78, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Blackstone", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-098"], "Related_IDs": [], "Disclosures": [{"ID": "REF-098", "Initiative_ID": "AI-051", "Title": "Blackstone to acquire Enverus (~$6.5B reported)", "Date": "2025-08", "Finding": "Blackstone funds agreed to acquire Enverus, the energy data analytics and SaaS platform, from Hellman & Friedman and Genstar Capital. Reuters reported the deal at roughly $6.5B.", "Metric": "~$6.5B (reported)", "Source_IDs": "S30", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Private equity pays a premium for proprietary energy data that genAI workflows depend on, so owning data is what drives valuation.", "Partners": "Blackstone, Hellman & Friedman, Genstar", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "ORIG-AI-051", "Initiative_ID": "AI-051", "Title": "Enverus acquisition and AI scaling", "Date": "2025-08-06 / 2026-04-23", "Finding": "Buyout announcement links the platform’s growth to GenAI, asset decisions and energy-market analytics.", "Metric": "Financial terms not disclosed", "Source_IDs": "S30; S47", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Enverus reported 8,000 customers across 50 countries. Blackstone’s April 2026 results subsequently include Enverus in first-quarter capital deployed, updating the original agreement-only status.", "Context": "Blackstone agreed in August 2025 to acquire Enverus from Hellman & Friedman and Genstar. The investment thesis connects energy intelligence with AI and rising power demand.", "Assessment": "The aggregate private-equity deployment figure is not the Enverus purchase price. Acquisition terms and an AI-only valuation remain undisclosed in these sources. See AI-065 for the later product launch."}], "Latest_disclosure": "2026-04-23", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-001"], "Research_search": " {\"key\": \"blackstone-enverus-agreement\", \"existing_ids\": [\"AI-051\"], \"related_existing_ids\": [], \"company\": \"Enverus\", \"title\": \"Enverus acquisition and reported capital deployment\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Agreement announced August 2025; Enverus included in Q1 2026 capital deployed\", \"practical_relevance\": \"Ownership transaction provides context to the energy software provider; it is not itself an AI implementation.\", \"business_problem\": \"Blackstone announced acquisition of energy data and analytics company Enverus from Hellman & Friedman and Genstar.\", \"workflow\": \"Investment/acquisition transaction only; no operating AI workflow is evidenced by the deal announcement.\", \"technology_data\": \"Not disclosed for the transaction. Enverus software is described broadly as data analytics and insights.\", \"integration_human_role\": \"Not applicable to an AI workflow.\", \"deployment_scope\": \"Enverus customer footprint is described by the acquirer announcement, but no product deployment is detailed.\", \"partners\": [\"Hellman & Friedman\", \"Genstar\"], \"disclosed_results\": [], \"economics\": \"Financial terms were not disclosed in the reviewed announcement.\", \"timeline\": [{\"date\": \"2025-08-06\", \"event\": \"Blackstone funds announced agreement to acquire Enverus.\", \"source_ids\": [\"S30\"]}, {\"date\": \"2026-04-23\", \"event\": \"Q1 results name Enverus among private-equity capital deployed during the quarter.\", \"source_ids\": [\"S47\"]}], \"limitations\": [\"The earnings disclosure supports capital deployment, but does not supply the exact legal closing date or a standalone Enverus transaction value.\", \"Investment ownership does not establish operating benefits from AI.\"], \"assessment\": \"Classify as capital/ownership context, separate from Enverus ONE product evidence.\", \"sources\": [{\"id\": \"S30\", \"title\": \"Blackstone announces agreement to acquire Enverus\", \"url\": \"https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/\", \"publisher\": \"Blackstone\", \"date\": \"2025-08-06\", \"type\": \"company announcement\", \"locator\": \"Transaction announcement\", \"review_status\": \"full text reviewed\", \"supports\": [\"Blackstone funds agreed to acquire Enverus from Hellman & Friedman and Genstar; this is a financial transaction, not an AI deployment.\", \"Enverus provides energy data analytics and insights used by energy companies and investors.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-001\"}, {\"id\": \"S47\", \"title\": \"Blackstone First Quarter 2026 Earnings Release\", \"url\": \"https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf\", \"publisher\": \"Blackstone\", \"date\": \"2026-04-23\", \"type\": \"Earnings release\", \"locator\": \"PDF page 15 / printed page 11, Private Equity capital deployed\", \"review_status\": \"Primary PDF section reviewed\", \"supports\": [\"Enverus named among Q1 2026 capital deployed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"bx-q126\"}], \"field_sources\": {\"business_problem\": [\"S30\"], \"workflow\": [\"S30\"], \"technology_data\": [\"S30\"], \"integration_human_role\": [], \"deployment_scope\": [\"S30\"], \"economics\": [\"S30\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"The 2025 agreement is supplemented with the Q1 2026 capital-deployment disclosure; neither the segment total nor media valuations are assigned as a verified Enverus purchase price.\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-001\", \"source_ids\": [\"S30\", \"S47\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-051\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-051", "company": "Blackstone / Enverus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf"], "findings": [{"severity": "medium", "field": "stage", "issue": "Blackstone’s August 2025 source is an acquisition agreement; the earnings material does not establish Enverus-specific capital deployed.", "correction": "Keep agreement/closing status separate from capital deployment; use no unallocated Blackstone fund metrics.", "source_url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-052", "Company": "Quantum Capital Group", "Segment": "Energy PE", "Theme": "Enterprise AI", "Initiative": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "Extracted_finding": "More operationally concrete than a generic investment thesis: it describes portfolio data infrastructure and specific well-level and field optimization workflows. The publication date of the 2024 ESG report is within window, but its accelerator operating date is not. No independent portfolio-level outcomes.", "Stage_at_disclosure": "Operating platform described; undated page; 2024 report published 2025", "Reported_metric": "More than 100 AI accelerators reported as of May 2024; historical scope predates the main window. Undated Lakebase detail: Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.", "Interpretation_limit": "The accelerator count is historical, not a new 2026 deployment result. The 2026 report updates the firm platform; per-company coverage and project costs remain undisclosed. Lakebase limits: No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution.", "Source_IDs": "S31; RS076; OCT9-QCG-2026; WATCH21PM-DBX", "Source_date": "2025-05-02", "Source_type": "company webpage", "Source_locator": "QEC, QIQ and AI at Work", "Source_URL": "https://www.quantumcap.com/digital-mission/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "Energy-focused private equity manager", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-05-02; see chronology for operating dates.", "Context": "Private equity asset teams and portfolio operators need common data access and repeatable, fast screening/operating decisions across varied energy assets.", "Implementation": "Quantum Energy Cloud connects portfolio financial, operational and market data using Microsoft and Databricks. The 2026 report identifies Claude Enterprise as the firm interface succeeding QIQ.\n\nUndated Lakebase detail: Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.", "Assessment": "More operationally concrete than a generic investment thesis: it describes portfolio data infrastructure and specific well-level and field optimization workflows. The publication date of the 2024 ESG report is within window, but its accelerator operating date is not. No independent portfolio-level outcomes.", "Detail_words": 61, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Private-equity operating platform", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Quantum Capital Group", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum Energy Cloud", "Date": "Undated; accessed 2026-09-26", "Finding": "Microsoft/Databricks-based foundation unifies portfolio financial, operating and market data for AI.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Quantum names Microsoft and Databricks as platform partners and describes a shared foundation for AI.", "Context": "Quantum Energy Cloud unifies financial, operational and market data across the sponsor’s portfolio.", "Assessment": "A shared data foundation can make portfolio analytics repeatable. The undated firm page does not establish when deployment began, how many companies participate, or whether a new initiative launched within this review window."}, {"ID": "OCT9-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "Date": "2026", "Finding": "Quantum Energy Cloud connects portfolio financial, operational and market data using Microsoft and Databricks. The 2026 report identifies Claude Enterprise as the firm interface succeeding QIQ.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "WATCH21PM-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum adds a relational layer to its deal-evaluation data", "Date": "Unknown", "Finding": "Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.", "Metric": "Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.", "Source_IDs": "WATCH21PM-DBX", "Source_URL": "https://www.databricks.com/customers/quantum-capital-group", "Evidence": "Original vendor customer story reviewed; date unknown; data foundation, not proof of AI outcomes", "Origin": "21 October evening Luna research; editor original-source review", "Assessment": "No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution."}], "Latest_disclosure": "2025-05-02", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch21pm-ai-052", "software_capital-003"], "Research_search": "{\"packet_id\": \"watch21pm-ai-052\", \"key\": \"watch21pm-ai-052\", \"record_ids\": [\"AI-052\"], \"company\": \"Quantum Capital Group\", \"title\": \"Quantum adds a relational layer to its deal-evaluation data\", \"category\": \"Business transformation foundation\", \"department\": \"Finance, investment & strategy\", \"report_date\": \"2026-10-21\", \"verification\": \"verified partner\", \"stage\": \"Vendor describes an adopted sponsor-level data layer\", \"business_problem\": \"Make geological, production and financial assumptions consistent across energy deal evaluation.\", \"workflow\": \"Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.\", \"technology_data\": \"Lakebase relational schemas, Postgres and PostGIS; Unity Catalog access controls and lineage; Lakeflow change-data capture; familiar SQL, Tableau and Spotfire tools.\", \"integration_human_role\": \"Engineers, analysts and deal teams query data. The source does not establish investment-approval rules, autonomous capital allocation or AI-agent execution rights.\", \"deployment_scope\": \"Quantum sponsor-level energy deal evaluation; asset and portfolio-company populations undisclosed. Undated source; adoption and evaluation periods unknown. Extends existing sponsor platform AI-052, separately from QB Energy’s field-planning workflow AI-054.\", \"economics\": \"Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.\", \"practical_relevance\": \"Can the team trace a deal assumption to its source and show whether shared data reduces reconciliation and review work?\", \"assessment\": \"Shared data can reduce inconsistent assumptions without proving AI value or investment performance. Test reconciliation, traceability and full review effort.\", \"limitations\": [\"No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution.\"], \"source_ids\": [\"WATCH21PM-DBX\"], \"partners\": [\"Databricks\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"Unknown\", \"event\": \"Vendor describes adopted relational data layer; publication and adoption dates unknown.\", \"source_ids\": [\"WATCH21PM-DBX\"]}], \"sources\": [{\"id\": \"WATCH21PM-DBX\", \"title\": \"Transforming energy deal evaluation with Lakebase\", \"url\": \"https://www.databricks.com/customers/quantum-capital-group\", \"date\": \"Unknown\", \"publisher\": \"Databricks\", \"locator\": \"Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs\", \"excerpt\": \"The firm now runs over 11,000 unique data queries per month\", \"access\": \"Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.\", \"type\": \"Vendor customer story\", \"review_status\": \"Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.\"}], \"field_sources\": {\"business_problem\": [\"WATCH21PM-DBX\"], \"workflow\": [\"WATCH21PM-DBX\"], \"technology_data\": [\"WATCH21PM-DBX\"], \"integration_human_role\": [\"WATCH21PM-DBX\"], \"deployment_scope\": [\"WATCH21PM-DBX\"], \"economics\": [\"WATCH21PM-DBX\"]}}", "Research_status": "Existing sponsor platform enriched 21 October 2026 with undated vendor data-foundation detail; no new outcome credit.", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "More than 100 portfolio AI accelerators represent in-window deployments.", "reason": "Official 2024 ESG PDF excerpt puts operating coverage at May 2024, before window; publication date in May 2025 does not change operational date. PDF could not be opened in full.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-052", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The accelerator count is historical, not a new 2026 deployment result. The 2026 report updates the firm platform; per-company coverage and project costs remain undisclosed. Lakebase limits: No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Current_source_check": "2026-10-21", "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-053", "Company": "Quantum Capital Group", "Segment": "Energy PE", "Theme": "Enterprise AI", "Initiative": "Claude Enterprise succeeds Quantum IQ", "Extracted_finding": "The 2026 sponsor report identifies Claude Enterprise as the current firm generative AI platform, succeeding Quantum IQ.", "Stage_at_disclosure": "Firm adoption described; portfolio-company deployment and measured usage not supplied.", "Reported_metric": "Not disclosed", "Interpretation_limit": "The undated digital webpage still describes QIQ. No usage denominator or measured return is disclosed; this is not proof of portfolio-wide Claude adoption.", "Source_IDs": "S31; OCT9-QCG-2026", "Source_date": "2026", "Source_type": "Sponsor stakeholder report", "Source_locator": "2026 Stakeholder Report page 12: Enterprise AI at Quantum", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence_status": "Primary source text reviewed", "PE_relationship": "Manager internal operations", "Coverage_bucket": "Requested source class", "Time_scope": "Undated current disclosure", "Context": "Quantum IQ is the sponsor’s proprietary generative AI platform for shared knowledge, analysis and workflow assistance.", "Implementation": "Quantum’s 2026 report says Claude Enterprise has replaced its earlier QIQ platform. Earlier QIQ disclosures remain in the history.", "Assessment": "This is a firm-described operating capability. Public disclosure does not supply usage, cost or independently tested output quality. The source is undated, so it is included as current context with a timing exception.", "Detail_words": 62, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Finance, investment & strategy", "Production & field operations"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Private-equity operating platform", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Quantum Capital Group", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-053", "Initiative_ID": "AI-053", "Title": "Quantum IQ", "Date": "Undated; accessed 2026-09-26", "Finding": "Proprietary GenAI platform answers questions, automates analysis and triggers workflows.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Quantum describes teams using QIQ to answer questions, fine-tune models and standardize practices.", "Context": "Quantum IQ is the sponsor’s proprietary generative AI platform for shared knowledge, analysis and workflow assistance.", "Assessment": "This is a firm-described operating capability. Public disclosure does not supply usage, cost or independently tested output quality. The source is undated, so it is included as current context with a timing exception."}, {"ID": "OCT9-AI-053", "Initiative_ID": "AI-053", "Title": "Claude Enterprise succeeds Quantum IQ", "Date": "2026", "Finding": "The 2026 sponsor report identifies Claude Enterprise as the current firm generative AI platform, succeeding Quantum IQ.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}], "Latest_disclosure": "Undated; accessed 2026-09-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "More than 100 portfolio AI accelerators represent in-window deployments.", "reason": "Official 2024 ESG PDF excerpt puts operating coverage at May 2024, before window; publication date in May 2025 does not change operational date. PDF could not be opened in full.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-053", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Current_source_check": "2026-10-09", "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-054", "Company": "Quantum Capital Group", "Segment": "Energy PE", "Theme": "Investment AI", "Initiative": "QB Energy Automated Pad Development with Quantum", "Extracted_finding": "A Quantum portfolio company in the Piceance Basin uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Reported_metric": "Microsoft reports a 20,000-acre scenario took 20 minutes versus three weeks; leasehold utilization rose from 50% to 85%. Production runs evaluated 2.3m pad candidates, 3.5m well possibilities and over 50m allocations; the accumulation period is unstated. Updated-version comparison: 50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.", "Interpretation_limit": "Microsoft comparisons remain vendor-reported. Quantum names QB Energy and reports 12 areas over nine months. The ~$10bn development program informed is not savings or value created. Engineers review plans; no autonomous capital approval is established. Solver attribution: Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model.", "Source_IDs": "S31; OCT5-S032; OCT9-QCG-2026; OCT5-S028", "Source_date": "2026-04-30", "Source_type": "Microsoft customer story", "Source_locator": "Field development planning solution; Improved speed, scale, and precision", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence_status": "Original source text reviewed on 6 October 2026", "PE_relationship": "Investment and portfolio operations", "Coverage_bucket": "Requested source class", "Time_scope": "Undated current disclosure", "Context": "The 2026 Quantum report names QB Energy behind the Piceance planning case. Earlier Microsoft reporting left the operator unnamed.", "Implementation": "QB Energy, identified in Quantum’s 2026 report, uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.", "Assessment": "Optimization explores feasible development choices under assumptions. It should not be interpreted as delegated investment authority, proven investment alpha or a substitute for reserves and land diligence. The public description is undated.", "Detail_words": 56, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["Reservoir engineering & field development", "Drilling & completions"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Private-equity operating platform", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Quantum Capital Group", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-054", "Initiative_ID": "AI-054", "Title": "Asset and capital planning", "Date": "Undated; accessed 2026-09-26", "Finding": "Models support single-well economics, scenarios, acreage planning and well/pad combinations.", "Metric": "Millions of combinations evaluated", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Single-well economics, geospatial analysis and constraint-based optimization evaluate millions of well and pad combinations.", "Context": "Quantum describes scenario models for asset acquisition and capital deployment.", "Assessment": "Optimization explores feasible development choices under assumptions. It should not be interpreted as delegated investment authority, proven investment alpha or a substitute for reserves and land diligence. The public description is undated."}, {"ID": "OCT5-AI-054-1", "Initiative_ID": "AI-054", "Title": "Asset and capital planning", "Date": "2026-04-30", "Finding": "A Quantum portfolio company in the Piceance Basin uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.", "Metric": "Microsoft reports a 20,000-acre scenario took 20 minutes versus three weeks; leasehold utilization rose from 50% to 85%. Production runs evaluated 2.3m pad candidates, 3.5m well possibilities and over 50m allocations; the accumulation period is unstated.", "Source_IDs": "OCT5-S032", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT9-AI-054", "Initiative_ID": "AI-054", "Title": "QB Energy Automated Pad Development with Quantum", "Date": "2026", "Finding": "The 2026 Quantum report names QB Energy behind the Piceance planning case. Earlier Microsoft reporting left the operator unnamed.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "WATCH21AM-AI-054", "Initiative_ID": "AI-054", "Title": "Quantum’s planning gain includes a separate algorithm improvement", "Date": "2026-04-30", "Finding": "Microsoft says algorithm changes in an updated version of Quantum’s field-planning tool reduced solve time by 50–65%. The integer-linear-programming optimizer selects feasible pads and wells. A Copilot Studio agent explains outputs. The numerical improvement belongs to the optimization algorithm; the source does not attribute it to the language model.", "Metric": "50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.", "Source_IDs": "OCT5-S032", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence": "Accessible original reviewed; issuer/vendor claims not independently validated", "Origin": "21 October morning Luna research; editor source and attribution review", "Assessment": "Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model."}], "Latest_disclosure": "Undated; accessed 2026-09-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete · Supplemented 6 October 2026", "Research_packets": ["oct6-gap-ai-054"], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "More than 100 portfolio AI accelerators represent in-window deployments.", "reason": "Official 2024 ESG PDF excerpt puts operating coverage at May 2024, before window; publication date in May 2025 does not change operational date. PDF could not be opened in full.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-054", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Research_verification": "verified partner", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Microsoft comparisons remain vendor-reported. Quantum names QB Energy and reports 12 areas over nine months. The ~$10bn development program informed is not savings or value created. Engineers review plans; no autonomous capital approval is established. Solver attribution: Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Research_search": "{\"packet_id\": \"oct6-gap-ai-054\", \"review_date\": \"2026-10-06\", \"title\": \"QB Energy geospatial field planning with Quantum\", \"workflow\": \"QB Energy, identified in Quantum’s 2026 report, uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Vendor/customer account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"Microsoft comparisons remain vendor-reported. Quantum names QB Energy and reports 12 areas over nine months. The ~$10bn development program informed is not savings or value created. Engineers review plans; no autonomous capital approval is established. Solver attribution: Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model.\"], \"verification\": \"verified partner\", \"source_review\": \"Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation.\", \"source_locator\": \"Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph\", \"timeline_date\": \"2026-04-30\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-054\", \"record_ids\": [\"AI-054\"], \"company\": \"Quantum Capital Group\", \"category\": \"Internal adoption\", \"department\": \"Reservoir engineering & field development\", \"source_ids\": [\"OCT5-S032\", \"OCT9-QCG-2026\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S032\"], \"workflow\": [\"OCT5-S032\", \"OCT9-QCG-2026\"], \"technology_data\": [\"OCT5-S032\"], \"deployment_scope\": [\"OCT5-S032\"], \"practical_relevance\": [\"OCT5-S032\"], \"economics\": [\"OCT5-S032\"], \"integration_human_role\": [\"OCT5-S032\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"Microsoft reports a 20,000-acre scenario took 20 minutes versus three weeks; leasehold utilization rose from 50% to 85%. Production runs evaluated 2.3m pad candidates, 3.5m well possibilities and over 50m allocations; the accumulation period is unstated.\", \"period\": \"2026-04-30\", \"attribution\": \"Vendor/customer account\", \"source_ids\": [\"OCT5-S032\"]}, {\"metric\": \"Updated solver-version comparison\", \"value\": \"50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.\", \"period\": \"Updated version; evaluation period unknown\", \"attribution\": \"Microsoft customer story; algorithm, not LLM result\", \"source_ids\": [\"OCT5-S032\"]}], \"timeline\": [{\"date\": \"2026-04-30\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"OCT5-S032\"]}], \"sources\": [{\"id\": \"OCT5-S032\", \"title\": \"Field development planning at Quantum Capital Group\", \"url\": \"https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service\", \"date\": \"2026-04-30\", \"type\": \"Microsoft customer story\", \"publisher\": \"Microsoft\", \"locator\": \"Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph\", \"review_status\": \"Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation.\"}, {\"id\": \"OCT9-QCG-2026\", \"title\": \"Quantum 2026 Stakeholder Report\", \"url\": \"https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf\", \"date\": \"2026\", \"type\": \"Sponsor stakeholder report\", \"publisher\": \"Quantum Capital Group\", \"locator\": \"Pages 11–15; page 68 disclaimer\", \"review_status\": \"Relevant primary sections reviewed\"}]}", "Operating_company": "QB Energy", "Current_source_check": "2026-10-21", "Evidence_review": {"source_read": "verified partner", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-055", "Company": "Quantum Capital Group", "Segment": "Energy PE", "Theme": "Operational AI", "Initiative": "Production optimization and surveillance", "Extracted_finding": "Models recommend completion designs, optimize lift and predict production anomalies and downtime.", "Stage_at_disclosure": "Firm-described use", "Reported_metric": "Not disclosed", "Interpretation_limit": "No named portfolio-company deployment metrics.", "Source_IDs": "S31", "Source_date": "Undated; accessed 2026-09-26", "Source_type": "PE operating-strategy disclosure", "Source_locator": "Quantum Energy Cloud; Quantum IQ; AI at Work", "Source_URL": "https://www.quantumcap.com/digital-mission/", "Evidence_status": "Primary source text reviewed", "PE_relationship": "Portfolio operations", "Coverage_bucket": "Requested source class", "Time_scope": "Undated current disclosure", "Context": "Quantum applies models to production optimization and surveillance across the investment and operating process.", "Implementation": "The firm describes completion recommendations, lift optimization, anomaly detection and downtime prediction.", "Assessment": "These functions could inform portfolio-company operations, but the page names no specific company, field or quantified outcome. The source is undated and should not be treated as a newly announced deployment in this window.", "Detail_words": 60, "Primary_department": "Production & field operations", "Secondary_departments": ["Drilling & completions", "Maintenance & reliability"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Private-equity operating platform", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Quantum Capital Group", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-055", "Initiative_ID": "AI-055", "Title": "Production optimization and surveillance", "Date": "Undated; accessed 2026-09-26", "Finding": "Models recommend completion designs, optimize lift and predict production anomalies and downtime.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The firm describes completion recommendations, lift optimization, anomaly detection and downtime prediction.", "Context": "Quantum applies models to production optimization and surveillance across the investment and operating process.", "Assessment": "These functions could inform portfolio-company operations, but the page names no specific company, field or quantified outcome. The source is undated and should not be treated as a newly announced deployment in this window."}], "Latest_disclosure": "Undated; accessed 2026-09-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "More than 100 portfolio AI accelerators represent in-window deployments.", "reason": "Official 2024 ESG PDF excerpt puts operating coverage at May 2024, before window; publication date in May 2025 does not change operational date. PDF could not be opened in full.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-055", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-056", "Company": "Quorum / Streamba", "Segment": "Energy software", "Theme": "AI products", "Initiative": "Streamba acquisition and VOR energy-logistics integration", "Extracted_finding": "Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.", "Stage_at_disclosure": "Acquired 2026-06-25; integration/agentic roadmap described", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No named live customer case, measured results or technical architecture detail.", "Source_IDs": "S32; S33", "Source_date": "2026-06-25", "Source_type": "company announcement", "Source_locator": "Streamba/VOR and DaWinci", "Source_URL": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "Evidence_status": "verified primary", "PE_relationship": "Francisco Partners lists Quorum as current buyout investment", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-06-25; see chronology for operating dates.", "Context": "Offshore operators manage cargo, personnel, vessels and changing plans across disconnected logistics tools; disruptions force manual cross-team replanning.", "Implementation": "Quorum added Streamba’s Visibility, Orchestration and Response product to DaWinci logistics for combined personnel/cargo planning and execution. VOR centralizes data and surfaces disruptions; Quorum describes agentic systems that can suggest/perform reassignment or plan updates, with human oversight.", "Assessment": "Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.", "Detail_words": 68, "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Quorum", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-056", "Initiative_ID": "AI-056", "Title": "AI logistics acquisition", "Date": "2026-06-25", "Finding": "Acquired VOR intelligence layer to connect planning, logistics and execution; agentic response is part of roadmap.", "Metric": "Terms not disclosed", "Source_IDs": "S32; S33", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Streamba’s Visibility, Orchestration and Response layer combines with DaWinci to cover personnel and cargo. The roadmap includes agents that respond to disruptions, reassign resources and revise plans under human oversight.", "Context": "Francisco Partners-backed Quorum acquired Streamba in June 2026 to connect logistics planning with field execution.", "Assessment": "The acquisition is disclosed as completed. Fully autonomous response remains a product-development ambition. Deal terms, integration timing and customer-level savings are not disclosed."}], "Latest_disclosure": "2026-06-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-007"], "Research_search": " {\"key\": \"quorum-streamba\", \"existing_ids\": [\"AI-056\"], \"related_existing_ids\": [], \"company\": \"Quorum Software\", \"title\": \"Streamba acquisition and VOR energy-logistics integration\", \"department\": \"Supply chain & logistics\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Acquired 2026-06-25; integration/agentic roadmap described\", \"practical_relevance\": \"Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.\", \"business_problem\": \"Offshore operators manage cargo, personnel, vessels and changing plans across disconnected logistics tools; disruptions force manual cross-team replanning.\", \"workflow\": \"Quorum added Streamba’s Visibility, Orchestration and Response product to DaWinci logistics for combined personnel/cargo planning and execution. VOR centralizes data and surfaces disruptions; Quorum describes agentic systems that can suggest/perform reassignment or plan updates, with human oversight.\", \"technology_data\": \"AI-native workflow orchestration is described, but release provides little model, training data or connectivity detail.\", \"integration_human_role\": \"Human oversight is explicit for autonomous response; detail on permissions and deployment gates not disclosed.\", \"deployment_scope\": \"Quorum customer product portfolio; this release does not name a customer rollout or specify installed base integration completion.\", \"partners\": [\"Streamba\"], \"disclosed_results\": [], \"economics\": \"Terms undisclosed; no operational ROI quantified.\", \"timeline\": [{\"date\": \"2026-06-25\", \"event\": \"Quorum announced Streamba acquisition to extend energy value-chain AI platform.\", \"source_ids\": [\"S32\"]}], \"limitations\": [\"No named live customer case, measured results or technical architecture detail.\"], \"assessment\": \"Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.\", \"sources\": [{\"id\": \"S32\", \"title\": \"Quorum Software acquires Streamba\", \"url\": \"https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/\", \"publisher\": \"Quorum Software\", \"date\": \"2026-06-25\", \"type\": \"company announcement\", \"locator\": \"Streamba/VOR and DaWinci\", \"review_status\": \"full text reviewed\", \"supports\": [\"Streamba AI-native supply chain platform and Visibility, Orchestration and Response system unify logistics data, workflows and teams.\", \"Quorum says agentic capabilities can identify disruption and propose or execute plan changes with human oversight.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-009\"}], \"field_sources\": {\"business_problem\": [\"S32\"], \"workflow\": [\"S32\"], \"technology_data\": [\"S32\"], \"integration_human_role\": [\"S32\"], \"deployment_scope\": [\"S32\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-007\", \"source_ids\": [\"S32\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-056\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-056", "company": "Quorum / Streamba", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "https://www.franciscopartners.com/investments/quorum-software"], "findings": [{"severity": "low", "field": "classification / deployment", "issue": "Acquisitions and planned integration enlarge a vendor portfolio but do not alone establish an integrated AI workflow in production.", "correction": "Describe deal/integration stage explicitly; avoid implying realized customer outcomes.", "source_url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-057", "Company": "PakEnergy", "Segment": "Energy software", "Theme": "AI products", "Initiative": "Embedded AI lease provision processing", "Extracted_finding": "Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.", "Stage_at_disclosure": "Launched 2025-09-25", "Reported_metric": "Lease provision task time: Hours to seconds (vendor claim) (Vendor-reported)", "Interpretation_limit": "No model or accuracy disclosure; vendor speed claim only.", "Source_IDs": "S34; S36", "Source_date": "2025-09-25", "Source_type": "company announcement", "Source_locator": "Land Management lease provision", "Source_URL": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "Charlesbank-backed portfolio company", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-09-25; see chronology for operating dates.", "Context": "Land teams process lease provisions and related research manually, potentially missing key dates/terms or delaying workflows.", "Implementation": "AI feature is embedded in PakEnergy Land Management to read/process lease provision work as part of existing land workflow.", "Assessment": "Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.", "Detail_words": 73, "Primary_department": "Land & legal", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "PakEnergy", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-057", "Initiative_ID": "AI-057", "Title": "Lease provision extraction", "Date": "2025-09-25", "Finding": "AI embedded in land management accelerates lease research and provision processing.", "Metric": "Vendor claim: hours to seconds", "Source_IDs": "S34; S36", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The function processes lease information inside the existing system. PakEnergy claims work previously taking hours can take seconds and positions this as the first release in a wider AI roadmap.", "Context": "Charlesbank-backed PakEnergy embedded AI lease-provision extraction in its Land Management product in September 2025.", "Assessment": "The time claim comes from the vendor, without a disclosed accuracy benchmark or customer sample. Faster extraction does not remove the need to verify obligations in the underlying lease."}], "Latest_disclosure": "2025-09-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-008"], "Research_search": " {\"key\": \"pak-lease-ai\", \"existing_ids\": [\"AI-057\"], \"related_existing_ids\": [], \"company\": \"PakEnergy\", \"title\": \"Embedded AI lease provision processing\", \"department\": \"Land & legal\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched 2025-09-25\", \"practical_relevance\": \"Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.\", \"business_problem\": \"Land teams process lease provisions and related research manually, potentially missing key dates/terms or delaying workflows.\", \"workflow\": \"AI feature is embedded in PakEnergy Land Management to read/process lease provision work as part of existing land workflow.\", \"technology_data\": \"Announcement describes embedded AI but not model, source-document types, validation methods or integration specifics.\", \"integration_human_role\": \"Land staff remain users; no detailed review/exception process or accuracy rates given.\", \"deployment_scope\": \"Feature within PakEnergy Land Management; availability described by vendor, no customer/site count.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Lease provision task time\", \"value\": \"Hours to seconds (vendor claim)\", \"period\": \"Launch announcement, 2025-09-25\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"S34\"]}], \"economics\": \"PakEnergy claims task duration falls from hours to seconds; no independent timing or cost validation.\", \"timeline\": [{\"date\": \"2025-09-25\", \"event\": \"PakEnergy announced embedded AI lease-provision capability.\", \"source_ids\": [\"S34\"]}], \"limitations\": [\"No model or accuracy disclosure; vendor speed claim only.\"], \"assessment\": \"Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.\", \"sources\": [{\"id\": \"S34\", \"title\": \"PakEnergy launches embedded AI-powered lease provision capability\", \"url\": \"https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability\", \"publisher\": \"PakEnergy\", \"date\": \"2025-09-25\", \"type\": \"company announcement\", \"locator\": \"Land Management lease provision\", \"review_status\": \"full text reviewed\", \"supports\": [\"Embedded AI processes lease provision research and data inside PakEnergy Land Management.\", \"Company says work previously taking hours can be done in seconds; no accuracy or independent outcome data disclosed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-010\"}], \"field_sources\": {\"business_problem\": [\"S34\"], \"workflow\": [\"S34\"], \"technology_data\": [\"S34\"], \"integration_human_role\": [\"S34\"], \"deployment_scope\": [\"S34\"], \"economics\": [\"S34\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-008\", \"source_ids\": [\"S34\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-057\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-057", "company": "PakEnergy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Ownership source confirms Charlesbank as PakEnergy private-equity partner and documents Petrosight well-lifecycle software acquisition; it does not support AI in PakEnergy accounting workflows. This is corroborating context only, not a factual AI correction.", "correction": "", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-058", "Company": "PakEnergy", "Segment": "Energy software", "Theme": "AI products", "Initiative": "PakCAPTURE ticket and receipt extraction", "Extracted_finding": "Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.", "Stage_at_disclosure": "Launched 2026-05-20", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No independent accuracy or economic results.", "Source_IDs": "RS079; S35; S36", "Source_date": "2026-05-20", "Source_type": "company announcement", "Source_locator": "PakCAPTURE workflow", "Source_URL": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "Charlesbank-backed portfolio company", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-05-20; see chronology for operating dates.", "Context": "Field drivers and back-office teams rekey receipts/paper ticket details and need to associate documents with loads rapidly.", "Implementation": "Driver photographs a receipt/ticket; PakAI extracts key fields, populates pickup/drop forms and links the record to a transportation load for office access.", "Assessment": "Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.", "Detail_words": 62, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "PakEnergy", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-058", "Initiative_ID": "AI-058", "Title": "PakCAPTURE ticket and receipt processing", "Date": "2026-05-20", "Finding": "Photo capture and AI extraction automate transportation paperwork and back-office entry.", "Metric": "No quantified return disclosed", "Source_IDs": "S35; S36", "Evidence": "Company-issued release text reviewed", "Origin": "Original register", "Implementation": "Drivers or staff capture document images so information can enter transportation and back-office workflows with less manual re-entry. The May 2026 release adds another product to PakEnergy’s AI rollout.", "Context": "PakCAPTURE brings AI document extraction into PakEnergy Transportation for tickets and receipts.", "Assessment": "The launch establishes availability. The reviewed announcement does not provide quantified payback, extraction accuracy or a measured reduction in billing disputes."}], "Latest_disclosure": "2026-05-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-009"], "Research_search": " {\"key\": \"pak-capture\", \"existing_ids\": [\"AI-058\"], \"related_existing_ids\": [], \"company\": \"PakEnergy\", \"title\": \"PakCAPTURE ticket and receipt extraction\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched 2026-05-20\", \"practical_relevance\": \"Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.\", \"business_problem\": \"Field drivers and back-office teams rekey receipts/paper ticket details and need to associate documents with loads rapidly.\", \"workflow\": \"Driver photographs a receipt/ticket; PakAI extracts key fields, populates pickup/drop forms and links the record to a transportation load for office access.\", \"technology_data\": \"AI extraction embedded in PakTransportation; underlying model/OCR, confidence scoring and exception handling are not disclosed.\", \"integration_human_role\": \"Driver captures record; staff can use transportation system record; no explicit human validation details.\", \"deployment_scope\": \"Feature available to PakTransportation customers; no usage numbers disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No quantified labor saving, accuracy, or customer economics disclosed.\", \"timeline\": [{\"date\": \"2026-05-20\", \"event\": \"PakEnergy announced PakCAPTURE.\", \"source_ids\": [\"RS079\"]}], \"limitations\": [\"No independent accuracy or economic results.\"], \"assessment\": \"Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.\", \"sources\": [{\"id\": \"RS079\", \"title\": \"PakEnergy launches AI-powered Ticket-N-Receipt Management\", \"url\": \"https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true\", \"publisher\": \"PakEnergy\", \"date\": \"2026-05-20\", \"type\": \"company announcement\", \"locator\": \"PakCAPTURE workflow\", \"review_status\": \"full text reviewed\", \"supports\": [\"Drivers photograph receipts or paper tickets; PakAI extracts data, populates pickup/drop forms and associates information with a load.\", \"Feature embedded in PakTransportation and available to customers.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-011\"}], \"field_sources\": {\"business_problem\": [\"RS079\"], \"workflow\": [\"RS079\"], \"technology_data\": [\"RS079\"], \"integration_human_role\": [\"RS079\"], \"deployment_scope\": [\"RS079\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-009\", \"source_ids\": [\"RS079\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-058\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-058", "company": "PakEnergy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The reopened acquisition release is ownership/deal context only and does not substantiate the PakCAPTURE ticket/receipt extraction workflow; retain the existing AI workflow on its own cited evidence.", "correction": "", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-059", "Company": "ECP / KKR / CyrusOne / Calpine", "Segment": "PE / gas-fired power", "Theme": "AI infrastructure", "Initiative": "ECP/KKR hyperscale data-center campus co-located with Calpine power", "Extracted_finding": "ECP’s role is investment in energy-adjacent data-center infrastructure; co-location with generation addresses power availability for hyperscale compute.", "Stage_at_disclosure": "Announced development; expected operations Q4 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Announcement is company/parent primary and forward-looking; development and operating schedule may change. AI-demand exposure is separate from AI adoption by ECP.", "Source_IDs": "RS107; S37", "Source_date": "2025-07-30", "Source_type": "company announcement", "Source_locator": null, "Source_URL": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "ECP and KKR strategic partnership", "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2025-07-30; see chronology for operating dates.", "Context": "ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations.", "Implementation": "ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations", "Assessment": "Material infrastructure exposure tied to hyperscale/AI demand; maintain it as distinct investment category and do not present as ECP internal AI.", "Detail_words": 81, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ECP", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-059", "Initiative_ID": "AI-059", "Title": "Bosque data-center campus", "Date": "2025-07-30", "Finding": "Campus adjacent to gas generation uses a dedicated long-term power contract.", "Metric": "190 MW campus; initial 144 MW IT; investment approaching $4bn", "Source_IDs": "S37", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "CyrusOne and ECP are developing a 190 MW campus with initial IT capacity of 144 MW, next to Calpine gas generation. July 2025 disclosure put investment near $4 billion and targeted Q4 2026 operation.", "Context": "ECP and KKR announced the Bosque campus as the first project under their broader $50 billion infrastructure partnership.", "Assessment": "Construction was underway at announcement. The $50 billion figure refers to the broader partnership, while $4 billion refers to this project. Campus and IT MW describe different capacity boundaries."}], "Latest_disclosure": "2025-07-30", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-014"], "Research_search": " {\"key\": \"ecp-texas-hyperscale-campus\", \"existing_ids\": [\"AI-059\"], \"related_existing_ids\": [], \"company\": \"ECP\", \"title\": \"ECP/KKR hyperscale data-center campus co-located with Calpine power\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Announced development; expected operations Q4 2026\", \"practical_relevance\": \"ECP’s role is investment in energy-adjacent data-center infrastructure; co-location with generation addresses power availability for hyperscale compute.\", \"business_problem\": \"ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations.\", \"workflow\": \"ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations\", \"technology_data\": \"Data-center campus, grid/power supply and natural gas generation; no internal AI tooling disclosed.\", \"integration_human_role\": \"ECP is financial sponsor; CyrusOne develops/operates data center; Calpine supplies generation.\", \"deployment_scope\": \"Single campus under development; planned capacity and schedule are forward-looking.\", \"partners\": [\"KKR\", \"CyrusOne\", \"Calpine\"], \"disclosed_results\": [], \"economics\": \"Reported total campus investment approaching $4bn; not AI adoption savings or ECP operating economics.\", \"timeline\": [{\"date\": \"2025-07-30\", \"event\": \"Joint campus development announced.\", \"source_ids\": [\"RS107\"]}], \"limitations\": [\"Announcement is company/parent primary and forward-looking; development and operating schedule may change.\", \"AI-demand exposure is separate from AI adoption by ECP.\"], \"assessment\": \"Material infrastructure exposure tied to hyperscale/AI demand; maintain it as distinct investment category and do not present as ECP internal AI.\", \"sources\": [{\"id\": \"RS107\", \"title\": \"ECP and KKR announce Texas hyperscale data-center campus\", \"url\": \"https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas\", \"publisher\": \"Bridgepoint Group / ECP\", \"date\": \"2025-07-30\", \"type\": \"company announcement\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"ECP/KKR/CyrusOne 190MW campus co-located with Calpine generation; AI-demand infrastructure.\"], \"original_id\": \"ecp\"}], \"field_sources\": {\"business_problem\": [\"RS107\"], \"workflow\": [\"RS107\"], \"technology_data\": [\"RS107\"], \"integration_human_role\": [\"RS107\"], \"deployment_scope\": [\"RS107\"], \"economics\": [\"RS107\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-014\", \"source_ids\": [\"RS107\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-059\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-059", "company": "ECP / KKR / CyrusOne / Calpine", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas"], "findings": [{"severity": "medium", "field": "classification", "issue": "Campus provides power/data-center capacity serving AI demand, not an ECP/KKR oil-and-gas operational AI workflow.", "correction": "Keep solely under AI-demand infrastructure.", "source_url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-060", "Company": "bp / Palantir / Infosys", "Segment": "Integrated", "Theme": "Automation context", "Initiative": "Refining digital / AI program", "Extracted_finding": "Management reported progress applying Palantir and Infosys to refining workflows.", "Stage_at_disclosure": "Early progress at disclosure", "Reported_metric": "No AI-only benefit disclosed", "Interpretation_limit": "Call says digital; AI context comes from partner announcement.", "Source_IDs": "S38; S39", "Source_date": "2024-10-29", "Source_type": "Earnings call transcript", "Source_locator": "Murray Auchincloss response to Lydia Rainforth", "Source_URL": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "Evidence_status": "Third-party-hosted transcript reviewed", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Within confirmed 24-month window", "Context": "bp’s Q3 2024 call described progress applying Palantir and Infosys to refining workflows.", "Implementation": "The call discusses digital improvement. A separate Palantir announcement on 9 September 2024 establishes the AI relationship but falls just outside the review window.", "Assessment": "The AI linkage uses corroborating partner evidence rather than explicit AI detail in every call remark. This distinction prevents a general digital program from becoming an unsupported claim of specific AI savings.", "Detail_words": 69, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": ["Maintenance & reliability", "IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Automation or digital workflow; explicit AI basis not established", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "bp", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-060", "Initiative_ID": "AI-060", "Title": "Refining digital / AI program", "Date": "2024-10-29", "Finding": "Management reported progress applying Palantir and Infosys to refining workflows.", "Metric": "No AI-only benefit disclosed", "Source_IDs": "S38; S39", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The call discusses digital improvement. A separate Palantir announcement on 9 September 2024 establishes the AI relationship but falls just outside the review window.", "Context": "bp’s Q3 2024 call described progress applying Palantir and Infosys to refining workflows.", "Assessment": "The AI linkage uses corroborating partner evidence rather than explicit AI detail in every call remark. This distinction prevents a general digital program from becoming an unsupported claim of specific AI savings."}], "Latest_disclosure": "2024-10-29", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Excluded / unverified AI claim", "Audit_flags": [{"claim": "2024 Palantir and Infosys refining digital implementation is established AI workflow.", "reason": "The in-window Q3 2024 call establishes digital/ontology progress but names no AI workflow; the Palantir AI relationship announcement is pre-window. Exclude from confirmed AI adoption absent a qualifying AI-specific source.", "source_url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter"}], "Research_verification": "unable to verify", "Audit_review": {"id": "AI-060", "company": "bp / Palantir / Infosys", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "https://palantir2020ipo.q4web.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/default.aspx"], "findings": [{"severity": "high", "field": "Research_category / AI basis", "issue": "The in-window Q3 2024 bp call describes digital progress with Palantir and Infosys at the refining system and says the ontology was built; it does not establish an AI workflow. Palantir’s AI-capabilities relationship announcement is dated 2024-09-09, before the window.", "correction": "Classify as excluded/unverified AI adoption (digital transformation only) unless a later AI-specific operational disclosure is found.", "source_url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter"}], "residual_gaps": ["No qualifying in-window source in this assignment establishes AI use in bp refining. Existing profile already flags the issue; classification patch makes it explicit."], "partition": 1, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "unable to verify", "claim_status": "Third-party-hosted transcript reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-061", "Company": "Cenovus", "Segment": "Integrated", "Theme": "Enterprise AI", "Initiative": "Governed productivity and decision-support pilots", "Extracted_finding": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "Stage_at_disclosure": "Governed pilot program; individual use cases not described", "Reported_metric": "Not disclosed", "Interpretation_limit": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.", "Source_IDs": "S40", "Source_date": "2026-02-18", "Source_type": "annual report", "Source_locator": "Printed MD&A page 47: Technology, Information Systems and Data Privacy; page 2: document date", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "Evidence_status": "Original filing text reviewed 17 October 2026", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.", "Context": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "Implementation": "It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.", "Assessment": "A verified program disclosure helps establish intent and pilot stage. It does not establish a named operating deployment.", "Detail_words": 61, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Enabling capability or strategic investment", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Cenovus", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-061", "Initiative_ID": "AI-061", "Title": "Productivity and decision-support pilots", "Date": "2026-02-18", "Finding": "Describes governed AI introduction beginning with productivity and decision-support use cases.", "Metric": "Not disclosed", "Source_IDs": "S40", "Evidence": "Original filing reverified 17 October 2026; program-level pilots only", "Origin": "Original register", "Implementation": "The 2025 Management’s Discussion and Analysis supplies the program-level disclosure, without naming individual applications or external vendors.", "Context": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "Assessment": "This supports the existence of a pilot program. More detailed implementation, operating scope and value claims would require additional disclosure. A generic program should not be expanded into invented use cases."}, {"ID": "WATCH17AM-AI-061", "Initiative_ID": "AI-061", "Title": "Governed productivity and decision-support pilots", "Date": "2026-02-18", "Finding": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots. It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.", "Metric": "No measured outcome, cost or ROI disclosed.", "Source_IDs": "S40", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "Evidence": "Original operator source read; results are not independently validated", "Origin": "17 October morning source review", "Assessment": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established."}], "Latest_disclosure": "2026-02-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch17am-ai-061", "north_america-011"], "Research_search": "{\"packet_id\": \"watch17am-ai-061\", \"key\": \"watch17am-ai-061\", \"record_ids\": [\"AI-061\"], \"company\": \"Cenovus\", \"title\": \"Governed productivity and decision-support pilots\", \"category\": \"Strategy only\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-17\", \"verification\": \"verified primary\", \"stage\": \"Governed pilot program; individual use cases not described\", \"business_problem\": \"Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.\", \"workflow\": \"It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.\", \"technology_data\": \"It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.\", \"deployment_scope\": \"Program-level pilots described in the 2025 MD&A dated 18 February 2026; no facility, user count or go-live date supplied.\", \"integration_human_role\": \"Governed introduction is stated; specific human review and approval boundaries are not disclosed.\", \"economics\": \"No measured outcome, cost or ROI disclosed.\", \"practical_relevance\": \"This resolves whether the company disclosed AI pilots. It leaves open which business tasks reached regular use.\", \"assessment\": \"This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.\", \"limitations\": [\"This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.\"], \"source_ids\": [\"S40\"], \"sources\": [{\"id\": \"S40\", \"title\": \"2025 Management’s Discussion and Analysis\", \"url\": \"https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm\", \"date\": \"2026-02-18\", \"publisher\": \"Cenovus Energy / SEC\", \"locator\": \"Printed MD&A page 47: Technology, Information Systems and Data Privacy; page 2: document date\", \"review_status\": \"Original source read by editor\", \"supports\": [\"Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.\", \"No measured outcome, cost or ROI disclosed.\"], \"accessed\": \"2026-10-17\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-02-18\", \"event\": \"Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.\", \"source_ids\": [\"S40\"]}], \"field_sources\": {\"business_problem\": [\"S40\"], \"workflow\": [\"S40\"], \"technology_data\": [\"S40\"], \"deployment_scope\": [\"S40\"], \"integration_human_role\": [\"S40\"], \"economics\": [\"S40\"]}}", "Research_status": "Source reverified 17 October 2026 · program-level pilots", "Research_category": "Strategy only", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-061", "company": "Cenovus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Cenovus MD&A describes governed productivity pilots; no specific measured value is attributed exclusively to AI.", "correction": "Retain as pilot portfolio and avoid treating broader productivity/automation results as AI-specific.", "source_url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm"}], "residual_gaps": ["Named use cases, scope and outcomes are not disclosed."], "partition": 2, "reviewer_outcome": "partially supported", "source_recheck_on": "2026-10-17"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Program-level pilot disclosure verified; individual use cases not named", "independent_validation": "Not established by this collection"}, "Earlier_source_review": {"as_of": "2026-09-27", "verification": "unable to verify", "flags": [{"claim": "Productivity and decision-support pilots.", "reason": "Cited MD&A text was not reopened; current evidence is insufficient to verify claim. Not disproved.", "source_url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm"}], "note": "Historical source-access limit resolved by original filing read on 17 October 2026."}, "Current_source_check": "2026-10-17", "Research_origin": ["research/watch-2026-10-17-am/README.md"]}, {"ID": "AI-062", "Company": "Energy Transfer / CloudBurst", "Segment": "Midstream", "Theme": "AI infrastructure", "Initiative": "CloudBurst San Marcos natural-gas supply for AI campus", "Extracted_finding": "Energy Transfer supplies fuel infrastructure to a proposed behind-the-meter power campus serving AI/data-center load; this is an AI demand exposure, not ET’s internal AI adoption.", "Stage_at_disclosure": "Agreement signed; subject to customer final investment decision; projected operation not confirmation", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Infrastructure that powers AI demand; category is deliberately separate from operating AI. Agreement is conditional on customer FID; announced volumes and projected date are not actual deliveries.", "Source_IDs": "S41", "Source_date": "2025-02-10", "Source_type": "company announcement", "Source_locator": null, "Source_URL": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-02-10; see chronology for operating dates.", "Context": "Company agreed to supply up to 450,000 MMBtu/day firm gas to a campus outside San Marcos, Texas, supporting up to approximately 1.2 GW of direct behind-the-meter power, subject to CloudBurst FID.", "Implementation": "Energy Transfer agreed to supply up to 450,000 MMBtu/day of firm gas to a proposed campus outside San Marcos, Texas, supporting up to approximately 1.2 GW of behind-the-meter power, subject to CloudBurst reaching a final investment decision with its customer.", "Assessment": "A material gas-demand contract with direct data-center linkage, but it is conventional energy infrastructure exposure rather than internal AI adoption.", "Detail_words": 77, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Energy Transfer", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-062", "Initiative_ID": "AI-062", "Title": "San Marcos gas supply", "Date": "2025-02-10", "Finding": "Long-term gas agreement for an AI-focused campus via Oasis Pipeline.", "Metric": "Up to 450,000 MMBtu/day", "Source_IDs": "S41", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The February 2025 announcement specifies up to 450,000 MMBtu/day, sufficient for roughly 1.2 GW of direct power, for at least ten years. Supply depends on CloudBurst reaching customer FID.", "Context": "Energy Transfer signed a long-term agreement to supply CloudBurst’s proposed AI-focused campus near San Marcos through Oasis Pipeline.", "Assessment": "Q3 2026 operation was an original conditional target, not a verified operating date in this review. Contract capacity is not actual gas consumption, and the source gives no realized revenue."}], "Latest_disclosure": "2025-02-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-010"], "Research_search": " {\"key\": \"et-cloudburst-ai-campus\", \"existing_ids\": [\"AI-062\"], \"related_existing_ids\": [], \"company\": \"Energy Transfer\", \"title\": \"CloudBurst San Marcos natural-gas supply for AI campus\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Agreement signed; subject to customer final investment decision; projected operation not confirmation\", \"practical_relevance\": \"Energy Transfer supplies fuel infrastructure to a proposed behind-the-meter power campus serving AI/data-center load; this is an AI demand exposure, not ET’s internal AI adoption.\", \"business_problem\": \"Company agreed to supply up to 450,000 MMBtu/day firm gas to a campus outside San Marcos, Texas, supporting up to approximately 1.2 GW of direct behind-the-meter power, subject to CloudBurst FID.\", \"workflow\": \"Company agreed to supply up to 450,000 MMBtu/day firm gas to a campus outside San Marcos, Texas, supporting up to approximately 1\", \"technology_data\": \"Conventional pipeline gas supply and power generation arrangements; no AI technology is being adopted by Energy Transfer.\", \"integration_human_role\": \"CloudBurst/data-center operator and power-generation parties would use the delivered gas; ET’s disclosed role is supply.\", \"deployment_scope\": \"A single proposed campus; long-term term and projected start are in company announcement; operations depend on FID.\", \"partners\": [\"CloudBurst\"], \"disclosed_results\": [], \"economics\": \"No AI-related returns disclosed; gas volumes/contract economics not publicly quantified beyond delivery ceiling.\", \"timeline\": [{\"date\": \"2025-02-10\", \"event\": \"Firm supply agreement announced for proposed campus.\", \"source_ids\": [\"S41\"]}], \"limitations\": [\"Infrastructure that powers AI demand; category is deliberately separate from operating AI.\", \"Agreement is conditional on customer FID; announced volumes and projected date are not actual deliveries.\"], \"assessment\": \"A material gas-demand contract with direct data-center linkage, but it is conventional energy infrastructure exposure rather than internal AI adoption.\", \"sources\": [{\"id\": \"S41\", \"title\": \"Energy Transfer and CloudBurst sign natural gas supply agreement\", \"url\": \"https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply\", \"publisher\": \"Energy Transfer\", \"date\": \"2025-02-10\", \"type\": \"company announcement\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Up to 450,000 MMBtu/day firm gas supply for campus near San Marcos, supporting up to ~1.2 GW behind-the-meter power, contingent on customer FID; AI-demand infrastructure.\"], \"original_id\": \"etcloud\"}], \"field_sources\": {\"business_problem\": [\"S41\"], \"workflow\": [\"S41\"], \"technology_data\": [\"S41\"], \"integration_human_role\": [\"S41\"], \"deployment_scope\": [\"S41\"], \"economics\": [\"S41\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Correct category is AI-demand infrastructure; no operating AI technology workflow.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-010\", \"source_ids\": [\"S41\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-062\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-062", "company": "Energy Transfer / CloudBurst", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-063", "Company": "EQT / Homer City Redevelopment", "Segment": "E&P / gas-fired power", "Theme": "AI infrastructure", "Initiative": "Homer City gas supply for AI data-centre power demand", "Extracted_finding": "Useful exposure to AI electricity demand, but it should not be counted as EQT adopting AI in its own operations.", "Stage_at_disclosure": "Not AI. EQT/Homer City agreement in principle is for EQT to source and supply natural gas to a planned 4.4-GW power facility for a data-center campus. It is an energy-supply relationship serving AI demand, not an AI workflow operated by EQT.", "Reported_metric": "4.4 GW gas facility; power targeted 2027", "Interpretation_limit": "Agreement in principle is not proof of final contract, construction completion or first power. This belongs in AI-demand infrastructure.", "Source_IDs": "S42", "Source_date": "2025-07", "Source_type": "company announcement", "Source_locator": "Agreement in principle", "Source_URL": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "Evidence_status": "Primary source text reviewed", "PE_relationship": "Knighthead leads project financing and holds equity in the project.", "Coverage_bucket": "Supplementary primary disclosure", "Time_scope": "Dated source: 2025-07; see chronology for operating dates.", "Context": "EQT reached an agreement in principle to source and supply gas for the 4.4 GW Homer City power facility serving an AI/HPC campus.", "Implementation": "The July 2025 release names Knighthead as financing sponsor, GE Vernova as turbine supplier and Kiewit as construction lead. It targeted first power in 2027.", "Assessment": "An agreement in principle differs from completed definitive supply terms. The 4.4 GW figure concerns generation capacity. Gas volumes, pricing and realized supply economics remain undisclosed here.", "Detail_words": 75, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "EQT", "Review_tier": "Reference pending review", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-063", "Initiative_ID": "AI-063", "Title": "Homer City gas supply", "Date": "2025-07-15", "Finding": "EQT selected in principle as exclusive gas sourcing and supply partner for AI/HPC campus power.", "Metric": "4.4 GW gas facility; power targeted 2027", "Source_IDs": "S42", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2025 release names Knighthead as financing sponsor, GE Vernova as turbine supplier and Kiewit as construction lead. It targeted first power in 2027.", "Context": "EQT reached an agreement in principle to source and supply gas for the 4.4 GW Homer City power facility serving an AI/HPC campus.", "Assessment": "An agreement in principle differs from completed definitive supply terms. The 4.4 GW figure concerns generation capacity. Gas volumes, pricing and realized supply economics remain undisclosed here."}], "Latest_disclosure": "2025-07", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["north_america-010"], "Research_search": " {\"key\": \"eqt-homer-city-ai-load\", \"existing_ids\": [\"AI-063\"], \"company\": \"EQT\", \"title\": \"Homer City gas supply for AI data-centre power demand\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Agreement in principle; proposed power project\", \"practical_relevance\": \"Useful exposure to AI electricity demand, but it should not be counted as EQT adopting AI in its own operations.\", \"business_problem\": \"Homer City redevelopment requires long-term gas supply to fuel large-scale power generation for AI/HPC loads.\", \"workflow\": \"EQT agreed in principle to supply gas to a planned 4.4 GW generation campus; this is commodity supply and generation infrastructure supporting AI demand, not an internal AI workflow.\", \"technology_data\": \"No AI system at EQT; project partners include Knighthead, GE Vernova and Kiewit in project roles.\", \"integration_human_role\": \"Supply agreement terms and delivery responsibilities require definitive documents.\", \"deployment_scope\": \"Homer City, Pennsylvania; target first power 2027 per existing record; current review found issuer/official project document link but did not re-open full terms.\", \"partners\": [\"Knighthead; GE Vernova; Kiewit.\"], \"disclosed_results\": [], \"economics\": \"Gas volumes/pricing in the agreement in principle were not disclosed in source reviewed; infrastructure economics are distinct from AI adoption.\", \"timeline\": [{\"date\": \"2025-07\", \"event\": \"EQT gas-supply agreement in principle announced for Homer City AI/HPC project.\", \"source_ids\": [\"S42\"]}], \"limitations\": [\"Agreement in principle is not proof of final contract, construction completion or first power.\", \"This belongs in AI-demand infrastructure.\"], \"assessment\": \"Useful exposure to AI electricity demand, but it should not be counted as EQT adopting AI in its own operations.\", \"sources\": [{\"id\": \"S42\", \"title\": \"Homer City Redevelopment announces agreement in principle for EQT Corporation to supply natural gas\", \"url\": \"https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp\", \"publisher\": \"Homer City Redevelopment\", \"date\": \"2025-07\", \"type\": \"company announcement\", \"locator\": \"Agreement in principle\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Project announcement describes EQT agreement in principle for gas supply to redevelopment targeting AI/HPC campus; not an executed definitive supply contract.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-016\"}], \"field_sources\": {\"business_problem\": [\"S42\"], \"workflow\": [\"S42\"], \"technology_data\": [\"S42\"], \"deployment_scope\": [\"S42\"], \"integration_human_role\": [], \"economics\": []}, \"verification\": \"unable to verify\", \"corrections_to_existing\": \"AI-063 category is correct; primary project source for this existing claim was not available in current search excerpts, so leave source review gap and do not certify the full gas-supply details.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-010\", \"source_ids\": [\"S42\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-063\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "unable to verify", "Audit_flags": [{"claim": "EQT gas supply is part of AI adoption.", "reason": "This is natural-gas supply infrastructure serving AI/HPC power demand; categorize separately from internal AI. Agreement was in principle.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp"}], "Audit_review": {"id": "AI-063", "company": "EQT / Homer City Redevelopment", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp"], "findings": [{"severity": "low", "field": "category / stage", "issue": "These records are infrastructure or supply arrangements responding to AI-load demand, not operational AI systems.", "correction": "Retain only in the AI-demand infrastructure category and do not describe gas infrastructure itself as AI adoption.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Not AI", "Stage_review": {"id": "AI-063", "stage": "Not AI", "stage_basis": "EQT/Homer City agreement in principle is for EQT to source and supply natural gas to a planned 4.4-GW power facility for a data-center campus. It is an energy-supply relationship serving AI demand, not an AI workflow operated by EQT.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "source_date": "2025-07-15; page updated 2025-07-21", "evidence_origin": "Counterparty", "review_correction": "Exclude from chart of AI workflow maturity or retain only with Not AI label; preserve the separate commercial stage as agreement-in-principle.", "limits": "Announcement is from project developer/counterparty; definitive supply contract, power operations, and data-center service start are not established by this source.", "source_id": "S42"}, "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Gas supply to an AI data-centre power project; not AI adoption.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "unable to verify", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-064", "Company": "Williams", "Segment": "Midstream / gas-fired power", "Theme": "AI infrastructure", "Initiative": "Socrates gas-to-power infrastructure for AI demand", "Extracted_finding": "Williams is commercializing integrated gas supply, pipeline and generation capacity for data-center demand; it is not an AI operating workflow at Williams.", "Stage_at_disclosure": "Under construction in 2025; progress presented in 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Project capex and contracted economics belong to infrastructure underwriting. Do not classify as internal AI use or infer that every phase is operational.", "Source_IDs": "RS106; S43", "Source_date": "2026-08-04", "Source_type": "earnings release", "Source_locator": null, "Source_URL": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-08-04; see chronology for operating dates.", "Context": "Company reported breaking ground on the $1.6bn Socrates project to serve AI demand; 2026 presentation describes gas supply, pipeline and generation components and phased capacity/service milestones.", "Implementation": "Williams' August 2026 presentation describes Socrates as integrated gas supply, pipeline and power-generation infrastructure. It reports 556 MW at ISO conditions, with Plato South in service and Plato North targeted for Q4 2026. These are project-capacity and phase-status claims, not AI operating results.", "Assessment": "High materiality infrastructure build responding to compute demand; its economics should be analyzed as power/gas infrastructure rather than AI productivity.", "Detail_words": 70, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "AI-related commercial infrastructure", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Williams", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-064", "Initiative_ID": "AI-064", "Title": "Socrates turnkey power", "Date": "2026-08-04", "Finding": "Gas supply, pipeline and generation solution serves data centers.", "Metric": "556 MW ISO; Plato South in service, North targeted Q4 2026", "Source_IDs": "S43", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The August 2026 presentation describes 556 MW at ISO conditions, with Plato South in service and Plato North targeted for the fourth quarter of 2026.", "Context": "Williams’ Socrates program combines gas supply, pipelines and power generation for data-center customers.", "Assessment": "Partial service is a more advanced milestone than a development announcement. ISO rating is a standardized capacity measure, and the presentation does not identify the share of customer load devoted to AI."}], "Latest_disclosure": "2026-08-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-013"], "Research_search": " {\"key\": \"williams-socrates-ai-demand\", \"existing_ids\": [\"AI-064\"], \"related_existing_ids\": [], \"company\": \"Williams\", \"title\": \"Socrates gas-to-power infrastructure for AI demand\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Under construction in 2025; progress presented in 2026\", \"practical_relevance\": \"Williams is commercializing integrated gas supply, pipeline and generation capacity for data-center demand; it is not an AI operating workflow at Williams.\", \"business_problem\": \"Company reported breaking ground on the $1.6bn Socrates project to serve AI demand; 2026 presentation describes gas supply, pipeline and generation components and phased capacity/service milestones.\", \"workflow\": \"Company reported breaking ground on the $1\", \"technology_data\": \"Natural gas infrastructure and power generation; AI is the customer demand driver, not deployed technology.\", \"integration_human_role\": \"Williams develops and operates the energy assets; data-center customer consumes power.\", \"deployment_scope\": \"Project-level infrastructure; 2026 deck reports South phase in service and North expected later in 2026 (planned).\", \"partners\": [\"Blackstone\"], \"disclosed_results\": [], \"economics\": \"Project investment announced at $1.6bn; not an AI savings metric.\", \"timeline\": [{\"date\": \"2025-08-04\", \"event\": \"Earnings release said construction began on Socrates.\", \"source_ids\": [\"RS106\"]}, {\"date\": \"2026-08-04\", \"event\": \"Investor presentation updated project phases/status.\", \"source_ids\": [\"S43\"]}], \"limitations\": [\"Project capex and contracted economics belong to infrastructure underwriting.\", \"Do not classify as internal AI use or infer that every phase is operational.\"], \"assessment\": \"High materiality infrastructure build responding to compute demand; its economics should be analyzed as power/gas infrastructure rather than AI productivity.\", \"sources\": [{\"id\": \"RS106\", \"title\": \"Williams Q2 2025 results\", \"url\": \"https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises\", \"publisher\": \"Williams\", \"date\": \"2025-08-04\", \"type\": \"earnings release\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Describes $1.6bn Socrates power innovation project to serve growing AI demand; infrastructure, not internal AI.\"], \"original_id\": \"wmsocrates\"}, {\"id\": \"S43\", \"title\": \"Williams Q2 2026 earnings presentation\", \"url\": \"https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc\", \"publisher\": \"Williams\", \"date\": \"2026-08-04\", \"type\": \"investor presentation\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Project details include gas supply, pipeline and generation; presented as power infrastructure for data-center load.\"], \"original_id\": \"wms26pres\"}], \"field_sources\": {\"business_problem\": [\"RS106\"], \"workflow\": [\"RS106\", \"S43\"], \"technology_data\": [\"RS106\"], \"integration_human_role\": [\"S43\"], \"deployment_scope\": [\"RS106\", \"S43\"], \"economics\": [\"RS106\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-013\", \"source_ids\": [\"RS106\", \"S43\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-064\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-064", "company": "Williams", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc"], "findings": [{"severity": "low", "field": "category / stage", "issue": "These records are infrastructure or supply arrangements responding to AI-load demand, not operational AI systems.", "correction": "Retain only in the AI-demand infrastructure category and do not describe gas infrastructure itself as AI adoption.", "source_url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "infrastructure", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-065", "Company": "Enverus / Blackstone", "Segment": "Energy software", "Theme": "AI products", "Initiative": "Enverus ONE: governed AI energy workflow platform", "Extracted_finding": "Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.", "Stage_at_disclosure": "Launched / available 2026-04-07", "Reported_metric": "AFE evaluation duration: Weeks to hours (vendor claim) (Vendor-reported)", "Interpretation_limit": "Acquisition is reported as an agreement; this source does not establish closing. No independent customer case, accuracy, scale, or realized ROI disclosed.", "Source_IDs": "S48; S30; R099", "Source_date": "2026-04-07", "Source_type": "company announcement", "Source_locator": "Launch, platform design and four Flows", "Source_URL": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": "Blackstone investment", "Coverage_bucket": "Supplemental primary disclosure", "Time_scope": "Dated source: 2026-04-07; see chronology for operating dates.", "Context": "Energy teams evaluate AFEs and producing assets using fragmented ownership, engineering, production and economic evidence; site-siting also requires multi-factor screening.", "Implementation": "Four product Flows at launch: AFE Evaluation ingests AFE, validates ownership and benchmarks offset economics; Current Production Valuation selects wells, loads production, forecasts and economics; Project Siting ranks parcels against criteria; QuickStart makes customer documents searchable through secure Q&A.", "Assessment": "Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.", "Detail_words": 72, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "PE-backed commercial workflow platform", "Scope_type": "Platform spanning workflows", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Enverus", "Review_tier": "Reviewed source", "Provenance": ["Original register", "User-supplied reference"], "Reference_IDs": ["REF-099", "REF-100"], "Related_IDs": [], "Disclosures": [{"ID": "REF-099", "Initiative_ID": "AI-065", "Title": "Enverus AI launched as 'AI operating system for energy'", "Date": "2025-12", "Finding": "Enverus unveiled Enverus AI, a governed system built on its energy datasets. It covers asset lifecycle management, document analysis, title verification, field safety monitoring and M&A deal screening.", "Metric": "8,000+ customer companies; claims '10X workforce impact'", "Source_IDs": "R099", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Title, document and deal-screening workflows are being built into the incumbent data vendor's product, so specialist startups there face bundling pressure.", "Partners": "Enverus", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-100", "Initiative_ID": "AI-065", "Title": "Enverus ONE governed AI platform with workflow 'Flows'", "Date": "2026-04", "Finding": "Enverus launched Enverus ONE, which pairs frontier models with Astra, its own energy model, as an execution layer for energy workflows. The first 'Flows' are AFE Evaluation, Current Production Valuation, Project Siting and QuickStart, running in a SOC 2 Type II private environment.", "Metric": "8,000+ companies in 50 countries; 2.7 PB data; 350M+ courthouse records; $500B+ annual transactions tracked", "Source_IDs": "S48", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Packaging agents as named, outcome-specific workflows (such as AFE evaluation) on proprietary data is the product pattern to borrow.", "Partners": "Astra, frontier LLMs", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "ORIG-AI-065", "Initiative_ID": "AI-065", "Title": "Enverus ONE workflow platform", "Date": "2026-04-07", "Finding": "Enverus ONE launched in April 2026 as an AI platform for energy workflows. It pairs frontier models with Enverus’ proprietary energy model, Astra.", "Metric": "Four named launch workflows", "Source_IDs": "S48", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Four launch workflows cover AFE evaluation, current-production valuation, project siting and secure document Q&A. The vendor describes private tenancy, role-based access and audit trails.", "Context": "Enverus ONE launched in April 2026 as an AI platform for energy workflows. It pairs frontier models with Enverus’ proprietary energy model, Astra.", "Assessment": "This is a named product launch within a PE-backed energy platform. Availability and vendor time-saving claims do not establish adoption, accuracy or independent customer returns."}], "Latest_disclosure": "2026-04-07", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-002"], "Research_search": " {\"key\": \"enverus-one\", \"existing_ids\": [\"AI-065\"], \"related_existing_ids\": [\"AI-051\"], \"company\": \"Enverus\", \"title\": \"Enverus ONE: governed AI energy workflow platform\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched / available 2026-04-07\", \"practical_relevance\": \"Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.\", \"business_problem\": \"Energy teams evaluate AFEs and producing assets using fragmented ownership, engineering, production and economic evidence; site-siting also requires multi-factor screening.\", \"workflow\": \"Four product Flows at launch: AFE Evaluation ingests AFE, validates ownership and benchmarks offset economics; Current Production Valuation selects wells, loads production, forecasts and economics; Project Siting ranks parcels against criteria; QuickStart makes customer documents searchable through secure Q&A.\", \"technology_data\": \"Astra proprietary energy model is paired with general frontier models and Enverus datasets; private tenancy, role-based controls and audit trails.\", \"integration_human_role\": \"Outputs are decision support; governance and audit features support human verification. Details of review gates for each Flow are not disclosed.\", \"deployment_scope\": \"Product launched for energy customers; no named customer adoption or counts disclosed in launch.\", \"partners\": [\"Blackstone funds (owner transaction)\", \"Astra (Enverus model)\"], \"disclosed_results\": [{\"metric\": \"AFE evaluation duration\", \"value\": \"Weeks to hours (vendor claim)\", \"period\": \"Launch disclosure, 2026-04-07\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"S48\"]}], \"economics\": \"Vendor says AFE review can fall from weeks to hours; no customer-specific baseline or independently verified savings.\", \"timeline\": [{\"date\": \"2025-08-06\", \"event\": \"Blackstone funds announced agreement to acquire Enverus from H&F and Genstar.\", \"source_ids\": [\"S30\"]}, {\"date\": \"2026-04-07\", \"event\": \"Enverus ONE launched and was available, with four Flows.\", \"source_ids\": [\"S48\"]}], \"limitations\": [\"Acquisition is reported as an agreement; this source does not establish closing.\", \"No independent customer case, accuracy, scale, or realized ROI disclosed.\"], \"assessment\": \"Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.\", \"sources\": [{\"id\": \"S48\", \"title\": \"Enverus ONE is live: governed AI platform built for energy industry\", \"url\": \"https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/\", \"publisher\": \"Enverus\", \"date\": \"2026-04-07\", \"type\": \"company announcement\", \"locator\": \"Launch, platform design and four Flows\", \"review_status\": \"full text reviewed\", \"supports\": [\"Enverus ONE launched with Astra energy model paired with general frontier AI, proprietary energy datasets and governed access controls.\", \"AFE Evaluation, Current Production Valuation, Project Siting and QuickStart workflows are available.\", \"AFE evaluation automates ingestion, ownership validation, offset economics and comparisons; vendor says weeks can become hours.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-002\"}, {\"id\": \"S30\", \"title\": \"Blackstone announces agreement to acquire Enverus\", \"url\": \"https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/\", \"publisher\": \"Blackstone\", \"date\": \"2025-08-06\", \"type\": \"company announcement\", \"locator\": \"Transaction announcement\", \"review_status\": \"full text reviewed\", \"supports\": [\"Blackstone funds agreed to acquire Enverus from Hellman & Friedman and Genstar; this is a financial transaction, not an AI deployment.\", \"Enverus provides energy data analytics and insights used by energy companies and investors.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-001\"}], \"field_sources\": {\"business_problem\": [\"S48\"], \"workflow\": [\"S48\"], \"technology_data\": [\"S48\"], \"integration_human_role\": [\"S48\"], \"deployment_scope\": [\"S48\"], \"economics\": [\"S48\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-002\", \"source_ids\": [\"S48\", \"S30\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-065\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-065", "company": "Enverus / Blackstone", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Enverus ONE launch source supports platform availability/capabilities, not customer use or outcomes.", "correction": "Call it a commercial AI platform launch; deployments not named.", "source_url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-066", "Company": "Kuwait Oil Company / SLB", "Segment": "Upstream / services", "Theme": "Operational AI", "Initiative": "Four AI collaborations for KOC Jurassic gas field", "Extracted_finding": "The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.", "Stage_at_disclosure": "Contract awards reported July 2026; implementation plans, not verified completed rollout", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "SLB partner disclosure only. Distinguish awarded projects from completed applications.", "Source_IDs": "S27", "Source_date": "2026-07-24", "Source_type": "earnings release", "Source_locator": "Quarterly highlights, contract awards and technology highlights", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Context": "Improve Jurassic-gas decisions and KOC AI skills.", "Implementation": "Four awards: AI rig scheduling, GenAI production analytics, log QC/reconstruction, 3-month academy.", "Assessment": "Analyst inference: The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.", "Detail_words": 56, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "Kuwait Oil Company", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-066", "Initiative_ID": "AI-066", "Title": "Jurassic gas AI collaborations", "Date": "2026-07-24", "Finding": "KOC awarded four Innovation Factori collaborations for Jurassic gas.", "Metric": "Four collaborations, including a three-month academy", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The scope covers rig scheduling, GenAI production analytics, log quality and reconstruction, plus a three-month AI academy.", "Context": "KOC awarded four Innovation Factori collaborations for Jurassic gas.", "Assessment": "This is a distinct program from KOC’s seismic-processing contract in AI-014. The award establishes intended work, while field scale, completion milestones and economics remain unreported in the cited results release."}], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-007"], "Research_search": " {\"key\": \"slb-koc-jurassic\", \"existing_ids\": [\"AI-066\"], \"related_existing_ids\": [], \"company\": \"Kuwait Oil Company\", \"title\": \"Four AI collaborations for KOC Jurassic gas field\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Contract awards reported July 2026; implementation plans, not verified completed rollout\", \"practical_relevance\": \"The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.\", \"business_problem\": \"Improve Jurassic-gas decisions and KOC AI skills.\", \"workflow\": \"Four awards: AI rig scheduling, GenAI production analytics, log QC/reconstruction, 3-month academy.\", \"technology_data\": \"Models and interfaces undisclosed.\", \"integration_human_role\": \"Academy trains KOC experts.\", \"deployment_scope\": \"Jurassic gas; breadth absent.\", \"partners\": [\"SLB\"], \"disclosed_results\": [], \"economics\": \"Value/outcomes absent.\", \"timeline\": [{\"date\": \"2026-07-24\", \"event\": \"Four AI collaborations disclosed in Q2 earnings\", \"source_ids\": [\"S27\"]}], \"limitations\": [\"SLB partner disclosure only. Distinguish awarded projects from completed applications.\"], \"assessment\": \"Analyst inference: The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.\", \"sources\": [{\"id\": \"S27\", \"title\": \"SLB Announces Second-Quarter 2026 Results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"earnings release\", \"locator\": \"Quarterly highlights, contract awards and technology highlights\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy).\"], \"original_id\": \"slb_q2\"}], \"field_sources\": {\"business_problem\": [\"S27\"], \"workflow\": [\"S27\"], \"technology_data\": [\"S27\"], \"integration_human_role\": [\"S27\"], \"deployment_scope\": [\"S27\"], \"economics\": [\"S27\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-007\", \"source_ids\": [\"S27\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-066\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "AI seismic processing, Jurassic gas AI and G Agent drilling outcomes", "reason": "No official KOC source review completed in this pass to substantiate these entries; the SLB earnings source is vendor disclosure and Kuwait Times is secondary. Retain as reported/unverified pending primary KOC technical or procurement evidence.", "source_url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-066", "company": "Kuwait Oil Company / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-067", "Company": "PTTEP / SLB", "Segment": "Upstream / services", "Theme": "Operational AI", "Initiative": "PTTEP offshore production data and AI advisory pilot", "Extracted_finding": "A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.", "Stage_at_disclosure": "Pilot award reported July 2026; not yet evidenced as completed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Partner announcement only; no PTTEP confirmation or pilot results reviewed.", "Source_IDs": "S27", "Source_date": "2026-07-24", "Source_type": "earnings release", "Source_locator": "Quarterly highlights, contract awards and technology highlights", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Context": "Improve offshore production assurance.", "Implementation": "Pilot: Cognite Data Fusion, OptiFlow, Tela advisory agents.", "Assessment": "Analyst inference: A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.", "Detail_words": 58, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "PTTEP", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-067", "Initiative_ID": "AI-067", "Title": "Offshore production AI pilot", "Date": "2026-07-24", "Finding": "PTTEP contracted SLB to pilot digital production tools at a domestic offshore asset.", "Metric": "One domestic offshore asset", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The pilot combines Cognite Data Fusion, OptiFlow and Tela advisory agents.", "Context": "PTTEP contracted SLB to pilot digital production tools at a domestic offshore asset.", "Assessment": "The defined pilot connects data preparation with production assurance and well optimization. The source gives no named asset, pilot result or measured uplift. A pilot award should not be presented as an enterprise rollout."}], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-009"], "Research_search": " {\"key\": \"slb-pttep-pilot\", \"existing_ids\": [\"AI-067\"], \"related_existing_ids\": [], \"company\": \"PTTEP\", \"title\": \"PTTEP offshore production data and AI advisory pilot\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Pilot award reported July 2026; not yet evidenced as completed\", \"practical_relevance\": \"A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.\", \"business_problem\": \"Improve offshore production assurance.\", \"workflow\": \"Pilot: Cognite Data Fusion, OptiFlow, Tela advisory agents.\", \"technology_data\": \"Models unspecified.\", \"integration_human_role\": \"Advisory; approvals absent.\", \"deployment_scope\": \"One offshore asset.\", \"partners\": [\"SLB\", \"Cognite\"], \"disclosed_results\": [], \"economics\": \"Value/results absent.\", \"timeline\": [{\"date\": \"2026-07-24\", \"event\": \"Pilot contract disclosed by SLB\", \"source_ids\": [\"S27\"]}], \"limitations\": [\"Partner announcement only; no PTTEP confirmation or pilot results reviewed.\"], \"assessment\": \"Analyst inference: A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.\", \"sources\": [{\"id\": \"S27\", \"title\": \"SLB Announces Second-Quarter 2026 Results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"earnings release\", \"locator\": \"Quarterly highlights, contract awards and technology highlights\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q2 disclosure: PTTEP one-asset offshore pilot using Cognite, OptiFlow and Tela agents.\"], \"original_id\": \"slb_q2\"}], \"field_sources\": {\"business_problem\": [\"S27\"], \"workflow\": [\"S27\"], \"technology_data\": [\"S27\"], \"integration_human_role\": [\"S27\"], \"deployment_scope\": [\"S27\"], \"economics\": [\"S27\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-009\", \"source_ids\": [\"S27\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-067\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Offshore production AI pilot", "reason": "Existing cited source is SLB earnings material; no PTTEP primary-source verification completed. Vendor-reported and not re-verified.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"}, {"claim": "Second-pass review", "reason": "Partner/supplier primary evidence only; no operator-side corroboration found in this source pass.", "source_url": ""}], "Audit_review": {"id": "AI-067", "company": "PTTEP / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [{"severity": "medium", "field": "attribution", "issue": "SLB Q2 2026 is supplier-side corroboration for PTTEP offshore pilot; no direct PTTEP corroboration located in this audit.", "correction": "Label vendor-reported and preserve operator-confirmation gap.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-068", "Company": "CNOOC subsidiary / SLB", "Segment": "Upstream / services", "Theme": "Operational AI", "Initiative": "Closed-loop process optimization at CNOOC Bohai offshore terminal", "Extracted_finding": "This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.", "Stage_at_disclosure": "Operating. SLB’s 2026-07-24 quarterly release describes an integrated closed-loop process-optimization solution in live operation at CNOOC Bohai offshore terminal.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "SLB vendor disclosure; no operating metric or CNOOC corroboration reviewed.", "Source_IDs": "S27", "Source_date": "2026-07-24", "Source_type": "earnings release", "Source_locator": "Quarterly highlights, contract awards and technology highlights", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Context": "Optimize terminal process economics.", "Implementation": "Symmetry twins link plant data to APC/DCS closed-loop optimization.", "Assessment": "Analyst inference: This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.", "Detail_words": 58, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "CNOOC", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-068", "Initiative_ID": "AI-068", "Title": "Offshore-terminal process optimization", "Date": "2026-07-24", "Finding": "SLB worked with CNOOC subsidiary China France Bohai Geoservices on offshore-terminal process optimization.", "Metric": "No quantified financial outcome", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Operational data and Symmetry digital twins connect with process-control systems for real-time economic optimization.", "Context": "SLB worked with CNOOC subsidiary China France Bohai Geoservices on offshore-terminal process optimization.", "Assessment": "The disclosure describes a closed-loop implementation. It does not quantify profit, energy savings or the AI contribution relative to the broader control system. The operating outcome remains a vendor-reported qualitative claim."}], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-010"], "Research_search": " {\"key\": \"slb-cnooc-terminal\", \"existing_ids\": [\"AI-068\"], \"related_existing_ids\": [], \"company\": \"CNOOC\", \"title\": \"Closed-loop process optimization at CNOOC Bohai offshore terminal\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Integrated solution reported in July 2026; live closed-loop operation described\", \"practical_relevance\": \"This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.\", \"business_problem\": \"Optimize terminal process economics.\", \"workflow\": \"Symmetry twins link plant data to APC/DCS closed-loop optimization.\", \"technology_data\": \"AI, simulation, APC/DCS.\", \"integration_human_role\": \"Oversight unspecified.\", \"deployment_scope\": \"One CNOOC subsidiary terminal.\", \"partners\": [\"SLB\", \"China France Bohai Geoservices\"], \"disclosed_results\": [], \"economics\": \"No quantified benefit.\", \"timeline\": [{\"date\": \"2026-07-24\", \"event\": \"Deployment described in SLB Q2 results\", \"source_ids\": [\"S27\"]}], \"limitations\": [\"SLB vendor disclosure; no operating metric or CNOOC corroboration reviewed.\"], \"assessment\": \"Analyst inference: This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.\", \"sources\": [{\"id\": \"S27\", \"title\": \"SLB Announces Second-Quarter 2026 Results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"earnings release\", \"locator\": \"Quarterly highlights, contract awards and technology highlights\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q2 disclosure: CNOOC subsidiary terminal connects Symmetry twins to APC/DCS.\"], \"original_id\": \"slb_q2\"}], \"field_sources\": {\"business_problem\": [\"S27\"], \"workflow\": [\"S27\"], \"technology_data\": [\"S27\"], \"integration_human_role\": [\"S27\"], \"deployment_scope\": [\"S27\"], \"economics\": [\"S27\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-010\", \"source_ids\": [\"S27\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-068\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-068", "company": "CNOOC subsidiary / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-068", "stage": "Operating", "stage_basis": "SLB’s 2026-07-24 quarterly release describes an integrated closed-loop process-optimization solution in live operation at CNOOC Bohai offshore terminal.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "source_date": "2026-07-24", "evidence_origin": "Vendor", "limits": "Single terminal and vendor-reported status; no operating metrics or CNOOC corroboration reviewed.", "source_id": "S27"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-069", "Company": "ExxonMobil Trinidad and Tobago / SLB", "Segment": "Upstream / services", "Theme": "Operational AI", "Initiative": "AI-enabled horizon interpretation in frontier seismic imaging", "Extracted_finding": "Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.", "Stage_at_disclosure": "Two-year award reported 2026-07-24", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "AI is a component of wider seismic processing contract; source does not isolate its contribution. No prospect or discovery result yet.", "Source_IDs": "S27", "Source_date": "2026-07-24", "Source_type": "partner/customer announcement", "Source_locator": "Digital/technology projects section", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Requested source class", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Context": "Sparse legacy 2D seismic data makes ultra-deepwater frontier imaging and prospect evaluation difficult.", "Implementation": "SLB awarded two-year data processing scope to generate first 3D image over nearly 6,000 km²; broader processing includes FWI, velocity modeling, source separation and multicomponent streamer data, with AI-enabled horizon interpretation.", "Assessment": "Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.", "Detail_words": 59, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational owner is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Defined operational or business workflow", "Ambition_assessment_basis": "Qualitative comparison within the consolidated research. Breadth, intended autonomy, integration depth and intended scale; assessed separately from evidence and achieved results.", "Company_primary": "ExxonMobil", "Review_tier": "Reviewed source", "Provenance": ["Original register"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "ORIG-AI-069", "Initiative_ID": "AI-069", "Title": "Frontier seismic AI interpretation", "Date": "2026-07-24", "Finding": "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic-processing project.", "Metric": "Nearly 6,000 km²; two-year project", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The nearly 6,000 km² program combines advanced imaging with AI-enabled horizon interpretation to create a first 3D image.", "Context": "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic-processing project.", "Assessment": "The AI component sits inside a larger imaging workflow. A processing award does not establish a discovery or commercial reserves, and the source does not isolate the AI contract value."}], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-009"], "Research_search": " {\"key\": \"exxon-tt-seismic\", \"existing_ids\": [\"AI-069\"], \"related_existing_ids\": [], \"company\": \"ExxonMobil Trinidad and Tobago / SLB\", \"title\": \"AI-enabled horizon interpretation in frontier seismic imaging\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Two-year award reported 2026-07-24\", \"practical_relevance\": \"Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.\", \"business_problem\": \"Sparse legacy 2D seismic data makes ultra-deepwater frontier imaging and prospect evaluation difficult.\", \"workflow\": \"SLB awarded two-year data processing scope to generate first 3D image over nearly 6,000 km²; broader processing includes FWI, velocity modeling, source separation and multicomponent streamer data, with AI-enabled horizon interpretation.\", \"technology_data\": \"SLB Omega geophysical processing software, geophysical imaging suite and AI-enabled horizon interpretation; source does not explain model architecture/training data.\", \"integration_human_role\": \"Output is intended to help ExxonMobil evaluate area; interpretation remains an exploration decision-support workflow, no autonomous decision claim.\", \"deployment_scope\": \"Named ExxonMobil Trinidad and Tobago Deepwater block; 6,000 km², two-year contract.\", \"partners\": [\"SLB\"], \"disclosed_results\": [], \"economics\": \"Contract value and commercial discoveries/returns not disclosed.\", \"timeline\": [{\"date\": \"2026-07-24\", \"event\": \"AI-enabled horizon interpretation in frontier seismic imaging documented or disclosed.\", \"source_ids\": [\"S27\"]}], \"limitations\": [\"AI is a component of wider seismic processing contract; source does not isolate its contribution.\", \"No prospect or discovery result yet.\"], \"assessment\": \"Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.\", \"sources\": [{\"id\": \"S27\", \"title\": \"SLB Q2 2026 results: ExxonMobil Trinidad & Tobago seismic project\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"partner/customer announcement\", \"locator\": \"Digital/technology projects section\", \"review_status\": \"full text reviewed\", \"supports\": [\"ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.\", \"Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.\", \"Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-011\"}], \"field_sources\": {\"business_problem\": [\"S27\"], \"workflow\": [\"S27\"], \"technology_data\": [\"S27\"], \"integration_human_role\": [\"S27\"], \"deployment_scope\": [\"S27\"], \"economics\": [\"S27\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-009\", \"source_ids\": [\"S27\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-069\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-069", "company": "ExxonMobil Trinidad and Tobago / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-070", "Company": "Saudi Aramco", "Company_primary": "Aramco", "Segment": "NOC", "HQ": "Saudi Arabia", "Theme": "Enterprise AI", "Initiative": "Microsoft industrial AI and digital-talent MoU", "Extracted_finding": "Enablement program whose future operating deployments require separate evidence.", "Context": "Scaling industrial AI with data-residency and governance requirements.", "Implementation": "The parties will explore Azure industrial solutions, cloud sovereignty, infrastructure optimization, an industry alliance and co-developed commercial AI systems.", "Assessment": "Broad scope, early commitment level. Track binding projects, named workflows, deployment milestones and user adoption before treating it as realized industrial transformation.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Non-binding memorandum; exploratory collaboration", "Interpretation_limit": "No implementation budget, production-user count or measured result disclosed.", "Source_IDs": "R003", "Source_date": "2026-02-12", "Source_type": "Company announcement / remarks", "Source_locator": "MoU scope and four focus areas", "Source_URL": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-02-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-003", "Initiative_ID": "AI-070", "Title": "Aramco-Microsoft MoU on industrial AI and digital talent", "Date": "2026-02", "Finding": "Aramco and Microsoft signed a non-binding MoU to accelerate industrial AI adoption on Azure with sovereign controls for Saudi data residency. Focus areas include operational efficiency, an industry alliance framework, industrial AI IP co-innovation, and workforce programs in AI engineering, cybersecurity, data governance and product management.", "Metric": "No quantified metric supplied", "Source_IDs": "R003", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Sovereignty/data residency is a gating requirement for Gulf NOCs; enablement offerings must pair tooling with local-cloud deployment and talent programs.", "Partners": "Microsoft, Azure", "Original_date": "2026-02", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-003"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["root_backfill-004"], "Research_search": " {\"key\": \"aramco-microsoft\", \"existing_ids\": [\"AI-070\"], \"company\": \"Aramco\", \"title\": \"Microsoft industrial AI and digital-talent MoU\", \"category\": \"Strategy only\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"stage\": \"Non-binding memorandum; exploratory collaboration\", \"verification\": \"verified primary\", \"partners\": [\"Microsoft\"], \"sources\": [{\"id\": \"R003\", \"title\": \"Aramco signs MoU with Microsoft to help advance industrial AI and digital talent transformation\", \"url\": \"https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation\", \"date\": \"2026-02-12\", \"locator\": \"MoU scope and four focus areas\", \"publisher\": \"Aramco\", \"type\": \"Company announcement / remarks\", \"review_status\": \"Full primary text reviewed\", \"supports\": [], \"accessed\": \"2026-09-27\", \"original_id\": \"AR4\"}], \"field_sources\": {\"business_problem\": [\"R003\"], \"workflow\": [\"R003\"], \"technology_data\": [\"R003\"], \"integration_human_role\": [\"R003\"], \"deployment_scope\": [\"R003\"], \"economics\": [\"R003\"]}, \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-02-12\", \"event\": \"Non-binding MoU signed.\", \"source_ids\": [\"R003\"]}], \"limitations\": [\"No implementation budget, production-user count or measured result disclosed.\"], \"economics\": \"Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.\", \"practical_relevance\": \"Enablement program whose future operating deployments require separate evidence.\", \"business_problem\": \"Scaling industrial AI with data-residency and governance requirements.\", \"workflow\": \"The parties will explore Azure industrial solutions, cloud sovereignty, infrastructure optimization, an industry alliance and co-developed commercial AI systems.\", \"technology_data\": \"Microsoft Azure and potential sovereign controls; solution architecture not disclosed.\", \"integration_human_role\": \"Proposed skills work covers AI engineering, cybersecurity, data governance and product management.\", \"deployment_scope\": \"Aramco operations and the Saudi industrial ecosystem are intended areas of focus. A named production rollout is not established.\", \"assessment\": \"Broad scope, early commitment level. Track binding projects, named workflows, deployment milestones and user adoption before treating it as realized industrial transformation.\", \"corrections_to_existing\": \"Makes the non-binding and exploratory status explicit.\", \"cohort\": \"root_backfill\", \"packet_id\": \"root_backfill-004\", \"source_ids\": [\"R003\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-070\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-070", "company": "Saudi Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation"], "findings": [{"severity": "medium", "field": "stage", "issue": "Aramco/Microsoft source is a non-binding MoU describing areas to explore.", "correction": "Stage as exploratory partnership, not deployed solution.", "source_url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-071", "Company": "ADNOC", "Company_primary": "ADNOC", "Segment": "NOC", "HQ": "UAE", "Theme": "Enterprise AI", "Initiative": "Enterprise AI workforce training and agentic AI capability building", "Extracted_finding": "The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.", "Context": "Develop employee AI literacy and practical capability as ADNOC expands enterprise AI use.", "Implementation": "ADNOC reports enterprise-wide generative AI rollout with Microsoft Copilot in Nov 2023; the 2025 sustainability report separately identifies Introduction to Agentic AI training, engaging over 18,000 participants and resulting in over 10,000 AI agents created across operations.", "Assessment": "Analyst inference: The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.", "Reported_metric": "Employees completing AI training: >40,000 (Company-reported aggregate); Monthly productivity hours gained: >70,000 hours/month (Company-reported; not isolated to training); Agentic AI training participants / agents created: >18,000 participants / >10,000 agents (Company-reported)", "Stage_at_disclosure": "By 2025, >40,000 employees completed AI training; 2025 report also records a separate agentic AI learning program", "Interpretation_limit": "40,000 trained employees, >90% utilization and 70,000 hours/month are corporate aggregate disclosures, not training-attributable outcomes. The reported >10,000 created agents are not evidence that all were deployed, used, or agentic to the same extent.", "Source_IDs": "RS030; RS031; R005", "Source_date": "2026", "Source_type": "company announcement", "Source_locator": "Copilot history, training and productivity aggregate", "Source_URL": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-005", "Initiative_ID": "AI-071", "Title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "Date": "2025-11", "Finding": "ADNOC deployed generative AI company-wide in November 2023; by ADIPEC 2025 more than 40,000 employees had completed AI training, with utilization above 90% and over 70,000 productivity hours gained per month. At the same time ADNOC signed an agreement with Microsoft to co-develop and deploy AI agents across its operations.", "Metric": "40,000+ trained; >90% utilization; 70,000+ hours/month saved", "Source_IDs": "R005", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Enterprise-wide training plus usage and hours-saved telemetry is the adoption playbook to copy; report utilization, not just licenses.", "Partners": "Microsoft, AIQ, Masdar, XRG, Gecko Robotics", "Original_date": "2025-11", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-005"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["nocs-017"], "Research_search": " {\"key\": \"adnoc-ai-workforce\", \"existing_ids\": [\"AI-071\"], \"related_existing_ids\": [], \"company\": \"ADNOC\", \"title\": \"Enterprise AI workforce training and agentic AI capability building\", \"department\": \"Workforce & learning\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"By 2025, >40,000 employees completed AI training; 2025 report also records a separate agentic AI learning program\", \"practical_relevance\": \"The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.\", \"business_problem\": \"Develop employee AI literacy and practical capability as ADNOC expands enterprise AI use.\", \"workflow\": \"ADNOC reports enterprise-wide generative AI rollout with Microsoft Copilot in Nov 2023; the 2025 sustainability report separately identifies Introduction to Agentic AI training, engaging over 18,000 participants and resulting in over 10,000 AI agents created across operations.\", \"technology_data\": \"Microsoft Copilot rollout, internal AI learning programs and agent-creation activity; detailed curriculum, agent architecture and usage controls not disclosed.\", \"integration_human_role\": \"Training and enablement are human capability measures; the sources do not attribute productivity gains specifically to the training program.\", \"deployment_scope\": \"ADNOC-wide workforce figures; 2025 report gives learner and agent counts but no breakdown by business unit or operational risk tier.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Employees completing AI training\", \"value\": \">40,000\", \"period\": \"As reported Nov. 2025\", \"attribution\": \"Company-reported aggregate\", \"source_ids\": [\"RS030\"]}, {\"metric\": \"Monthly productivity hours gained\", \"value\": \">70,000 hours/month\", \"period\": \"As reported Nov. 2025 since 2023 enterprise Copilot rollout\", \"attribution\": \"Company-reported; not isolated to training\", \"source_ids\": [\"RS030\"]}, {\"metric\": \"Agentic AI training participants / agents created\", \"value\": \">18,000 participants / >10,000 agents\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS031\"]}], \"economics\": \"No training cost, controlled productivity baseline or causal attribution to training disclosed.\", \"timeline\": [{\"date\": \"2023-11\", \"event\": \"ADNOC says it rolled out enterprise generative AI with Microsoft Copilot; pre-window context.\", \"source_ids\": [\"RS030\"]}, {\"date\": \"2025-11-02\", \"event\": \"ADNOC reports training and productivity aggregates alongside a new co-development agreement with Microsoft.\", \"source_ids\": [\"RS030\"]}, {\"date\": \"2025\", \"event\": \"Sustainability report documents agentic AI learning program activity.\", \"source_ids\": [\"RS031\"]}], \"limitations\": [\"40,000 trained employees, >90% utilization and 70,000 hours/month are corporate aggregate disclosures, not training-attributable outcomes.\", \"The reported >10,000 created agents are not evidence that all were deployed, used, or agentic to the same extent.\"], \"assessment\": \"Analyst inference: The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.\", \"sources\": [{\"id\": \"RS030\", \"title\": \"ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI\", \"url\": \"https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai\", \"publisher\": \"ADNOC\", \"date\": \"2025-11-02\", \"type\": \"company announcement\", \"locator\": \"Copilot history, training and productivity aggregate\", \"review_status\": \"full text reviewed\", \"supports\": [\"ADNOC reports enterprise Copilot rollout in 2023, >40,000 employees completing AI training, utilization above 90%, and >70,000 monthly productivity hours; no training causal attribution stated.\"], \"original_id\": \"adnoc_msft_ai\"}, {\"id\": \"RS031\", \"title\": \"ADNOC Sustainability Report 2025\", \"url\": \"https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42\", \"publisher\": \"ADNOC\", \"date\": \"2026\", \"type\": \"annual report\", \"locator\": \"AI and digital capabilities / workforce training\", \"review_status\": \"full text reviewed\", \"supports\": [\"In 2025, ADNOC says its Introduction to Agentic AI program engaged >18,000 participants and resulted in >10,000 agents created across operations.\"], \"original_id\": \"adnoc_sustainability\"}], \"field_sources\": {\"business_problem\": [\"RS030\", \"RS031\"], \"workflow\": [\"RS030\", \"RS031\"], \"technology_data\": [\"RS030\", \"RS031\"], \"integration_human_role\": [\"RS030\", \"RS031\"], \"deployment_scope\": [\"RS030\", \"RS031\"], \"economics\": [\"RS030\", \"RS031\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Verified primary source for the workforce training/enablement record; it is separate from ENERGYai (AI-040).\", \"cohort\": \"nocs\", \"packet_id\": \"nocs-017\", \"source_ids\": [\"RS030\", \"RS031\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-071\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "40,000+ employees AI-trained; 70,000 productivity hours per month", "reason": "Existing profile cites a news source rather than primary company evidence; not re-verified in reviewed primary disclosures. Do not treat as disproved.", "source_url": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-071", "company": "ADNOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "40,000+ employees trained", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Productivity hours are company-reported and not isolated to the training programme.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-072", "Company": "ExxonMobil", "Company_primary": "ExxonMobil", "Segment": "Supermajor", "HQ": "USA", "Theme": "Operational AI", "Initiative": "AI advisory and closed-loop autonomous drilling in Guyana", "Extracted_finding": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.", "Context": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.", "Implementation": "Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "Assessment": "How much well-delivery time changed after automation, and how often did shore teams intervene?", "Reported_metric": "SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "Stage_at_disclosure": "Existing operating profile; supplier-reported results now visible", "Interpretation_limit": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.", "Source_IDs": "RS003; R010; SW7-014", "Source_date": "Undated (2026 URL)", "Source_type": "Vendor customer case", "Source_locator": "Results: Autonomous drilling and rig automation drive record-setting performance in Guyana", "Source_URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Review_tier": "Reviewed source", "Time_scope": "Undated case in the supplier’s 2026 URL directory. Exact start, observation period, well count and definition of the operations denominator are not supplied.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-010", "Initiative_ID": "AI-072", "Title": "First fully autonomous AI-drilled well offshore Guyana", "Date": "2026-03", "Finding": "ExxonMobil completed what it called the industry's first fully closed-loop automated well placement with full rig automation in the Stabroek Block. VP of Exploration John Ardill said no driller touched the controls and AI handled drilling and geosteering through the reservoir section.", "Metric": "No quantified metric supplied", "Source_IDs": "R010", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Closed-loop autonomy is arriving first in drilling, where outcomes are measurable and operations are repeatable; human oversight shifts to remote centers.", "Partners": "Halliburton (LOGIX, EarthStar), Sekal (DrillTronics), Noble", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "SW7-D-AI-072", "Initiative_ID": "AI-072", "Title": "Guyana Neuro and DrillOps directional-drilling automation", "Date": "2026 (source URL; exact date unknown)", "Finding": "Neuro controls trajectory and downlink requests; DrillOps automates surface equipment and drilling tasks. Shore teams monitor operations.", "Metric": "SLB reports 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "Source_IDs": "SW7-014", "Evidence": "Original full page opened/read; deployment, named products, functions, 48-km path, autonomy share, and downlink metrics appear in final case-study section. SLB is the supplier and reports the performance claims.", "Origin": "Seventh research swarm"}, {"ID": "OCT14C-AI-072", "Initiative_ID": "AI-072", "Title": "Autonomous drilling in Guyana, with supplier-reported results", "Date": "Undated (2026 URL)", "Finding": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work. Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "Metric": "SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "Source_IDs": "SW7-014", "Source_URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained."}], "Related_IDs": [], "Reference_IDs": ["REF-010"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["swarm7-ai-072", "majors-003", "oct14c-ai-072"], "Research_search": "{\"packet_id\": \"oct14c-ai-072\", \"key\": \"oct14c-ai-072\", \"record_ids\": [\"AI-072\"], \"company\": \"ExxonMobil\", \"title\": \"Autonomous drilling in Guyana, with supplier-reported results\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Existing operating profile; supplier-reported results now visible\", \"business_problem\": \"SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.\", \"workflow\": \"Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.\", \"technology_data\": \"Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.\", \"deployment_scope\": \"Undated case in the supplier’s 2026 URL directory. Exact start, observation period, well count and definition of the operations denominator are not supplied.\", \"integration_human_role\": \"Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.\", \"economics\": \"SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.\", \"practical_relevance\": \"How much well-delivery time changed after automation, and how often did shore teams intervene?\", \"assessment\": \"How much well-delivery time changed after automation, and how often did shore teams intervene?\", \"limitations\": [\"Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.\"], \"source_ids\": [\"SW7-014\"], \"sources\": [{\"id\": \"SW7-014\", \"title\": \"ExxonMobil Guyana: drilling automation in deepwater operations\", \"url\": \"https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations\", \"date\": \"Undated (2026 URL)\", \"type\": \"Vendor customer case\", \"publisher\": \"SLB\", \"locator\": \"Results: Autonomous drilling and rig automation drive record-setting performance in Guyana\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"Undated (2026 URL)\", \"event\": \"SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.\", \"source_ids\": [\"SW7-014\"]}], \"field_sources\": {\"business_problem\": [\"SW7-014\"], \"workflow\": [\"SW7-014\"], \"technology_data\": [\"SW7-014\"], \"deployment_scope\": [\"SW7-014\"], \"economics\": [\"SW7-014\"]}}", "Research_status": "Clarified 14 October 2026 · third pass", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [{"claim": "First fully autonomous AI-drilled well offshore Guyana (2026).", "reason": "Official page confirms current AI advisory and closed-loop drilling in Guyana, but source does not verify the specific first-well milestone or timing.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies"}], "Audit_review": {"id": "AI-072", "company": "ExxonMobil", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-072", "stage": "Operating", "stage_basis": "ExxonMobil’s public digital-technologies page states its proprietary AI drilling-advisory system is currently used in Guyana and supports closed-loop automation without human intervention. The corporate page is undated, so this is the page’s present-tense status, not a date-verified change within the window.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "source_date": "Undated; checked 2026-10-04", "evidence_origin": "Operator", "review_correction": "Keep the original operator workflow and its undated source date. Do not assign the 2026 URL date from the separate SLB supplemental case to ExxonMobil’s operator disclosure or treat it as exact publication date.", "limits": "Page says currently used but gives no deployment date, scale, performance measures, autonomy boundaries, or last-updated date. Current statement alone cannot establish presence by 2026-10-01.", "source_id": "RS003"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "SLB reports 48+ km and 93%+ autonomy; exact operations denominator and well count undisclosed", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.", "Outcome_counted": true, "Outcome_count_note": "Supplier-reported operational results; not net cash ROI.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Existing operating profile; supplier-reported results now visible", "independent_validation": "Not established"}, "Current_source_check": "2026-10-14", "Research_origin": ["research/gap-swarm-oct14c/README.md"]}, {"ID": "AI-073", "Company": "ExxonMobil", "Company_primary": "ExxonMobil", "Segment": "Supermajor", "HQ": "USA", "Theme": "Operational AI", "Initiative": "AI identifies four new Guyana exploration opportunities", "Extracted_finding": "ExxonMobil said it used AI, including deep learning, machine learning, HPC and advanced seismic imaging, applied to historical discoveries and drilling results to identify four new exploration opportunities in Guyana's Stabroek Block. It is expanding use of these tools to evaluate prospects that were previously hard to assess.", "Context": "ExxonMobil said it used AI, including deep learning, machine learning, HPC and advanced seismic imaging, applied to historical discoveries and drilling results to identify four new exploration opportunities in Guyana's Stabroek Block. It is expanding use of these tools to evaluate prospects that were previously hard to assess.", "Implementation": "Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Reinterpreting legacy subsurface data with AI can surface new prospects in mature acreage; data readiness is the prerequisite.", "Reported_metric": "4 new opportunities identified", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Imported claim and interpretation from the supplied reference. Source text, date, scope and outcome attribution require independent verification.", "Source_IDs": "R011", "Source_date": "2026-08", "Source_type": "World Oil", "Source_locator": "See source and disclosure timeline", "Source_URL": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Reference date within window; date not independently verified", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-011", "Initiative_ID": "AI-073", "Title": "AI identifies four new Guyana exploration opportunities", "Date": "2026-08", "Finding": "ExxonMobil said it used AI, including deep learning, machine learning, HPC and advanced seismic imaging, applied to historical discoveries and drilling results to identify four new exploration opportunities in Guyana's Stabroek Block. It is expanding use of these tools to evaluate prospects that were previously hard to assess.", "Metric": "4 new opportunities identified", "Source_IDs": "R011", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Reinterpreting legacy subsurface data with AI can surface new prospects in mature acreage; data readiness is the prerequisite.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-011"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-08", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "AI identifies four new Guyana exploration opportunities (2026).", "reason": "Secondary news result found; did not locate/open matching ExxonMobil primary release in this pass. Do not certify without primary corroboration.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai"}, {"claim": "AI identified four new exploration opportunities in Guyana.", "reason": "Only secondary reports/call reproductions surfaced; no primary Exxon release was located. Treat as secondary-only pending direct company confirmation. The claim is not disproved.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-073", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "scale_only", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-074", "Company": "Chevron", "Company_primary": "Chevron", "Segment": "Supermajor", "HQ": "USA", "Theme": "Operational AI", "Initiative": "ApEX multi-agent exploration search", "Extracted_finding": "A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.", "Context": "Exploration screening requires geoscientists to locate and reconcile relevant information across large, heterogeneous subsurface archives.", "Implementation": "Explorationists query a chat interface; multiple specialized search agents retrieve relevant materials such as seismic interpretations, logs, maps, reports and prior plans for expert review.", "Assessment": "Analyst inference: A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.", "Reported_metric": "Corpus size: >1 million files / 4 TB (Company-reported)", "Stage_at_disclosure": "Deployed since Aug 2024; update disclosed Nov 2025", "Interpretation_limit": "Source is company description; no independent accuracy or decision-impact evidence.", "Source_IDs": "RS001; R013", "Source_date": "2025-11-12", "Source_type": "Chevron newsroom article", "Source_locator": "How does ApEX work? / Advantages in exploration", "Source_URL": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-11-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-013", "Initiative_ID": "AI-074", "Title": "ApEX multi-agent GenAI searches more than one million exploration files", "Date": "2025-11-12", "Finding": "Chevron says its ApEX generative-AI platform launched in August 2024 and was first used in Gulf of America deepwater exploration. Explorationists query a chat interface; multiple specialized search agents search more than one million files. Chevron reports a 4 TB corpus and says results surface in seconds.", "Metric": "More than 1 million files; 4 TB corpus; search responses in seconds (company-reported)", "Source_IDs": "R013; RS001", "Evidence": "Verified primary; company-reported capabilities", "Origin": "Chevron newsroom article and redistributed copy", "Supplied_interpretation": "Analyst inference: specialist agents can make heterogeneous exploration archives easier to search; no independently measured drilling or study-time result disclosed.", "Partners": "In-house", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-013"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["majors-001"], "Research_search": " {\"key\": \"chevron-apex\", \"existing_ids\": [\"AI-074\"], \"company\": \"Chevron\", \"title\": \"ApEX multi-agent exploration search\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Deployed since Aug 2024; update disclosed Nov 2025\", \"practical_relevance\": \"A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.\", \"business_problem\": \"Exploration screening requires geoscientists to locate and reconcile relevant information across large, heterogeneous subsurface archives.\", \"workflow\": \"Explorationists query a chat interface; multiple specialized search agents retrieve relevant materials such as seismic interpretations, logs, maps, reports and prior plans for expert review.\", \"technology_data\": \"In-house multi-agent generative AI; described corpus exceeds 1 million files and 4 TB. Specific model, RAG configuration and integration systems not disclosed.\", \"integration_human_role\": \"Explorationists formulate queries, inspect the supporting materials and retain prospect/economic judgment.\", \"deployment_scope\": \"First used in Gulf of America deepwater exploration; company says it can support decisions around acreage investment. Further assets unreported.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Corpus size\", \"value\": \">1 million files / 4 TB\", \"period\": \"2025 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS001\"]}], \"economics\": \"Chevron says ApEX is designed to improve speed and information access; no time saved, exploration success, or return metric disclosed.\", \"timeline\": [{\"date\": \"2024-08\", \"event\": \"ApEX launched and first used in Gulf deepwater exploration.\", \"source_ids\": [\"RS001\"]}, {\"date\": \"2025-11-12\", \"event\": \"Chevron disclosed current corpus size and workflow.\", \"source_ids\": [\"RS001\"]}], \"limitations\": [\"Source is company description; no independent accuracy or decision-impact evidence.\"], \"assessment\": \"Analyst inference: A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.\", \"sources\": [{\"id\": \"RS001\", \"title\": \"A smarter way to prospect for oil and gas — Chevron ApEX\", \"url\": \"https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas\", \"publisher\": \"Chevron\", \"date\": \"2025-11-12\", \"type\": \"company announcement\", \"locator\": \"How does ApEX work? / Advantages in exploration\", \"review_status\": \"full text reviewed\", \"supports\": [\"ApEX launched in August 2024 and was first used for deepwater exploration in the Gulf of America.\", \"The in-house multi-agent GenAI system searches more than one million files (4 TB) for explorationists, combining data from surveys, well logs, maps and technical reports.\", \"Use supports prospect screening and investment decisions by making subsurface materials easier to retrieve; experts remain the users and decision-makers.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-001\"}], \"field_sources\": {\"business_problem\": [\"RS001\"], \"workflow\": [\"RS001\"], \"technology_data\": [\"RS001\"], \"integration_human_role\": [\"RS001\"], \"deployment_scope\": [\"RS001\"], \"economics\": [\"RS001\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"AI-074 relied on a secondary article/vendor context; this official 2025 disclosure verifies ApEX and adds technical/workflow specifics.\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-001\", \"source_ids\": [\"RS001\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-074\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-074", "company": "Chevron", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c"], "findings": [{"severity": "medium", "field": "date / metric", "issue": "Issuer article is dated 2025-11-12 and supports >1 million files, 4 TB and seconds-fast retrieval; it does not quantify pre-drill studies falling from weeks to minutes.", "correction": "Use the exact Nov. 12, 2025 date and remove the weeks-to-minutes result unless another source supports it.", "source_url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas"}, {"severity": "medium", "field": "Supplemental source review", "issue": "Chevron's own article is dated 2025-11-12, not the supplied December month. It states >1 million files, 4 TB, and seconds-fast retrieval; it does not say studies fell from weeks to minutes.", "correction": "Date REF-013 to 2025-11-12 and remove 'weeks to minutes' as an achieved metric unless another source directly supports it. Keep the corpus size and retrieval workflow.", "source_url": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "supported with metric/date correction"}, "Audit_outcome": "corrected", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-075", "Company": "Chevron", "Company_primary": "Chevron", "Segment": "Supermajor", "HQ": "USA", "Theme": "Operational AI", "Initiative": "Edge agentic assessment for robotic inspections", "Extracted_finding": "Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.", "Context": "Aerial and terrestrial robot inspection missions require operator-by-operator permit checks, adding time and variability before work starts.", "Implementation": "Safe Start agents assess operating conditions before and during robotic missions and generate plans; specialized sensor-analysis and geospatial agents coordinate. Chevron describes moving from manual point-in-time checks toward continuous assessment while inspection proceeds.", "Assessment": "Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Collaboration announced 2026-06-02", "Interpretation_limit": "Partner announcement; deployment scale and outcomes not independently reported. AI-assisted missions are not stated as human-free inspection.", "Source_IDs": "R015", "Source_date": "2026-06-02", "Source_type": "partner case study", "Source_locator": "Safe Start agent, robot missions and edge systems", "Source_URL": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-06-02; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-015", "Initiative_ID": "AI-075", "Title": "Agentic AI for robotic inspections", "Date": "2026-06-02", "Finding": "DataRobot and Chevron described edge agents for robotic mission planning and continuous Safe Start assessment using gas sensors, vision and local AI.", "Metric": "No quantified financial benefit disclosed", "Source_IDs": "R015", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Agentic AI adds most value where it removes human approval bottlenecks on already automated physical workflows, with governance built in.", "Partners": "DataRobot, NVIDIA NIM", "Original_date": "2026-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-015"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-06-02", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-002"], "Research_search": " {\"key\": \"chevron-inspections\", \"existing_ids\": [\"AI-075\"], \"related_existing_ids\": [], \"company\": \"Chevron / DataRobot\", \"title\": \"Edge agentic assessment for robotic inspections\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Collaboration announced 2026-06-02\", \"practical_relevance\": \"Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.\", \"business_problem\": \"Aerial and terrestrial robot inspection missions require operator-by-operator permit checks, adding time and variability before work starts.\", \"workflow\": \"Safe Start agents assess operating conditions before and during robotic missions and generate plans; specialized sensor-analysis and geospatial agents coordinate. Chevron describes moving from manual point-in-time checks toward continuous assessment while inspection proceeds.\", \"technology_data\": \"DataRobot Agent Workforce Platform, NVIDIA NIM, NVIDIA AI compute, existing hard-wired gas sensors plus vision systems; edge/local assessment and integration into existing Chevron systems.\", \"integration_human_role\": \"System maintains oversight/control; human intervention is intended to decrease while approved operational standards, security, reliability and governance remain. Exact stop/go authority and fallback protocol not disclosed.\", \"deployment_scope\": \"Collaboration at Chevron facilities; no fleet/site count, rollout status beyond announced collaboration, or quantified result.\", \"partners\": [\"DataRobot\", \"NVIDIA\"], \"disclosed_results\": [], \"economics\": \"No quantified safety, time or financial outcomes. Avoid “autonomous inspections at scale” as established result.\", \"timeline\": [{\"date\": \"2026-06-02\", \"event\": \"Edge agentic assessment for robotic inspections documented or disclosed.\", \"source_ids\": [\"R015\"]}], \"limitations\": [\"Partner announcement; deployment scale and outcomes not independently reported.\", \"AI-assisted missions are not stated as human-free inspection.\"], \"assessment\": \"Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.\", \"sources\": [{\"id\": \"R015\", \"title\": \"DataRobot and Chevron Collaborate to Advance Agentic AI for Autonomous Inspections\", \"url\": \"https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/\", \"publisher\": \"DataRobot\", \"date\": \"2026-06-02\", \"type\": \"partner case study\", \"locator\": \"Safe Start agent, robot missions and edge systems\", \"review_status\": \"full text reviewed\", \"supports\": [\"Chevron and DataRobot collaboration applies edge agents to support autonomous inspection at Chevron facilities.\", \"Safe Start assesses conditions before/during robotic missions using gas sensors, vision and locally running models; integrates with existing operating systems and NVIDIA NIM.\", \"Vendor and Chevron program manager describe reduced manual point-in-time checks and human oversight/control; no measured outcomes.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-002\"}], \"field_sources\": {\"business_problem\": [\"R015\"], \"workflow\": [\"R015\"], \"technology_data\": [\"R015\"], \"integration_human_role\": [\"R015\"], \"deployment_scope\": [\"R015\"], \"economics\": []}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-002\", \"source_ids\": [\"R015\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-075\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-075", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-076", "Company": "Shell", "Company_primary": "Shell", "Segment": "Supermajor", "HQ": "UK", "Theme": "Operational AI", "Initiative": "Agentic root-cause analysis in global reliability program", "Extracted_finding": "Extends predictive maintenance beyond anomaly detection toward analysis and remediation of reliability issues across asset operations.", "Context": "Shell needs to identify equipment anomalies and determine causes/remediation across a large distributed asset base.", "Implementation": "C3 AI reliability platform monitors equipment for predictive maintenance; new agreement will add agent-based root-cause analysis and remediation.", "Assessment": "Analyst inference: this is a substantial global reliability program with clearly named predictive-maintenance footprint; agentic remediation scope is only newly contracted.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "New multi-year agreement announced 2026-06-04; implementation planned", "Interpretation_limit": "Expansion announcement is prospective; no results for the added agent functionality. 13,000 equipment coverage and hundreds of millions value are vendor-reported, not independent.", "Source_IDs": "RS013; R016", "Source_date": "2026-06-04", "Source_type": "partner case study", "Source_locator": "News Details, lines 285–297", "Source_URL": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-06-04; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Production & field operations"], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-016", "Initiative_ID": "AI-076", "Title": "Shell extends C3 AI with agent-based root-cause analysis", "Date": "2026-06", "Finding": "Shell signed a new multi-year agreement extending its C3 AI Reliability program, which has monitored equipment across its assets since 2018. The expansion adds AI agent-based root-cause analysis and remediation on top of anomaly detection, running on Microsoft Azure.", "Metric": "13,000+ pieces of equipment monitored; program since 2018", "Source_IDs": "R016", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Mature predictive-maintenance estates are the natural base for agents: move from 'alert' to 'diagnose and recommend fix.'", "Partners": "C3 AI, Microsoft Azure", "Original_date": "2026-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-016"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-06-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["majors-018"], "Research_search": " {\"key\": \"shell-agentic-reliability\", \"existing_ids\": [\"AI-076\"], \"company\": \"Shell\", \"title\": \"Agentic root-cause analysis in global reliability program\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"Production & field operations\"], \"category\": \"Internal adoption\", \"stage\": \"New multi-year agreement announced 2026-06-04; implementation planned\", \"practical_relevance\": \"Extends predictive maintenance beyond anomaly detection toward analysis and remediation of reliability issues across asset operations.\", \"business_problem\": \"Shell needs to identify equipment anomalies and determine causes/remediation across a large distributed asset base.\", \"workflow\": \"C3 AI reliability platform monitors equipment for predictive maintenance; new agreement will add agent-based root-cause analysis and remediation.\", \"technology_data\": \"C3 AI Reliability and Agentic AI Platform on Microsoft Azure; partner says historical deployment monitors >13,000 equipment items.\", \"integration_human_role\": \"Decision rights and human review in agent RCA are not disclosed.\", \"deployment_scope\": \"Global operations; multi-year expansion. Existing predictive maintenance ran since 2018; new agentic scope announced for future rollout.\", \"partners\": [\"C3 AI\", \"Microsoft Azure\"], \"disclosed_results\": [], \"economics\": \"C3 AI executive claims hundreds of millions of dollars in economic value, but basis and Shell validation are not supplied.\", \"timeline\": [{\"date\": \"2026-06-04\", \"event\": \"New multi-year expansion agreement announced; agentic RCA planned.\", \"source_ids\": [\"RS013\"]}], \"limitations\": [\"Expansion announcement is prospective; no results for the added agent functionality.\", \"13,000 equipment coverage and hundreds of millions value are vendor-reported, not independent.\"], \"assessment\": \"Analyst inference: this is a substantial global reliability program with clearly named predictive-maintenance footprint; agentic remediation scope is only newly contracted.\", \"sources\": [{\"id\": \"RS013\", \"title\": \"C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations\", \"url\": \"https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai\", \"publisher\": \"C3 AI\", \"date\": \"2026-06-04\", \"accessed\": \"2026-09-27\", \"type\": \"partner case study\", \"locator\": \"News Details, lines 285–297\", \"review_status\": \"full text reviewed\", \"supports\": [\"C3 AI says it has worked with Shell since 2018 on predictive maintenance monitoring over 13,000 equipment items.\", \"A new multi-year agreement announced June 2026 extends reliability AI and introduces agent-based root cause analysis and remediation; the agreement is prospective.\", \"Program uses C3 AI Reliability and C3 Agentic AI Platform on Microsoft Azure.\", \"Hundreds of millions of dollars value is a C3 AI executive claim, not independent or Shell-attributed validation.\"], \"original_id\": \"majors-018\"}], \"field_sources\": {\"business_problem\": [\"RS013\"], \"workflow\": [\"RS013\"], \"technology_data\": [\"RS013\"], \"integration_human_role\": [\"RS013\"], \"deployment_scope\": [\"RS013\"], \"economics\": [\"RS013\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"Official Shell/C3 partner disclosure corroborates AI-076 with vendor evidence and reveals prior predictive maintenance footprint; classify the latest agentic scope as prospective.\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-018\", \"source_ids\": [\"RS013\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-076\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "C3 AI and Shell expand root-cause analysis across global asset operations (2026).", "reason": "Existing record links BusinessWire; official Shell announcement not found/opened in pass. Secondary release needs verification; listed official Shell AI pages establish generic Shell e/model platform but not this named expansion.", "source_url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-076", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "C3 AI’s June 4, 2026 release confirms existing Shell predictive maintenance on 13,000+ pieces of equipment, deployed on Azure, and says the new multi-year agreement will introduce agent-based root cause analysis/remediation. Do not describe the new agentic component as already deployed; “hundreds of millions” is C3 AI executive-attributed, not independently validated.", "correction": "", "source_url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-077", "Company": "TotalEnergies", "Company_primary": "TotalEnergies", "Segment": "Supermajor", "HQ": "France", "Theme": "Operational AI", "Initiative": "Frontier AI models for reservoir exploration and engineering", "Extracted_finding": "TotalEnergies and Mistral announced a three-year program to develop frontier models for geoscience/reservoir workflows. This is existing disclosure REF-021, not a new project.", "Context": "Subsurface experts need to interpret extensive seismic, geological and reservoir data and consider multiple development scenarios for prospects and existing projects.", "Implementation": "Joint scientific lab will develop models to support exploration, reservoir characterization and development, generate multiple scenarios for exploration opportunities, and help optimize and extend existing projects.", "Assessment": "Analyst inference: Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.", "Reported_metric": "Announced three-year joint development program representing more than €100m investment; no deployed-model result or realized return disclosed.", "Stage_at_disclosure": "Three-year joint development program announced; no field deployment or return disclosed.", "Interpretation_limit": "Joint investment allocation not stated. Source says “next to 10 petaflops of data,” dimensionally invalid as data-volume measure; omit from numeric comparison. Program announced, not deployed/benchmarked/monetized.", "Source_IDs": "R021; OCT5-S045", "Source_date": "2026-09-15", "Source_type": "Operator press release", "Source_locator": "Release body, “Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data”", "Source_URL": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "Evidence_status": "Relevant original operator release independently reviewed on 2026-10-06.", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-09-15; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-021", "Initiative_ID": "AI-077", "Title": "TotalEnergies and Mistral launch €100M frontier subsurface AI program", "Date": "2026-09", "Finding": "TotalEnergies and Mistral AI announced a three-year joint program worth more than €100M to build frontier, agentic AI models for geoscientists in exploration, reservoir characterization and development. A joint lab pairs TotalEnergies subsurface experts with Mistral ML researchers on about 10 petaflops of data.", "Metric": ">€100M over 3 years", "Source_IDs": "R021", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "European majors are backing sovereign model vendors to co-train domain models on proprietary subsurface data.", "Partners": "Mistral AI", "Original_date": "2026-09", "Related_IDs": []}, {"ID": "OCT5-AI-077-1", "Initiative_ID": "AI-077", "Title": "TotalEnergies and Mistral announce a three-year frontier subsurface AI program", "Date": "2026-09-15", "Finding": "TotalEnergies and Mistral announced a three-year joint program representing more than €100m investment to develop frontier AI models for exploration, reservoir characterization and development. A joint scientific laboratory is planned. This duplicates linked disclosure REF-021 in the current record; update the existing case summary only.", "Metric": "Announced joint-program investment >€100m over three years; no deployed-model outcome or return reported.", "Source_IDs": "OCT5-S045", "Source_URL": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Related_IDs": [], "Reference_IDs": ["REF-021"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-09-15", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["oct6-executive-ai-077", "majors-017"], "Research_search": "{\"packet_id\": \"oct6-executive-ai-077\", \"review_date\": \"2026-10-06\", \"title\": \"TotalEnergies and Mistral announce a three-year frontier subsurface AI program\", \"publisher\": \"TotalEnergies\", \"source_title\": \"TotalEnergies and Mistral announce a three-year frontier subsurface AI program\", \"source_date\": \"2026-09-15\", \"source_type\": \"Operator press release\", \"source_locator\": \"Release body, “Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data”\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"TotalEnergies and Mistral plan to build frontier models for exploration, reservoir characterization and development through a joint scientific laboratory.\", \"technology_data\": \"The program combines AI, the operator’s geoscience expertise and subsurface data. Model benchmarks, field deployment and partner budget allocations have not been disclosed.\", \"deployment_scope\": \"Announced joint development program; deployment and partner budget allocations not reported.\", \"practical_relevance\": \"This is a commitment to build subsurface models. Track validation against existing methods and field deployment before assuming an exploration advantage.\", \"business_problem\": \"Developing models that can help interpret subsurface data and inform reservoir decisions.\", \"economics\": \"Not cash already spent or realized benefit; investment allocation between partners is unstated.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Existing linked disclosure; not a new project/profile. Exclude dimensionally invalid “petaflops of data” metric.\"], \"timeline_event\": \"Joint program announced 2026-09-15; model deployment date not established.\", \"timeline_date\": \"2026-09-15\", \"source_ids\": [\"R021\"], \"stage\": \"Three-year joint development program announced; no field deployment or return disclosed.\", \"key\": \"oct6-executive-ai-077\", \"record_ids\": [\"AI-077\"], \"company\": \"TotalEnergies\", \"category\": \"Internal adoption\", \"department\": \"Exploration & geoscience\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"R021\"], \"workflow\": [\"R021\"], \"technology_data\": [\"R021\"], \"deployment_scope\": [\"R021\"], \"practical_relevance\": [\"R021\"], \"economics\": [\"R021\"], \"integration_human_role\": [\"R021\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \">€100m investment commitment over three years; no deployed-model result or return.\", \"period\": \"2026-09-15\", \"attribution\": \"Operator account\", \"source_ids\": [\"R021\"]}], \"timeline\": [{\"date\": \"2026-09-15\", \"event\": \"Joint program announced 2026-09-15; model deployment date not established.\", \"source_ids\": [\"R021\"]}], \"sources\": [{\"id\": \"R021\", \"title\": \"TotalEnergies and Mistral launch €100M frontier subsurface AI program\", \"url\": \"https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng\", \"date\": \"2026-09-15\", \"type\": \"TotalEnergies\", \"publisher\": \"TotalEnergies\", \"locator\": \"Release body, “Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data”\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-077", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Existing REF-021 already records this announcement. Do not duplicate profile. >€100m is announced joint program investment, not realized benefit; omit dimensionally invalid petaflops-as-data-volume phrase.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Relevant original operator release independently reviewed on 2026-10-06.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-078", "Company": "TotalEnergies", "Company_primary": "TotalEnergies", "Segment": "Supermajor", "HQ": "France", "Theme": "Operational AI", "Initiative": "Industrial data and AI foundation across operated upstream", "Extracted_finding": "TotalEnergies says an AspenTech Inmation real-time data layer and Cognite contextualization layer are being deployed across 40 upstream and 16 refining/chemical sites, with a tenfold asset data-point objective and full deployment targeted for end-2027. Integrated Power is already done; AI is a later layer. India target is at least 500 engineers across HSE, digital plant and Integrated Power modeling programs.", "Context": "Industrial operations need contextualized, accessible sensor and engineering data to monitor assets, improve decisions and deploy AI solutions.", "Implementation": "AspenTech Inmation collects real-time data; Cognite transforms, enriches, stores and exposes it. The AI layer is planned after the data layer. Integrated Power implementation was already complete at disclosure.", "Assessment": "Analyst inference: Large, concrete industrial data foundation and deployment scope; business outcome remains an objective.", "Reported_metric": "Target scope: 40 upstream and 16 refining/chemical sites; 10x asset data points; full deployment end-2027. India center: at least 500 engineers by 2027 across all named programs. No quantified AI outcome.", "Stage_at_disclosure": "Signed data-platform rollout in progress; end-2027 full-deployment objective.", "Interpretation_limit": "Deployment target and enabling data foundation; not 56 completed AI deployments. The 500-engineer target spans multiple programs and is not AI-only. No costs, acceptance criteria, or realized return disclosed.", "Source_IDs": "RS004; RS005; R024; OCT5-S046", "Source_date": "2026-02-11", "Source_type": "Operator results-call transcript", "Source_locator": "PDF pp. 9–10, transcript lines 358–385", "Source_URL": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf", "Evidence_status": "Relevant original operator transcript independently reviewed on 2026-10-06.", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-02; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Asset-wide data and AI integration", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-024", "Initiative_ID": "AI-078", "Title": "Industrial AI across operated upstream assets", "Date": "2025-09-26", "Finding": "TotalEnergies and Cognite announced a three-year deployment of industrial data and AI capabilities across all TotalEnergies-operated upstream assets worldwide, spanning drilling through production.", "Metric": "Three-year program / all operated upstream assets", "Source_IDs": "R024", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "A contextualized industrial data layer comes before agents; budget for it explicitly.", "Partners": "Cognite (Data Fusion)", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "OCT5-AI-078-1", "Initiative_ID": "AI-078", "Title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "Date": "2026-02-11", "Finding": "Management says a signed AspenTech contract will deploy Inmation for real-time collection across 40 upstream and 16 refining/chemical sites; integrated power was already done. Cognite will transform, enrich, store and expose the data. Management aims to multiply asset data points by ten, with full deployment targeted by end-2027; a later AI layer is intended to improve production and availability. At least 500 engineers by 2027 is a critical-mass target for all three programs (digital HSE, digital plant and Integrated Power modelling).", "Metric": "Planned data-layer scope: 40 upstream + 16 refining/chemical sites; 10x data-point objective; end-2027 deployment target. India competency-center target: ≥500 engineers by 2027 across programs.", "Source_IDs": "OCT5-S046", "Source_URL": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Related_IDs": [], "Reference_IDs": ["REF-024"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai-078", "majors-004"], "Research_search": "{\"packet_id\": \"oct6-executive-ai-078\", \"review_date\": \"2026-10-06\", \"title\": \"TotalEnergies 2025 Results & 2026 Objectives presentation transcript\", \"publisher\": \"TotalEnergies\", \"source_title\": \"TotalEnergies 2025 Results & 2026 Objectives presentation transcript\", \"source_date\": \"2026-02-11\", \"source_type\": \"Operator results-call transcript\", \"source_locator\": \"PDF pp. 9–10 / lines 358–385\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"TotalEnergies is rolling out real-time asset-data collection and a common data layer, with a subsequent AI layer intended to improve production and availability.\", \"technology_data\": \"AspenTech Inmation collects real-time data; Cognite transforms, enriches, stores and exposes it. Management targets ten times as many asset data points and full deployment by end-2027.\", \"deployment_scope\": \"Signed platform rollout underway; full-deployment target end-2027; not completed AI deployment.\", \"practical_relevance\": \"This shows the data work behind a large operator’s AI plans. Compare data capture, interoperability and asset rollout; the number of sites does not measure AI value.\", \"business_problem\": \"Making operating data accessible across 40 upstream and 16 refining and chemical sites.\", \"economics\": \"No cost or realized return disclosed.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Enabling data layer; figures are targets. The 500-engineer goal covers HSE, digital plant and Integrated Power modelling, not AI-only.\"], \"timeline_event\": \"February 2026 call describes ongoing rollout and end-2027 target.\", \"timeline_date\": \"2026-02-11\", \"source_ids\": [\"OCT5-S046\"], \"stage\": \"Signed data-platform rollout in progress; end-2027 full-deployment objective.\", \"key\": \"oct6-executive-ai-078\", \"record_ids\": [\"AI-078\"], \"company\": \"TotalEnergies\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S046\"], \"workflow\": [\"OCT5-S046\"], \"technology_data\": [\"OCT5-S046\"], \"deployment_scope\": [\"OCT5-S046\"], \"practical_relevance\": [\"OCT5-S046\"], \"economics\": [\"OCT5-S046\"], \"integration_human_role\": [\"OCT5-S046\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \"No realized AI result reported; site counts/data expansion/headcount are management objectives.\", \"period\": \"2026-02-11\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S046\"]}], \"timeline\": [{\"date\": \"2026-02-11\", \"event\": \"February 2026 call describes ongoing rollout and end-2027 target.\", \"source_ids\": [\"OCT5-S046\"]}], \"sources\": [{\"id\": \"OCT5-S046\", \"title\": \"TotalEnergies 2025 Results & 2026 Objectives presentation transcript\", \"url\": \"https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf\", \"date\": \"2026-02-11\", \"type\": \"Operator results-call transcript\", \"publisher\": \"TotalEnergies\", \"locator\": \"PDF pp. 9–10 / lines 358–385\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-078", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "TotalEnergies’ Sep 26, 2025 announcement plans three-year scale-out of Cognite data/AI across all operated upstream assets. It describes scope and intended value chain but is an expansion/scale-up announcement, not proof that worldwide rollout was already complete.", "correction": "", "source_url": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Scope/technology implementation targets only; not realized AI results. 500-engineer India center is for all three programs, not AI-only.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Relevant original operator transcript independently reviewed on 2026-10-06.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-079", "Company": "Petronas", "Company_primary": "PETRONAS", "Segment": "NOC", "HQ": "Malaysia", "Theme": "Enterprise AI", "Initiative": "Microsoft collaboration on agentic AI and Copilot", "Extracted_finding": "PETRONAS and Microsoft signed an MoU to explore agentic AI, Copilot, analytics, cloud and cybersecurity across the value chain, supported by the Malaysia West cloud region.", "Context": "The collaboration connects PETRONAS digital transformation with Malaysia’s local AI and cloud ecosystem.", "Implementation": "The partners will investigate applications using Microsoft’s Malaysia West cloud region and pursue joint ecosystem programs.", "Assessment": "Broad partnership intent with limited operating detail. Specific applications and adoption evidence remain necessary to assess execution.", "Reported_metric": "No deployment or financial metric disclosed", "Stage_at_disclosure": "Memorandum of understanding / exploration", "Interpretation_limit": "The release describes areas to explore. It does not establish a completed rollout, user count or realized benefit.", "Source_IDs": "R031", "Source_date": "2025-07-31", "Source_type": "Petronas", "Source_locator": "See source and disclosure timeline", "Source_URL": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "Evidence_status": "Primary source text reviewed", "Review_tier": "Reviewed source", "Time_scope": "Within confirmed 24-month window", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-031", "Initiative_ID": "AI-079", "Title": "Microsoft collaboration on agentic AI and Copilot", "Date": "2025-07-31", "Finding": "PETRONAS and Microsoft signed an MoU to explore agentic AI, Copilot, analytics, cloud and cybersecurity across the value chain, supported by the Malaysia West cloud region.", "Metric": "No deployment or financial metric disclosed", "Source_IDs": "R031", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "In-country cloud regions are unlocking Copilot/agent rollouts at Asian NOCs.", "Partners": "Microsoft Copilot, Azure (Malaysia West)", "Original_date": "2025-07", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-031"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-07-31", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Microsoft MoU proves agentic AI/Copilot deployment across PETRONAS value chain.", "reason": "Primary release says partners will explore initiatives and technologies; deployment scope/results are not described. AIngineer is separately profiled as a product/use-case collaboration.", "source_url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem"}], "Audit_review": {"id": "AI-079", "company": "Petronas", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem"], "findings": [{"severity": "medium", "field": "stage", "issue": "PETRONAS/Microsoft source is an ecosystem collaboration/MoU; it does not itself prove Copilot or agent deployments at PETRONAS.", "correction": "Keep as partnership/strategy; do not call operational adoption.", "source_url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Primary source text reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-080", "Company": "Petronas", "Company_primary": "PETRONAS", "Segment": "NOC", "HQ": "Malaysia", "Theme": "Enterprise AI", "Initiative": "Agentic AI enhancement to Malaysia’s myPROdata upstream investment platform", "Extracted_finding": "The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.", "Context": "Help investors and industry participants navigate Malaysia’s upstream opportunity data and perform faster E&P screening/evaluation.", "Implementation": "Agentic AI is intended to connect and analyze geological, geophysical, engineering and other datasets, supporting prospective-area discovery, resource assessment, development evaluation and commerciality analysis.", "Assessment": "Analyst inference: The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Agreement signed Sept. 3, 2026; platform enhancement/rollout planned", "Interpretation_limit": "Agreement and intended capabilities are disclosed; production deployment, usage and achieved decision speed are not yet reported. Investment and cycle-time goals are national upstream targets, not attribution to the myPROdata AI enhancement.", "Source_IDs": "RS032; R032", "Source_date": "2026-09-03", "Source_type": "company announcement", "Source_locator": "Agreement, platform workflow and MPM targets", "Source_URL": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-09-03; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-032", "Initiative_ID": "AI-080", "Title": "Agentic AI added to myPROdata upstream platform", "Date": "2026-09", "Finding": "Petronas, through Malaysia Petroleum Management, partnered with Iraya Energies to add agentic AI to its myPROdata upstream data platform so it can autonomously analyze and connect G&G and engineering data for investors. The aim is to speed E&P evaluation and support RM50-60B a year of upstream investment.", "Metric": "RM50-60B/yr investment target; discovery-to-first-oil goal cut from 100 to 50 months", "Source_IDs": "R032", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Agentic AI is being pointed outward at investors and partners; data-room agents are a monetizable NOC use case.", "Partners": "Iraya Energies", "Original_date": "2026-09", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-032"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-09-03", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["nocs-018"], "Research_search": " {\"key\": \"petronas-myprodata-agentic\", \"existing_ids\": [\"AI-080\"], \"related_existing_ids\": [\"AI-050\"], \"company\": \"PETRONAS\", \"title\": \"Agentic AI enhancement to Malaysia’s myPROdata upstream investment platform\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Agreement signed Sept. 3, 2026; platform enhancement/rollout planned\", \"practical_relevance\": \"The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.\", \"business_problem\": \"Help investors and industry participants navigate Malaysia’s upstream opportunity data and perform faster E&P screening/evaluation.\", \"workflow\": \"Agentic AI is intended to connect and analyze geological, geophysical, engineering and other datasets, supporting prospective-area discovery, resource assessment, development evaluation and commerciality analysis.\", \"technology_data\": \"Existing web myPROdata platform with subsurface/surface data; Iraya Energies to integrate the agentic-AI capability and establish technical, commercial and security foundations.\", \"integration_human_role\": \"Investors remain decision makers; public release says agents analyze and connect data but does not define human approval controls or autonomous investment decisions.\", \"deployment_scope\": \"Malaysia upstream data platform offered to investors and industry; agreement signed, broad rollout not yet evidenced.\", \"partners\": [\"Iraya Energies\"], \"disclosed_results\": [], \"economics\": \"RM50-60 billion/year upstream investment and reducing discovery-to-first-hydrocarbon timeline from 100 to 50 months are PETRONAS/MPM national ambitions, not project-specific AI savings or outcomes.\", \"timeline\": [{\"date\": \"2026-09-03\", \"event\": \"MPM/PETRONAS signed an agreement with Iraya Energies to add agentic AI to myPROdata.\", \"source_ids\": [\"RS032\"]}], \"limitations\": [\"Agreement and intended capabilities are disclosed; production deployment, usage and achieved decision speed are not yet reported.\", \"Investment and cycle-time goals are national upstream targets, not attribution to the myPROdata AI enhancement.\"], \"assessment\": \"Analyst inference: The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.\", \"sources\": [{\"id\": \"RS032\", \"title\": \"PETRONAS Advances Malaysia’s Upstream Data Platform with Agentic AI to Accelerate E&P Investment\", \"url\": \"https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep\", \"publisher\": \"PETRONAS\", \"date\": \"2026-09-03\", \"type\": \"company announcement\", \"locator\": \"Agreement, platform workflow and MPM targets\", \"review_status\": \"full text reviewed\", \"supports\": [\"PETRONAS/MPM signed an agreement with Iraya Energies to add agentic AI to myPROdata for Malaysia exploration and production data evaluation.\", \"Iraya is lead integrator; the release states investment and cycle-time targets as broader national ambitions, not achieved product outcomes.\"], \"original_id\": \"petronas_myprodata_ai\"}], \"field_sources\": {\"business_problem\": [\"RS032\"], \"workflow\": [\"RS032\"], \"technology_data\": [\"RS032\"], \"integration_human_role\": [\"RS032\"], \"deployment_scope\": [\"RS032\"], \"economics\": [\"RS032\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"nocs\", \"packet_id\": \"nocs-018\", \"source_ids\": [\"RS032\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-080\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-080", "company": "Petronas", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep"], "findings": [{"severity": "medium", "field": "stage", "issue": "PETRONAS release supports agentic-AI enhancement to myPROdata, but deployment/production results and investment impact are not quantified.", "correction": "Report platform development/rollout language as disclosed; omit inferred economic return.", "source_url": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-081", "Company": "PetroChina / CNPC", "Company_primary": "PetroChina", "Segment": "NOC", "HQ": "China", "Theme": "Enterprise AI", "Initiative": "Kunlun large model with value-chain applications", "Extracted_finding": "The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.", "Context": "Apply a petroleum-domain language/model platform to analysis and tasks across exploration, development, refining and planning.", "Implementation": "2026 example: crude procurement question answered with an evaluation, processing plan and risk notes in five minutes versus several hours or days of manual analysis. CNPC describes autonomous planning/task decomposition, tool execution, calculation, prediction and explainability capabilities.", "Assessment": "Analyst inference: The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.", "Reported_metric": "crude purchase evaluation cycle: 5 minutes vs hours or days (Company-reported); deployed scenarios: 152 (Company-reported)", "Stage_at_disclosure": "70B model and 43 petroleum scenarios in 2024; 2026 update reports 152 deployed scenarios", "Interpretation_limit": "2024 launch predates window start; included only to establish baseline, while 2025-26 capability and scale updates fall inside the window. 2026 source reviewed in full.", "Source_IDs": "RS019; RS020; R033", "Source_date": "2026-05-29", "Source_type": "annual report", "Source_locator": "Launch of 70-billion-parameter Kunlun Large Model", "Source_URL": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-05-29; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-033", "Initiative_ID": "AI-081", "Title": "CNPC releases 300B-parameter Kunlun energy model", "Date": "2025-05", "Finding": "CNPC released an upgraded 300B-parameter Kunlun large model, which it described as the first large model approved by regulators in China's energy and chemical sector (approved August 2024). The industry LLM scaled from 70B to 300B parameters, alongside larger vision (4.4B) and multimodal (80B) models.", "Metric": "LLM 70B to 300B; vision 300M to 4.4B; multimodal 16B to 80B", "Source_IDs": "R033", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Chinese NOCs are building sovereign domain foundation models; Western vendors are largely shut out of that stack.", "Partners": "China Mobile, Huawei, iFlytek", "Original_date": "2025-05", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-033"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-29", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["nocs-007"], "Research_search": " {\"key\": \"cnpc-kunlun\", \"existing_ids\": [\"AI-081\"], \"company\": \"PetroChina\", \"title\": \"Kunlun large model with value-chain applications\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"70B model and 43 petroleum scenarios in 2024; 2026 update reports 152 deployed scenarios\", \"practical_relevance\": \"The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.\", \"business_problem\": \"Apply a petroleum-domain language/model platform to analysis and tasks across exploration, development, refining and planning.\", \"workflow\": \"2026 example: crude procurement question answered with an evaluation, processing plan and risk notes in five minutes versus several hours or days of manual analysis. CNPC describes autonomous planning/task decomposition, tool execution, calculation, prediction and explainability capabilities.\", \"technology_data\": \"Company-developed platform, grown from 70B parameters in 2024 to 300B by May 2025, per official reports; joint development with China Mobile, Huawei and iFLYTEK. Uses sector-specific data, AI middleware and computing environment.\", \"integration_human_role\": \"Domain teams apply outputs; example comes from a refinery planning manager. Governance, model boundaries and human review are not detailed in reviewed disclosures.\", \"deployment_scope\": \"CNPC says 152 scenarios across the value chain are in scaled application by May 2026; scenario-level deployment counts not given.\", \"partners\": [\"China Mobile\", \"Huawei\", \"iFLYTEK\"], \"disclosed_results\": [{\"metric\": \"crude purchase evaluation cycle\", \"value\": \"5 minutes vs hours or days\", \"period\": \"May 2026 example\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS020\"]}, {\"metric\": \"deployed scenarios\", \"value\": \"152\", \"period\": \"May 2026\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS020\"]}], \"economics\": \"No consolidated financial impact disclosed. One refinery manager says crude assessment shortened from hours/days to five minutes; not independently validated.\", \"timeline\": [{\"date\": \"2024-08\", \"event\": \"Kunlun launched (outside window; baseline detail in 2024 annual report)\", \"source_ids\": [\"RS019\"]}, {\"date\": \"2025-05\", \"event\": \"300B parameter milestone per CNPC announcement; report source documents current scale-up\", \"source_ids\": [\"RS020\"]}, {\"date\": \"2026-05-28\", \"event\": \"152 application scenarios and higher-order AI capabilities announced\", \"source_ids\": [\"RS020\"]}], \"limitations\": [\"2024 launch predates window start; included only to establish baseline, while 2025-26 capability and scale updates fall inside the window. 2026 source reviewed in full.\"], \"assessment\": \"Analyst inference: The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.\", \"sources\": [{\"id\": \"RS019\", \"title\": \"2024 Annual Report\", \"url\": \"https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf\", \"publisher\": \"CNPC\", \"date\": \"2025-08-01\", \"type\": \"annual report\", \"locator\": \"Launch of 70-billion-parameter Kunlun Large Model\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"CNPC reports launching a 70-billion-parameter Kunlun model in 2024 with 43 professional/general scenarios for petroleum.\", \"Report names a joint development group of CNPC, China Mobile, Huawei and iFLYTEK; platform included eight trained large models, 18 application scenarios, specialized datasets, AI middleware and computing environment.\"], \"original_id\": \"cnpc_annual\"}, {\"id\": \"RS020\", \"title\": \"中国石油昆仑大模型迈入主动智能新阶段\", \"url\": \"https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml\", \"publisher\": \"CNPC News Center\", \"date\": \"2026-05-29\", \"type\": \"company announcement\", \"locator\": \"Release overview and application examples\", \"review_status\": \"full text reviewed\", \"supports\": [\"CNPC says Kunlun reached 152 deployed scenarios across the energy and chemical value chain in May 2026.\", \"Specific example: crude purchase evaluation with processing plans and risk advice in five minutes versus hours or days of manual analysis, as reported by a refinery manager.\"], \"original_id\": \"cnpc_2026\"}], \"field_sources\": {\"business_problem\": [\"RS019\", \"RS020\"], \"workflow\": [\"RS019\", \"RS020\"], \"technology_data\": [\"RS019\", \"RS020\"], \"integration_human_role\": [\"RS019\", \"RS020\"], \"deployment_scope\": [\"RS019\", \"RS020\"], \"economics\": [\"RS019\", \"RS020\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-007\", \"source_ids\": [\"RS019\", \"RS020\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-081\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Direct 2026 article fetch failed (encoding/internal error); an official search-index excerpt and official 2025 annual report were reviewed. More operating-unit detail and independent validation remain undisclosed.", "source_url": ""}], "Audit_review": {"id": "AI-081", "company": "PetroChina / CNPC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml"], "findings": [{"severity": "medium", "field": "stage / metrics", "issue": "CNPC’s official portal search-indexed article describes 152 Kunlun AI scenarios as scaled/deployed and reports selected speed/cost/accuracy results; the Chinese direct page could not be fetched, so this is excerpt-level verification. The separate 2025 annual report discusses earlier model/scenario scale.", "correction": "Retain reported scale/results as company-reported, but mark 2026 article excerpt-only and do not merge the 2025 model disclosures with later deployment dates.", "source_url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml"}], "residual_gaps": ["Direct 2026 article fetch failed (encoding/internal error); an official search-index excerpt and official 2025 annual report were reviewed. More operating-unit detail and independent validation remain undisclosed."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and the deployed-scenario count was located in the source.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-082", "Company": "CNOOC", "Company_primary": "CNOOC", "Segment": "NOC", "HQ": "China", "Theme": "Operational AI", "Initiative": "Offshore predictive maintenance cuts failures ~20%", "Extracted_finding": "CNOOC deployed AI predictive maintenance across offshore platforms including Bohai Oilfield, using real-time sensor data from subsea pumps and turbines. China Daily reported roughly a 20% cut in critical equipment failure rates; CNPC's Changqing field reported ~30% shorter exploration cycles and 10%+ higher per-well output.", "Context": "CNOOC deployed AI predictive maintenance across offshore platforms including Bohai Oilfield, using real-time sensor data from subsea pumps and turbines. China Daily reported roughly a 20% cut in critical equipment failure rates; CNPC's Changqing field reported ~30% shorter exploration cycles and 10%+ higher per-well output.", "Implementation": "Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Rotating-equipment predictive maintenance remains the most consistently quantified AI win across all regions.", "Reported_metric": "~20% fewer critical equipment failures (CNOOC, as reported by China Daily). The ~30% shorter exploration cycle and >10% per-well output in the same article belong to CNPC's Changqing field (PetroChina group), not CNOOC, and are not attributed to this profile.", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Imported claim and interpretation from the supplied reference. Source text, date, scope and outcome attribution require independent verification.", "Source_IDs": "R034", "Source_date": "2026-04", "Source_type": "China Daily", "Source_locator": "See source and disclosure timeline", "Source_URL": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Reference date within window; date not independently verified", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-034", "Initiative_ID": "AI-082", "Title": "Offshore predictive maintenance cuts failures ~20%", "Date": "2026-04", "Finding": "CNOOC deployed AI predictive maintenance across offshore platforms including Bohai Oilfield, using real-time sensor data from subsea pumps and turbines. China Daily reported roughly a 20% cut in critical equipment failure rates; CNPC's Changqing field reported ~30% shorter exploration cycles and 10%+ higher per-well output.", "Metric": "~20% fewer critical equipment failures (CNOOC); ~30% shorter exploration cycle, >10% per-well output (CNPC Changqing)", "Source_IDs": "R034", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Rotating-equipment predictive maintenance remains the most consistently quantified AI win across all regions.", "Partners": "", "Original_date": "2026-04", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-034"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Internal adoption", "Audit_flags": [{"claim": "Offshore predictive maintenance reduces failures by approximately 20%", "reason": "Existing claim cites secondary China Daily coverage; the reviewed official interim results discuss AI+ scenarios but do not substantiate this specific metric. Unverified, not disproved.", "source_url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-082", "company": "CNOOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "realized", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Attribution corrected 5 Oct 2026: Changqing metrics belong to CNPC/PetroChina, not CNOOC. Pending review; secondary source only.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-083", "Company": "Santos", "Company_primary": "Santos", "Segment": "Integrated", "HQ": "Australia", "Theme": "Operational AI", "Initiative": "Integrated production-system model for Eastern Australia and PNG", "Extracted_finding": "Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.", "Context": "Continuous production optimization needs a joined view from reservoir through facility inlet across geographically distributed assets.", "Implementation": "Xecta says Santos implemented an Integrated Production System Model to generate ongoing optimization insights across Eastern Australia and Papua New Guinea; legacy profile also claims expansion from Cooper Basin and five-year agreement, not independently verified in reviewed Xecta index.", "Assessment": "Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Xecta source index lists deployment in June/July 2025; exact release date not available", "Interpretation_limit": "Official vendor news index reviewed; exact article body and date inaccessible. Claims about initial Cooper Basin rollout, 5-year contract and Kelvin.ai autonomous control not reverified.", "Source_IDs": "RS114; R036", "Source_date": "Undated source; see chronology", "Source_type": "company webpage", "Source_locator": "Santos Eastern Australia and PNG digital oilfield listing", "Source_URL": "https://xecta.com/news-resources/", "Evidence_status": "verified partner", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-036", "Initiative_ID": "AI-083", "Title": "Xecta AI digital twin expanded to East Australia and PNG", "Date": "2025-07", "Finding": "After a successful rollout in the Cooper Basin, Santos signed a five-year agreement with Xecta to build an AI-driven digital twin for production optimization across its Eastern Australia and Papua New Guinea operations. This builds on its 2023 Kelvin.ai agreement for autonomous control of wellsite and facility operations.", "Metric": "5-year agreement", "Source_IDs": "R036", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Mid-size integrateds scale AI by proving value in one basin and then signing enterprise-wide multi-year deals.", "Partners": "Xecta, Kelvin.ai, SIG ML", "Original_date": "2025-07", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-036"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["majors_backfill-012"], "Research_search": " {\"key\": \"santos-xecta\", \"existing_ids\": [\"AI-083\"], \"related_existing_ids\": [], \"company\": \"Santos / Xecta\", \"title\": \"Integrated production-system model for Eastern Australia and PNG\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Xecta source index lists deployment in June/July 2025; exact release date not available\", \"practical_relevance\": \"Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.\", \"business_problem\": \"Continuous production optimization needs a joined view from reservoir through facility inlet across geographically distributed assets.\", \"workflow\": \"Xecta says Santos implemented an Integrated Production System Model to generate ongoing optimization insights across Eastern Australia and Papua New Guinea; legacy profile also claims expansion from Cooper Basin and five-year agreement, not independently verified in reviewed Xecta index.\", \"technology_data\": \"Specific model/data architecture could not be reviewed because linked release was unavailable; Xecta index names the integrated production model.\", \"integration_human_role\": \"Decision rights/control automation are not disclosed by accessible primary page.\", \"deployment_scope\": \"Named deployment in Eastern Australia and PNG, but no operating well/facility counts found.\", \"partners\": [\"Xecta\"], \"disclosed_results\": [], \"economics\": \"No quantitative production/economic results in accessible primary listing.\", \"timeline\": [{\"date\": \"June/July 2025 listing; exact publication date unavailable\", \"event\": \"Integrated production-system model for Eastern Australia and PNG documented or disclosed.\", \"source_ids\": [\"RS114\"]}], \"limitations\": [\"Official vendor news index reviewed; exact article body and date inaccessible.\", \"Claims about initial Cooper Basin rollout, 5-year contract and Kelvin.ai autonomous control not reverified.\"], \"assessment\": \"Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.\", \"sources\": [{\"id\": \"RS114\", \"title\": \"Xecta news and resources index\", \"url\": \"https://xecta.com/news-resources/\", \"publisher\": \"Xecta\", \"date\": \"June/July 2025 listing; exact publication date unavailable\", \"type\": \"company webpage\", \"locator\": \"Santos Eastern Australia and PNG digital oilfield listing\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Xecta index identifies Santos integrated production-system model/digital oilfield across Eastern Australia and PNG.\", \"Underlying dated technical release did not open; current confirmation limited to vendor page index.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-015\"}], \"field_sources\": {\"business_problem\": [\"RS114\"], \"workflow\": [\"RS114\"], \"technology_data\": [\"RS114\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS114\"], \"economics\": []}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-012\", \"source_ids\": [\"RS114\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-083\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [{"claim": "Xecta AI digital twin expanded to East Australia and PNG (2025).", "reason": "Existing source is trade press. Official 2026 investor briefing excerpt reviewed does not independently verify Xecta rollout claim.", "source_url": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta"}, {"claim": "Five-year Xecta contract, Cooper Basin rollout and Kelvin.ai autonomous control are independently confirmed.", "reason": "Xecta official index supports only Santos IPS across Eastern Australia/PNG in accessible material. Original source body unavailable; other details not reverified. This is a gap, not disproval.", "source_url": "https://xecta.com/news-resources/"}], "Audit_review": {"id": "AI-083", "company": "Santos", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://xecta.com/news-resources/", "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-084", "Company": "OMV", "Company_primary": "OMV", "Segment": "Integrated", "HQ": "Austria", "Theme": "Enterprise AI", "Initiative": "AI Hub and enterprise assistants", "Extracted_finding": "Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.", "Context": "Employees need secure tools for document analysis, information retrieval and navigation of internal procedures.", "Implementation": "Central Hub provides chatbots, assistants and agents; GPT assistant retrieves/analyzes documents; Regulations Assistant guides users through internal policies.", "Assessment": "Analyst inference: Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.", "Reported_metric": "Monthly active users: >2,500 employees (Company-reported); GenAI training reach: >10,000 employees (Company-reported)", "Stage_at_disclosure": "AI Hub launched 2025; 2025 annual report reviewed 2026-04-02", "Interpretation_limit": "Employee-use count and project pipeline are company-reported.", "Source_IDs": "R037", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Digitalization; AI Across the Value Chain", "Source_URL": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Enterprise adoption with named workflows", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-037", "Initiative_ID": "AI-084", "Title": "AI Hub and enterprise adoption", "Date": "2026-04-02", "Finding": "OMV’s 2025 report describes an Azure-based AI Hub launched in 2025, with over 2,500 monthly GPT users and over 10,000 staff trained.", "Metric": "50+ projects in use / 40+ in development / 210+ ideas", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "A single AI front door with a visible pipeline (ideas to development to live) makes adoption measurable.", "Partners": "Microsoft Azure, SAP, Salesforce, SLB, TerraAI", "Original_date": "2026-03", "Related_IDs": []}], "Related_IDs": ["AI-132", "AI-133", "AI-134", "AI-135"], "Reference_IDs": ["REF-037"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-008"], "Research_search": " {\"key\": \"omv-ai-hub\", \"existing_ids\": [\"AI-084\"], \"company\": \"OMV\", \"title\": \"AI Hub and enterprise assistants\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"AI Hub launched 2025; 2025 annual report reviewed 2026-04-02\", \"practical_relevance\": \"Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.\", \"business_problem\": \"Employees need secure tools for document analysis, information retrieval and navigation of internal procedures.\", \"workflow\": \"Central Hub provides chatbots, assistants and agents; GPT assistant retrieves/analyzes documents; Regulations Assistant guides users through internal policies.\", \"technology_data\": \"Built on Azure and integrated in OMV digital environment; details of model, retrieval architecture and guardrails not disclosed.\", \"integration_human_role\": \"Employees use outputs for daily work; policies assistant surfaces rules; review remains with staff.\", \"deployment_scope\": \"Group-wide hub; in-house GPT used monthly by >2,500 employees; 10,000+ attended GenAI training.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Monthly active users\", \"value\": \">2,500 employees\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R037\"]}, {\"metric\": \"GenAI training reach\", \"value\": \">10,000 employees\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R037\"]}], \"economics\": \"No directly attributable productivity or financial metric; adoption and training counts reported.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"AI Hub launched; 2025 usage metrics reported in annual report.\", \"source_ids\": [\"R037\"]}], \"limitations\": [\"Employee-use count and project pipeline are company-reported.\"], \"assessment\": \"Analyst inference: Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.\", \"sources\": [{\"id\": \"R037\", \"title\": \"Digitalization — OMV Combined Annual Report 2025\", \"url\": \"https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html\", \"publisher\": \"OMV\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Digitalization; AI Across the Value Chain\", \"review_status\": \"full text reviewed\", \"supports\": [\"OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.\", \"OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.\", \"Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.\", \"iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.\", \"Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-012\"}], \"field_sources\": {\"business_problem\": [\"R037\"], \"workflow\": [\"R037\"], \"technology_data\": [\"R037\"], \"integration_human_role\": [\"R037\"], \"deployment_scope\": [\"R037\"], \"economics\": [\"R037\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-008\", \"source_ids\": [\"R037\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-084\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-084", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "2,500+ monthly active users", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-085", "Company": "Kuwait Oil Company", "Company_primary": "Kuwait Oil Company", "Segment": "NOC", "HQ": "Kuwait", "Theme": "Operational AI", "Initiative": "G Agent AI drilling-rig resource management and scheduling", "Extracted_finding": "This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.", "Context": "Improve rig resource allocation, drilling planning, system coordination and real-time operational decision-making.", "Implementation": "G Agent supports real-time data analysis and rig resource management; KOC separately describes an agentic AI rig-scheduling project that it says improved productivity, planning and operational quality.", "Assessment": "Analyst inference: This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "KOC’s Oct–Dec 2025 digest describes G Agent and an agentic rig-scheduling project as operational initiatives at its AI Innovation Center", "Interpretation_limit": "The KOC company publication confirms the named platform and workflow, but gives no measured results or rollout scope. The bulletin describes G Agent and agentic scheduling; exact technical architecture and whether they are one system are not clear.", "Source_IDs": "RS033; R038", "Source_date": "Undated source; see chronology", "Source_type": "company periodical", "Source_locator": "PDF pp. 14–16, AI Innovation Center / G Agent", "Source_URL": "https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-038", "Initiative_ID": "AI-085", "Title": "AI Innovation Center delivers G Agent drilling win", "Date": "2025-09", "Finding": "KOC opened its first AI Innovation Center in August 2025 with Microsoft, Halliburton and Ghaia.ai. Its first result is an AI-powered rig drilling project on Ghaia.ai's G Agent platform, a human-agent collaboration mesh that improves productivity and planning of drilling rig operations.", "Metric": "No quantified metric supplied", "Source_IDs": "R038", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Co-located innovation centers with a hyperscaler, an OFS firm and an agent startup can get a first production use case within weeks.", "Partners": "Microsoft, Halliburton, Ghaia.ai (G Agent)", "Original_date": "2025-09", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-038"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["nocs-019"], "Research_search": " {\"key\": \"koc-g-agent-drilling\", \"existing_ids\": [\"AI-085\"], \"related_existing_ids\": [], \"company\": \"Kuwait Oil Company\", \"title\": \"G Agent AI drilling-rig resource management and scheduling\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"KOC’s Oct–Dec 2025 digest describes G Agent and an agentic rig-scheduling project as operational initiatives at its AI Innovation Center\", \"practical_relevance\": \"This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.\", \"business_problem\": \"Improve rig resource allocation, drilling planning, system coordination and real-time operational decision-making.\", \"workflow\": \"G Agent supports real-time data analysis and rig resource management; KOC separately describes an agentic AI rig-scheduling project that it says improved productivity, planning and operational quality.\", \"technology_data\": \"AI platform arising from KOC partnership with Microsoft and Halliburton; Microsoft Azure and licenses are part of broader government collaboration; Ghaia AI supports technology-program/training management at the Center.\", \"integration_human_role\": \"KOC says AI agents support decision-making; human role and authorization boundary for rig-scheduling outputs are not specified.\", \"deployment_scope\": \"KOC AI Innovation Center and drilling operations; no rig count, field/site or deployment breadth quantified in the reviewed KOC bulletin.\", \"partners\": [\"Microsoft\", \"Halliburton\", \"Ghaia AI\"], \"disclosed_results\": [], \"economics\": \"No quantified operational result, contract value or AI-specific cost/savings disclosed.\", \"timeline\": [{\"date\": \"2025-10 to 2025-12\", \"event\": \"KOC bulletin describes AI Innovation Center, G Agent and agentic rig scheduling.\", \"source_ids\": [\"RS033\"]}], \"limitations\": [\"The KOC company publication confirms the named platform and workflow, but gives no measured results or rollout scope.\", \"The bulletin describes G Agent and agentic scheduling; exact technical architecture and whether they are one system are not clear.\"], \"assessment\": \"Analyst inference: This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.\", \"sources\": [{\"id\": \"RS033\", \"title\": \"The Kuwait Digest 15 (October–December 2025)\", \"url\": \"https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf\", \"publisher\": \"Kuwait Oil Company\", \"date\": \"2025-Q4\", \"type\": \"company periodical\", \"locator\": \"PDF pp. 14–16, AI Innovation Center / G Agent\", \"review_status\": \"full text reviewed\", \"supports\": [\"KOC says G Agent uses AI for drilling-rig resource management, systems integration and real-time data analysis; it also describes an agentic drilling-rig scheduling project.\", \"KOC names Microsoft and Halliburton partnership and says Ghaia AI manages technology programs/training at the Center.\"], \"original_id\": \"koc_digest\"}], \"field_sources\": {\"business_problem\": [\"RS033\"], \"workflow\": [\"RS033\"], \"technology_data\": [\"RS033\"], \"integration_human_role\": [\"RS033\"], \"deployment_scope\": [\"RS033\"], \"economics\": [\"RS033\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"nocs\", \"packet_id\": \"nocs-019\", \"source_ids\": [\"RS033\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-085\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "AI seismic processing, Jurassic gas AI and G Agent drilling outcomes", "reason": "No official KOC source review completed in this pass to substantiate these entries; the SLB earnings source is vendor disclosure and Kuwait Times is secondary. Retain as reported/unverified pending primary KOC technical or procurement evidence.", "source_url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-085", "company": "Kuwait Oil Company", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-086", "Company": "Devon Energy", "Company_primary": "Devon Energy", "Segment": "Independent E&P", "HQ": "US", "Theme": "Enterprise AI", "Initiative": "ChatDVN internal company-data assistant", "Extracted_finding": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data.", "Context": "Reduce time spent finding company data and turn it into useful analysis.", "Implementation": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.", "Assessment": "The issuer confirms an operating internal workflow. Ask what employees complete through the tool and measure quality and full review effort before crediting a productivity or cash benefit.", "Reported_metric": "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.", "Stage_at_disclosure": "Operating internal workflow; issuer-described daily use", "Interpretation_limit": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status. Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch.", "Source_IDs": "FU001; R039", "Source_date": "2026-05-06", "Source_type": "Issuer-hosted earnings transcript", "Source_locator": "PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI", "Source_URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Evidence_status": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.", "Review_tier": "Reviewed source", "Time_scope": "Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Workforce & learning"], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Internal adoption", "Disclosures": [{"ID": "REF-039", "Initiative_ID": "AI-086", "Title": "ChatDVN internal AI platform plus AI-101 training", "Date": "2025-07", "Finding": "Devon built ChatDVN, a proprietary platform combining AI tools, machine learning and LLMs, and paired it with a company-wide AI-101 course. CTO Trey Lowe cited productivity gains among engineers, software developers and geoscientists, and use cases such as gas-lift optimization and simul-frac planning done in hours. Devon is standardizing wellsite sensors and cameras so AI can scale, including planned video-to-data analytics in 2026.", "Metric": "About 1,000 employees completed AI-101 by end of June 2025; 15-30% productivity gain for engineers, developers and geoscientists; 5% oil uplift at one Delaware gas-lift site", "Source_IDs": "R039", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A branded internal assistant plus a mandatory foundational course gets adoption started faster than rolling out tools on their own.", "Partners": "ChatDVN (in-house)", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "WATCH21EX-AI-086", "Initiative_ID": "AI-086", "Title": "Devon confirms ChatDVN is part of daily work", "Date": "2026-05-06", "Finding": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.", "Metric": "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.", "Source_IDs": "FU001", "Source_URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Evidence": "Issuer-hosted corrected call transcript, relevant original page read", "Origin": "21 October additional Luna review; editor primary-source admission", "Assessment": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status."}], "Related_IDs": [], "Reference_IDs": ["REF-039"], "Provenance": ["User-supplied reference", "21 October issuer-hosted call review; pending status resolved for daily-use claim only"], "Latest_disclosure": "2026-05-06", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["watch21extra-ai-086", "oct6-followup-ai-086", "north_america-002"], "Research_search": "{\"packet_id\": \"watch21extra-ai-086\", \"key\": \"watch21extra-ai-086\", \"record_ids\": [\"AI-086\"], \"company\": \"Devon Energy\", \"title\": \"Devon confirms ChatDVN is part of daily work\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-21\", \"verification\": \"verified primary\", \"stage\": \"Operating internal workflow; issuer-described daily use\", \"business_problem\": \"Reduce time spent finding company data and turn it into useful analysis.\", \"workflow\": \"Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.\", \"technology_data\": \"Firewalled internal AI tool and organized company data. Model provider, retrieval architecture and permissions are not disclosed.\", \"integration_human_role\": \"Engineers act on operating insights. The call does not establish ChatDVN execution rights, approval rules or any autonomous executive authority.\", \"deployment_scope\": \"Internal company-data workflow; user and asset populations undisclosed. Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch.\", \"economics\": \"No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.\", \"practical_relevance\": \"Which jobs do employees finish through ChatDVN, and does the full review process improve time, accuracy and follow-through?\", \"assessment\": \"The issuer confirms an operating internal workflow. Ask what employees complete through the tool and measure quality and full review effort before crediting a productivity or cash benefit.\", \"limitations\": [\"Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status.\", \"No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.\"], \"source_ids\": [\"FU001\"], \"partners\": [], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-05-06\", \"event\": \"Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data.\", \"source_ids\": [\"FU001\"]}], \"sources\": [{\"id\": \"FU001\", \"title\": \"Devon Q1 2026 earnings call — internal ChatDVN workflow\", \"url\": \"https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf\", \"date\": \"2026-05-06\", \"publisher\": \"Devon Energy / FactSet CallStreet\", \"locator\": \"PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI\", \"excerpt\": \"a fully-firewalled internal tool called ChatDVN\", \"access\": \"Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.\", \"type\": \"Issuer-hosted earnings transcript\", \"review_status\": \"Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.\"}], \"field_sources\": {\"business_problem\": [\"FU001\"], \"workflow\": [\"FU001\"], \"technology_data\": [\"FU001\"], \"integration_human_role\": [\"FU001\"], \"deployment_scope\": [\"FU001\"], \"economics\": [\"FU001\"]}}", "Research_status": "Primary daily-use evidence linked 21 October 2026; older interview claims remain separate.", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-086", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "Daily workflow described; no active-user denominator or verified company-wide rollout.", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status.", "Outcome_counted": false, "Outcome_count_note": "Operating use verified from issuer-hosted call; older productivity/training figures are not corroborated by that source and receive no result credit.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.", "claim_status": "Operating daily use described by issuer; no quantified tool-level outcome verified", "independent_validation": "Not established"}, "Current_source_check": "2026-10-21", "Research_origin": ["research/watch-2026-10-21-extra/README.md"]}, {"ID": "AI-087", "Company": "Coterra Energy", "Company_primary": "Coterra Energy", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "ML well prediction informs capital planning", "Extracted_finding": "Named capital-planning application; implementation and value remain unclear.", "Context": "According to Devon's CTO after the merger closed, Coterra had invested heavily in AI and machine-learning models for well prediction that informed its capital plans. Those models are now being combined with Devon's operations-focused AI systems.", "Implementation": "Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Well-performance prediction models used in capital allocation are a high-value, low-integration-risk AI use case that survives M&A.", "Reported_metric": "No quantified metric supplied", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Third-party-hosted transcript; no realized value disclosed.", "Source_IDs": "RS034; R042", "Source_date": "2026-08", "Source_type": "secondary", "Source_locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted", "Source_URL": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-08; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-044", "Initiative_ID": "AI-087", "Title": "ML well-prediction models fed capital planning", "Date": "2026-08", "Finding": "According to Devon's CTO after the merger closed, Coterra had invested heavily in AI and machine-learning models for well prediction that informed its capital plans. Those models are now being combined with Devon's operations-focused AI systems.", "Metric": "No quantified metric supplied", "Source_IDs": "R042", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Well-performance prediction models used in capital allocation are a high-value, low-integration-risk AI use case that survives M&A.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-044"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-08", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-015"], "Research_search": " {\"key\": \"coterra-ml-capital-planning\", \"existing_ids\": [\"AI-087\"], \"company\": \"Coterra Energy\", \"title\": \"ML well prediction informs capital planning\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Reported by Devon after merger close; pre-merger scale and integration state not fully disclosed\", \"practical_relevance\": \"Named capital-planning application; implementation and value remain unclear.\", \"business_problem\": \"Improve well forecasts used in capital choices.\", \"workflow\": \"CEO said ML well predictions inform capital planning.\", \"technology_data\": \"ML named; inputs, validation and architecture undisclosed.\", \"integration_human_role\": \"Capital-planning use is reported; review controls are unknown.\", \"deployment_scope\": \"Legacy Coterra assets; well coverage is undisclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific financial outcome disclosed.\", \"timeline\": [{\"date\": \"2026-08\", \"event\": \"CEO said models were being integrated after the merger.\", \"source_ids\": [\"RS034\"]}], \"limitations\": [\"Third-party-hosted transcript; no realized value disclosed.\"], \"assessment\": \"Distinct planning use from Devon lift controls; integration is ongoing and secondary-reported.\", \"sources\": [{\"id\": \"RS034\", \"title\": \"Devon Energy Q2 2026 Earnings Call Transcript\", \"url\": \"https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html\", \"publisher\": \"Yahoo Finance\", \"date\": \"2026-08\", \"type\": \"secondary\", \"locator\": \"CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"CEO said Coterra ML models support capital planning.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-003\"}], \"field_sources\": {\"business_problem\": [\"RS034\"], \"workflow\": [\"RS034\"], \"technology_data\": [\"RS034\"], \"integration_human_role\": [\"RS034\"], \"deployment_scope\": [\"RS034\"], \"economics\": []}, \"verification\": \"secondary only\", \"corrections_to_existing\": \"AI-087 exactly matches the Coterra well-prediction/capital-planning workflow. This post-close disclosure is secondary-hosted and needs issuer transcript re-verification.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-015\", \"source_ids\": [\"RS034\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-087\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": {"id": "AI-087", "company": "Coterra Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html"], "findings": [{"severity": "high", "field": "source quality", "issue": "Cited Q2 2026 Coterra details are from a Yahoo-transcribed call and the URL returned 429; no issuer transcript corroboration was accessible.", "correction": "Mark claims unverified pending direct Coterra filing/transcript; this access failure is not proof the claim is false.", "source_url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html"}], "residual_gaps": ["Cited secondary transcript returned HTTP 429; claim is not disproved, but issuer-source verification remains outstanding."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-088", "Company": "EOG Resources", "Company_primary": "EOG Resources", "Segment": "Independent E&P", "HQ": "US", "Theme": "Enterprise AI", "Initiative": "Proprietary generative AI for collaboration and operational data", "Extracted_finding": "EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.", "Context": "EOG said it uses 'quite a bit of generative AI' to organize geologic data and find hidden trends, alongside real-time drilling optimization, predictive maintenance and some autonomous field operations. The work builds on almost 10 years of smart centralized gas-lift technology and newer deep-learning tools that capture human observation and experiential learning. EOG said it is 'not quite to true agentic intelligence' and continues to favor building apps in-house next to field operations.", "Implementation": "EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.", "Assessment": "Tie in-house software to proprietary field sensors and downhole tools, and capture expert tacit knowledge as training data.", "Reported_metric": "No AI-specific outcome or rollout count disclosed.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Interpretation_limit": "Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage.", "Source_IDs": "R045; R046; OCT5-S033", "Source_date": "2026 annual report covering 2025; exact publication day not confirmed", "Source_type": "Operator annual report", "Source_locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence_status": "Original source text reviewed on 6 October 2026", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-02-25; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-045", "Initiative_ID": "AI-088", "Title": "Generative AI across geology, drilling and safety", "Date": "2025-11", "Finding": "EOG said it uses 'quite a bit of generative AI' to organize geologic data and find hidden trends, alongside real-time drilling optimization, predictive maintenance and some autonomous field operations. The work builds on almost 10 years of smart centralized gas-lift technology and newer deep-learning tools that capture human observation and experiential learning. EOG said it is 'not quite to true agentic intelligence' and continues to favor building apps in-house next to field operations.", "Metric": "About 10 years of ML-driven gas-lift optimization", "Source_IDs": "R045", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Tie in-house software to proprietary field sensors and downhole tools, and capture expert tacit knowledge as training data.", "Partners": "In-house apps", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "OCT5-AI-088-1", "Initiative_ID": "AI-088", "Title": "Generative AI for geology, drilling and safety", "Date": "2026 annual report covering 2025; exact publication day not confirmed", "Finding": "EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.", "Metric": "No AI-specific outcome or rollout count disclosed.", "Source_IDs": "OCT5-S033", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Related_IDs": [], "Reference_IDs": ["REF-045"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-25", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["oct6-gap-ai-088", "north_america-003"], "Research_search": "{\"packet_id\": \"oct6-gap-ai-088\", \"review_date\": \"2026-10-06\", \"title\": \"Proprietary generative AI for collaboration and operational data\", \"workflow\": \"EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Operator account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage.\"], \"verification\": \"verified primary\", \"source_review\": \"Full original document reviewed on 6 October 2026\", \"source_locator\": \"PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time\", \"timeline_date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-088\", \"record_ids\": [\"AI-088\"], \"company\": \"EOG Resources\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"source_ids\": [\"OCT5-S033\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S033\"], \"workflow\": [\"OCT5-S033\"], \"technology_data\": [\"OCT5-S033\"], \"deployment_scope\": [\"OCT5-S033\"], \"practical_relevance\": [\"OCT5-S033\"], \"economics\": [\"OCT5-S033\"], \"integration_human_role\": [\"OCT5-S033\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"No AI-specific outcome or rollout count disclosed.\", \"period\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S033\"]}], \"timeline\": [{\"date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"OCT5-S033\"]}], \"sources\": [{\"id\": \"OCT5-S033\", \"title\": \"Proprietary generative AI for collaboration and operational data\", \"url\": \"https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf\", \"date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"type\": \"Operator annual report\", \"publisher\": \"EOG Resources\", \"locator\": \"PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time\", \"review_status\": \"Full original document reviewed on 6 October 2026\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Generative AI, production optimizers, remote control deployment.", "reason": "Existing citations are third-party transcript copies; issuer-source corroboration outstanding. Not disproved.", "source_url": "https://investors.eogresources.com/Events-and-Presentations"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-088", "company": "EOG Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence type as a management earnings-call transcript and retain exact speaker-attributed claims.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-089", "Company": "EOG Resources", "Company_primary": "EOG Resources", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "Proprietary machine-learning production optimizer", "Extracted_finding": "EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary.", "Context": "EOG described its proprietary production optimizers program, which uses machine learning to optimize base production for better run time and lower cost. EOG is also rolling automation across acquired Utica operations so wells can be managed remotely from its 24-hour control room.", "Implementation": "EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary. Related August 2026 context: EOG’s August 2026 deck reports 5% improved Utica base performance from its production-optimizer suite after the Encino acquisition. The deck does not attribute that result specifically to AI or machine learning.", "Assessment": "Standardize automation and remote monitoring on acquired assets right away so the ML optimizers have data to work on.", "Reported_metric": "No AI-specific outcome or rollout count disclosed.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Interpretation_limit": "Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage. Related Utica result: Related operating-technology context only. The deck does not define base performance or disclose the well count, measurement period, method, costs or counterfactual. The annual report’s description of machine-learning tools does not establish AI causality for this 5% result.", "Source_IDs": "R046; OCT5-S034; OCT5-S033; OCT10-EOG-UTICA", "Source_date": "2026 annual report covering 2025; exact publication day not confirmed", "Source_type": "Operator annual report", "Source_locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence_status": "Original source text reviewed on 6 October 2026", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-02-25; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-046", "Initiative_ID": "AI-089", "Title": "ML 'production optimizers' and 24-hour control room", "Date": "2026-02", "Finding": "EOG described its proprietary production optimizers program, which uses machine learning to optimize base production for better run time and lower cost. EOG is also rolling automation across acquired Utica operations so wells can be managed remotely from its 24-hour control room.", "Metric": "No quantified metric supplied", "Source_IDs": "R046", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardize automation and remote monitoring on acquired assets right away so the ML optimizers have data to work on.", "Partners": "Production optimizers (in-house)", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "OCT5-AI-089-1", "Initiative_ID": "AI-089", "Title": "ML production optimizers and remote Utica operations", "Date": "2026 annual report covering 2025; exact publication day not confirmed", "Finding": "EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary.", "Metric": "No AI-specific outcome or rollout count disclosed.", "Source_IDs": "OCT5-S034", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT10-AI-089-UTICA", "Initiative_ID": "AI-089", "Title": "Utica production optimizer: related technology result, not AI attribution", "Date": "2026-08-04", "Finding": "EOG’s August 2026 deck reports 5% improved Utica base performance from its production-optimizer suite after the Encino acquisition. The deck does not attribute that result specifically to AI or machine learning.", "Metric": "5% improved Utica base performance; issuer-reported operating-technology result, not counted as an AI result.", "Source_IDs": "OCT10-EOG-UTICA", "Source_URL": "https://investors.eogresources.com/image/EOG_0826.pdf", "Evidence": "Original issuer deck reviewed; AI attribution not established", "Origin": "10 October 2026 field-case review", "Assessment": "Related operating-technology context only. The deck does not define base performance or disclose the well count, measurement period, method, costs or counterfactual. The annual report’s description of machine-learning tools does not establish AI causality for this 5% result."}], "Related_IDs": [], "Reference_IDs": ["REF-046"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-08-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["oct6-gap-ai-089", "north_america-016"], "Research_search": "{\"packet_id\": \"oct6-gap-ai-089\", \"review_date\": \"2026-10-06\", \"title\": \"Proprietary machine-learning production optimizer\", \"workflow\": \"EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Operator account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage.\"], \"verification\": \"verified primary\", \"source_review\": \"Full original document reviewed on 6 October 2026\", \"source_locator\": \"PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time\", \"timeline_date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-089\", \"record_ids\": [\"AI-089\"], \"company\": \"EOG Resources\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"source_ids\": [\"OCT5-S033\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S033\"], \"workflow\": [\"OCT5-S033\"], \"technology_data\": [\"OCT5-S033\"], \"deployment_scope\": [\"OCT5-S033\"], \"practical_relevance\": [\"OCT5-S033\"], \"economics\": [\"OCT5-S033\"], \"integration_human_role\": [\"OCT5-S033\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"No AI-specific outcome or rollout count disclosed.\", \"period\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S033\"]}], \"timeline\": [{\"date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"OCT5-S033\"]}], \"sources\": [{\"id\": \"OCT5-S033\", \"title\": \"Proprietary generative AI for collaboration and operational data\", \"url\": \"https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf\", \"date\": \"2026 annual report covering 2025; exact publication day not confirmed\", \"type\": \"Operator annual report\", \"publisher\": \"EOG Resources\", \"locator\": \"PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time\", \"review_status\": \"Full original document reviewed on 6 October 2026\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Generative AI, production optimizers, remote control deployment.", "reason": "Existing citations are third-party transcript copies; issuer-source corroboration outstanding. Not disproved.", "source_url": "https://investors.eogresources.com/Events-and-Presentations"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-089", "company": "EOG Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence type as a management earnings-call transcript and retain exact speaker-attributed claims.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage. Related Utica result: Related operating-technology context only. The deck does not define base performance or disclose the well count, measurement period, method, costs or counterfactual. The annual report’s description of machine-learning tools does not establish AI causality for this 5% result.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-10"}, {"ID": "AI-090", "Company": "ConocoPhillips", "Company_primary": "ConocoPhillips", "Segment": "Independent E&P", "HQ": "US", "Theme": "Enterprise AI", "Initiative": "Enterprise AI strategy to embed AI in core processes", "Extracted_finding": "Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.", "Context": "Manual effort and inconsistent decision cycles across geographically diverse operations.", "Implementation": "CEO said the company is embedding AI in core business processes to reduce manual work, speed decisions and improve consistency.", "Assessment": "Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Enterprise framework described; named workflows not disclosed", "Interpretation_limit": "Disclosure is high-level; no practical named workflow or results identified.", "Source_IDs": "R049", "Source_date": "2026", "Source_type": "company announcement", "Source_locator": "CEO answer on AI", "Source_URL": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-049", "Initiative_ID": "AI-090", "Title": "CEO: embedding AI into core processes at scale", "Date": "2026-05", "Finding": "Answering a shareholder question at the 2026 annual meeting, CEO Ryan Lance called AI 'an increasingly important enabler'. He said it builds on long-standing analytics and that the company is focused on embedding AI into core business processes to cut manual effort, speed decision cycles and improve consistency across geographically diverse operations. He said all applications run under an enterprise framework for governance, security and responsible use.", "Metric": "No quantified metric supplied", "Source_IDs": "R049", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Enterprise framing (governance first, scale proven use cases, redeploy people to higher-value work) is the template large E&Ps are signaling to shareholders.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-049"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-004"], "Research_search": " {\"key\": \"conocophillips-enterprise-ai\", \"existing_ids\": [\"AI-090\"], \"company\": \"ConocoPhillips\", \"title\": \"Enterprise AI strategy to embed AI in core processes\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Strategy only\", \"stage\": \"Enterprise framework described; named workflows not disclosed\", \"practical_relevance\": \"Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.\", \"business_problem\": \"Manual effort and inconsistent decision cycles across geographically diverse operations.\", \"workflow\": \"CEO said the company is embedding AI in core business processes to reduce manual work, speed decisions and improve consistency.\", \"technology_data\": \"No named models, data, platform or partner disclosed.\", \"integration_human_role\": \"Enterprise governance, security and responsible-use framework; human roles and specific controls not detailed.\", \"deployment_scope\": \"Companywide ambition; no concrete deployment footprint reported.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No spend, savings, productivity or operating metrics disclosed.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"CEO described enterprise AI framework at annual meeting.\", \"source_ids\": [\"R049\"]}], \"limitations\": [\"Disclosure is high-level; no practical named workflow or results identified.\"], \"assessment\": \"Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.\", \"sources\": [{\"id\": \"R049\", \"title\": \"ConocoPhillips Annual Meeting 2026 Transcript\", \"url\": \"https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf\", \"publisher\": \"ConocoPhillips\", \"date\": \"2026\", \"type\": \"company announcement\", \"locator\": \"CEO answer on AI\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"CEO described focus on embedding AI into core processes to reduce manual work, accelerate decisions and improve consistency under enterprise governance/security/responsible-use framework; no named application or outcome disclosed.\"], \"accessed\": \"2026-09-27\", \"publication_date_status\": \"year only / exact publication date not established in reviewed page\", \"original_id\": \"na-006\"}], \"field_sources\": {\"business_problem\": [\"R049\"], \"workflow\": [\"R049\"], \"technology_data\": [\"R049\"], \"integration_human_role\": [\"R049\"], \"deployment_scope\": [\"R049\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"AI-090 is correctly characterized as an enterprise-level statement; it does not establish deployed applications or benefits.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-004\", \"source_ids\": [\"R049\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-090\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-090", "company": "ConocoPhillips", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf"], "findings": [{"severity": "medium", "field": "stage / workflow", "issue": "Enterprise AI intent is a strategy disclosure unless a specific workflow and deployment are documented.", "correction": "Do not imply operational deployment or measurable results beyond the disclosed strategy.", "source_url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-091", "Company": "Occidental Petroleum", "Company_primary": "Occidental", "Segment": "Independent E&P", "HQ": "US", "Theme": "Enterprise AI", "Initiative": "AI in workforce efficiency and process review", "Extracted_finding": "A stated workforce-efficiency objective, with no named workflow or measured result.", "Context": "On the Q1 2026 earnings call, Oxy leaders connected AI with workforce efficiency and process review. They gave no named AI workflow, deployment scale, or AI-specific productivity result.", "Implementation": "Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Expect AI to be tied explicitly to workforce redesign and process re-engineering, not only to technical use cases.", "Reported_metric": "No AI-specific quantified productivity, savings, or financial result disclosed", "Stage_at_disclosure": "Management described AI as taking a larger role across disciplines; specific deployments were not identified in this call.", "Interpretation_limit": "This source supports a general workforce-efficiency strategy statement, not an operating workflow or an AI-attributed cash-flow target.", "Source_IDs": "S25; R052", "Source_date": "2026-05-06", "Source_type": "issuer-hosted earnings-call transcript (with third-party transcript rendition reviewed)", "Source_locator": "COO answer on workforce efficiency", "Source_URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "Evidence_status": "Primary earnings-call transcript excerpt reviewed", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-05-06; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-052", "Initiative_ID": "AI-091", "Title": "AI identified as part of workforce-efficiency and process review", "Date": "2026-05-06", "Finding": "Management described AI taking a larger role across disciplines and workforce/process efficiency; this transcript does not disclose an AI-specific cash-flow target.", "Metric": "No AI-specific quantified result disclosed", "Source_IDs": "R052; S25", "Evidence": "Official transcript excerpt reviewed; third-party transcript is corroborating rendering", "Origin": "Q1 2026 earnings-call remarks", "Supplied_interpretation": "Management statement indicates strategic intent; it does not establish measured workforce savings or causation.", "Partners": "None disclosed", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-052"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-06", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-021"], "Research_search": " {\"key\": \"oxy-workforce-ai\", \"existing_ids\": [\"AI-091\"], \"company\": \"Occidental\", \"title\": \"AI in workforce efficiency and process review\", \"department\": \"Workforce & learning\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Center of Excellence established; individual workflow maturity not disclosed\", \"practical_relevance\": \"A stated workforce-efficiency objective, with no named workflow or measured result.\", \"business_problem\": \"Management linked AI with workforce efficiency and process review.\", \"workflow\": \"COO said AI was expanding across disciplines; no process was specified.\", \"technology_data\": \"Model, data and platform are undisclosed.\", \"integration_human_role\": \"Review/control process is undisclosed.\", \"deployment_scope\": \"No user or asset scope disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific savings disclosed.\", \"timeline\": [{\"date\": \"2026-05\", \"event\": \"On Q1 2026 call, COO described AI as taking a larger role across disciplines and connected it to workforce efficiency; no named process or result was disclosed.\", \"source_ids\": [\"S25\"]}], \"limitations\": [\"The claim does not establish a named deployment or measurable effect.\"], \"assessment\": \"A strategic direction, not a verified workflow or labor saving.\", \"sources\": [{\"id\": \"S25\", \"title\": \"Q1 2026 Occidental Earnings Call Transcript\", \"url\": \"https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf\", \"publisher\": \"Occidental\", \"date\": \"2026-05-06\", \"type\": \"earnings call\", \"locator\": \"COO answer on workforce efficiency\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"COO said AI is taking a larger role across company disciplines as part of workforce efficiency; no workflow, deployed scale or AI-only savings quantified.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-005\"}], \"field_sources\": {\"business_problem\": [\"S25\"], \"workflow\": [\"S25\"], \"technology_data\": [], \"integration_human_role\": [], \"deployment_scope\": [], \"economics\": []}, \"verification\": \"verified primary, strategy-level\", \"corrections_to_existing\": \"AI-091 exactly matches the workforce-efficiency point. No named process, adoption scale or AI-specific productivity outcome was disclosed.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-021\", \"source_ids\": [\"S25\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-091\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary, strategy-level", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-091", "company": "Occidental Petroleum", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf"], "findings": [{"severity": "high", "field": "Supplemental source review", "issue": "R052 confirms general management statements about AI's growing role in workforce efficiency, but it does not substantiate a $4B target or connect that target to AI. The Motley Fool URL is a third-party transcript; the official Oxy-hosted transcript corroborates the remarks but provides no AI-specific economics.", "correction": "Remove the $4B figure from the AI profile's Reported_metric and from the R052 disclosure finding; state that no AI-specific quantified outcome was disclosed. Limit the record to management's stated strategic direction and its generic workforce/process-efficiency context.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Issuer transcript excerpt reviewed. General workforce-efficiency strategy only; no named deployment or AI-specific result. The earlier pending classification is superseded by the existing transcript review.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary, strategy-level", "claim_status": "Primary earnings-call transcript excerpt reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-092", "Company": "Ovintiv", "Company_primary": "Ovintiv", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "Proprietary AI for drilling and centralized operations", "Extracted_finding": "Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.", "Context": "CEO Brendan McCracken said Ovintiv is 'just at the very front end' of applying AI, and is active in three areas: production operations (uptime, artificial-lift optimization), drilling times and costs, and completions. For the NuVista acquisition, Ovintiv planned to connect wells to its Grand Prairie operations control center and use in-house AI tools and algorithms to optimize lift setpoints, with remote shut-in and restart in minutes. It listed AI as both a near-term and a long-term synergy bucket.", "Implementation": "Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.", "Assessment": "Put AI and automation in the acquisition synergy case and connect acquired wells to the control center on day one.", "Reported_metric": "No quantified AI-specific outcome disclosed.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Interpretation_limit": "The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.", "Source_IDs": "R053; R054; R055; OCT5-S035", "Source_date": "2026-02-23", "Source_type": "SEC-filed operator annual report", "Source_locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence_status": "Original source text reviewed on 6 October 2026", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-05-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-053", "Initiative_ID": "AI-092", "Title": "In-house AI across production, drilling and completions", "Date": "2025-11", "Finding": "CEO Brendan McCracken said Ovintiv is 'just at the very front end' of applying AI, and is active in three areas: production operations (uptime, artificial-lift optimization), drilling times and costs, and completions. For the NuVista acquisition, Ovintiv planned to connect wells to its Grand Prairie operations control center and use in-house AI tools and algorithms to optimize lift setpoints, with remote shut-in and restart in minutes. It listed AI as both a near-term and a long-term synergy bucket.", "Metric": "No quantified metric supplied", "Source_IDs": "R053", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Put AI and automation in the acquisition synergy case and connect acquired wells to the control center on day one.", "Partners": "Operations control center; in-house AI algorithms", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "OCT5-AI-092-1", "Initiative_ID": "AI-092", "Title": "In-house AI for artificial lift, drilling and completions", "Date": "2026-02-23", "Finding": "Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.", "Metric": "No quantified AI-specific outcome disclosed.", "Source_IDs": "OCT5-S035", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Related_IDs": [], "Reference_IDs": ["REF-053"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["oct6-gap-ai-092", "north_america-012"], "Research_search": "{\"packet_id\": \"oct6-gap-ai-092\", \"review_date\": \"2026-10-06\", \"title\": \"Proprietary AI for drilling and centralized operations\", \"workflow\": \"Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Operator account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.\"], \"verification\": \"verified primary\", \"source_review\": \"Full original document reviewed on 6 October 2026\", \"source_locator\": \"2025 Form 10-K, Business and Properties, Montney section, printed page 13\", \"timeline_date\": \"2026-02-23\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-092\", \"record_ids\": [\"AI-092\"], \"company\": \"Ovintiv\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"source_ids\": [\"OCT5-S035\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S035\"], \"workflow\": [\"OCT5-S035\"], \"technology_data\": [\"OCT5-S035\"], \"deployment_scope\": [\"OCT5-S035\"], \"practical_relevance\": [\"OCT5-S035\"], \"economics\": [\"OCT5-S035\"], \"integration_human_role\": [\"OCT5-S035\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"No quantified AI-specific outcome disclosed.\", \"period\": \"2026-02-23\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S035\"]}], \"timeline\": [{\"date\": \"2026-02-23\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"OCT5-S035\"]}], \"sources\": [{\"id\": \"OCT5-S035\", \"title\": \"Proprietary AI for drilling and centralized operations\", \"url\": \"https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm\", \"date\": \"2026-02-23\", \"type\": \"SEC-filed operator annual report\", \"publisher\": \"Ovintiv\", \"locator\": \"2025 Form 10-K, Business and Properties, Montney section, printed page 13\", \"review_status\": \"Full original document reviewed on 6 October 2026\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "AI use cases and benefits across production, drilling and completions.", "reason": "Detailed workflow claims are currently supported by third-party transcript excerpts only; no primary issuer transcript was reviewed. Not disproved.", "source_url": "https://investors.ovintiv.com/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-092", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-093", "Company": "Ovintiv", "Company_primary": "Ovintiv", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "DRIVE Center applies AI to acquired rigs", "Extracted_finding": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "Context": "Weeks after closing an acquisition, Ovintiv had linked all producing wells to its operations control center and the drilling rigs to its DRIVE Center, an optimization tool that uses AI to improve drilling performance. Management said this lets synergies start immediately.", "Implementation": "Disclosed partners / technologies: DRIVE Center (in-house). Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "A centralized AI drilling-optimization center makes acquired rigs productive in weeks, which strengthens bid economics.", "Reported_metric": "About $1M per well synergy savings built into guidance", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Available transcript sources are third-party hosted; exact claims need issuer-hosted transcript/audio corroboration. Q1 2026 mentions combined operational performance; attribution to AI alone is not established.", "Source_IDs": "R054", "Source_date": "2026-02-24", "Source_type": "secondary", "Source_locator": "Management remarks", "Source_URL": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-02-24; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-054", "Initiative_ID": "AI-093", "Title": "DRIVE Center AI optimizes rigs from day one", "Date": "2026-02", "Finding": "Weeks after closing an acquisition, Ovintiv had linked all producing wells to its operations control center and the drilling rigs to its DRIVE Center, an optimization tool that uses AI to improve drilling performance. Management said this lets synergies start immediately.", "Metric": "About $1M per well synergy savings built into guidance", "Source_IDs": "R054", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A centralized AI drilling-optimization center makes acquired rigs productive in weeks, which strengthens bid economics.", "Partners": "DRIVE Center (in-house)", "Original_date": "2026-02", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-054"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-24", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-017"], "Research_search": " {\"key\": \"ovintiv-drive-center\", \"existing_ids\": [\"AI-093\"], \"company\": \"Ovintiv\", \"title\": \"DRIVE Center applies AI to acquired rigs\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Reported use cases and scale; primary transcript confirmation outstanding\", \"practical_relevance\": \"The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.\", \"business_problem\": \"Improve drilling/frac performance and operational cost through AI optimization.\", \"workflow\": \"Ovintiv’s Q4 2025 third-party transcript says recently acquired drilling rigs were linked to the DRIVE Center, an AI optimization tool, within weeks of closing.\", \"technology_data\": \"In-house algorithms paired with a proprietary/private data set; no architecture or named external partners disclosed.\", \"integration_human_role\": \"Operations and drilling optimization centers use company algorithms; transcript describes rapid acquired-asset integration. Human review/process not fully described.\", \"deployment_scope\": \"Claims span acquired wells/rigs, production operations centers and frac jobs; no well/rig totals verified.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-attributable cost or production uplift verified; management discussed combined cost/well results and synergy context without allocation.\", \"timeline\": [{\"date\": \"2026-02-24\", \"event\": \"Ovintiv’s Q4 2025 third-party transcript says recently acquired drilling rigs were linked to the DRIVE Center, an AI optimization tool, within weeks of closing.\", \"source_ids\": [\"R054\"]}], \"limitations\": [\"Available transcript sources are third-party hosted; exact claims need issuer-hosted transcript/audio corroboration.\", \"Q1 2026 mentions combined operational performance; attribution to AI alone is not established.\"], \"assessment\": \"The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.\", \"sources\": [{\"id\": \"R054\", \"title\": \"Ovintiv Q4 2025 Earnings Call Transcript\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/\", \"publisher\": \"The Motley Fool\", \"date\": \"2026-02-24\", \"type\": \"secondary\", \"locator\": \"Management remarks\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-019\"}], \"field_sources\": {\"business_problem\": [\"R054\"], \"workflow\": [\"R054\"], \"technology_data\": [\"R054\"], \"integration_human_role\": [\"R054\"], \"deployment_scope\": [\"R054\"], \"economics\": []}, \"verification\": \"secondary only\", \"corrections_to_existing\": \"AI-093 exactly matches this workflow. Evidence is a third-party transcript excerpt; primary issuer confirmation remains outstanding.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-017\", \"source_ids\": [\"R054\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-093\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [{"claim": "AI use cases and benefits across production, drilling and completions.", "reason": "Detailed workflow claims are currently supported by third-party transcript excerpts only; no primary issuer transcript was reviewed. Not disproved.", "source_url": "https://investors.ovintiv.com/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-093", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "realized", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-094", "Company": "Ovintiv", "Company_primary": "Ovintiv", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "Real-time AI optimization of fracturing", "Extracted_finding": "Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.", "Context": "McCracken called AI paired with Ovintiv's private dataset 'the big technology frontier'. The company develops in-house algorithms for production operations centers (uptime, artificial lift) and for frac design, tuning 'the 70 odd input design factors' of each frac job in real time.", "Implementation": "Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.", "Assessment": "A proprietary operating dataset plus in-house models is presented as a moat that competitors cannot copy.", "Reported_metric": "No quantified AI-specific outcome disclosed.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Interpretation_limit": "The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.", "Source_IDs": "R055; OCT5-S036; OCT5-S035", "Source_date": "2026-02-23", "Source_type": "SEC-filed operator annual report", "Source_locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence_status": "Original source text reviewed on 6 October 2026", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-05-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-055", "Initiative_ID": "AI-094", "Title": "AI tunes 70-odd frac design inputs in real time", "Date": "2026-05", "Finding": "McCracken called AI paired with Ovintiv's private dataset 'the big technology frontier'. The company develops in-house algorithms for production operations centers (uptime, artificial lift) and for frac design, tuning 'the 70 odd input design factors' of each frac job in real time.", "Metric": "About 70 frac design input factors optimized in real time", "Source_IDs": "R055", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A proprietary operating dataset plus in-house models is presented as a moat that competitors cannot copy.", "Partners": "In-house algorithms", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "OCT5-AI-094-1", "Initiative_ID": "AI-094", "Title": "AI tunes frac-design factors in real time", "Date": "2026-02-23", "Finding": "Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.", "Metric": "No quantified AI-specific outcome disclosed.", "Source_IDs": "OCT5-S036", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Related_IDs": [], "Reference_IDs": ["REF-055"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["oct6-gap-ai-094", "north_america-018"], "Research_search": "{\"packet_id\": \"oct6-gap-ai-094\", \"review_date\": \"2026-10-06\", \"title\": \"Real-time AI optimization of fracturing\", \"workflow\": \"Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Operator account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.\"], \"verification\": \"verified primary\", \"source_review\": \"Full original document reviewed on 6 October 2026\", \"source_locator\": \"2025 Form 10-K, Business and Properties, Montney section, printed page 13\", \"timeline_date\": \"2026-02-23\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-094\", \"record_ids\": [\"AI-094\"], \"company\": \"Ovintiv\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"source_ids\": [\"OCT5-S035\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S035\"], \"workflow\": [\"OCT5-S035\"], \"technology_data\": [\"OCT5-S035\"], \"deployment_scope\": [\"OCT5-S035\"], \"practical_relevance\": [\"OCT5-S035\"], \"economics\": [\"OCT5-S035\"], \"integration_human_role\": [\"OCT5-S035\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"No quantified AI-specific outcome disclosed.\", \"period\": \"2026-02-23\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S035\"]}], \"timeline\": [{\"date\": \"2026-02-23\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"OCT5-S035\"]}], \"sources\": [{\"id\": \"OCT5-S035\", \"title\": \"Proprietary AI for drilling and centralized operations\", \"url\": \"https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm\", \"date\": \"2026-02-23\", \"type\": \"SEC-filed operator annual report\", \"publisher\": \"Ovintiv\", \"locator\": \"2025 Form 10-K, Business and Properties, Montney section, printed page 13\", \"review_status\": \"Full original document reviewed on 6 October 2026\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "AI use cases and benefits across production, drilling and completions.", "reason": "Detailed workflow claims are currently supported by third-party transcript excerpts only; no primary issuer transcript was reviewed. Not disproved.", "source_url": "https://investors.ovintiv.com/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-094", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-095", "Company": "EQT", "Company_primary": "EQT", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "AI-assisted service rebids and procurement targeting", "Extracted_finding": "Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return.", "Context": "CEO Toby Rice discussed AI-assisted service-rebid targeting and a prospective procurement cost opportunity. His separate mention of back-office purchasing does not establish AI use there.", "Implementation": "Toby Rice described AI tools that could pinpoint service rebid efforts. He separately noted a procurement focus beyond the field; the source does not explicitly extend AI use to back-office purchasing. No automated award process was specified.", "Assessment": "The use case is specific, but the public statement provides no spend denominator, named system, implementation detail or achieved outcome.", "Reported_metric": "Prospective low-single-digit procurement cost reduction; unit and spend denominator unspecified", "Stage_at_disclosure": "Unclear. EQT CEO said during the 2026-02-18 earnings call that the company had AI tools that 'would allow' it to pinpoint service-rebid efforts, while describing aggressive future rebidding and a possible low-single-digit procurement opportunity. The quote establishes tool availability, but not that this particular targeting workflow is live or piloted.", "Interpretation_limit": "Low-single-digit procurement cost reduction is an opportunity estimate, not a realized saving or an isolated AI effect. Tool, vendor, scope and baseline remain undisclosed.", "Source_IDs": "EQTD01; R056", "Source_date": "2026-02-18", "Source_type": "Issuer-hosted corrected earnings-call transcript", "Source_locator": "Printed pp. 20–21; Toby Rice, service rebids and procurement", "Source_URL": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "Evidence_status": "Issuer-hosted transcript reviewed on 29 September 2026", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-02-18; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-056", "Initiative_ID": "AI-095", "Title": "AI tools to target service rebids and procurement", "Date": "2026-02", "Finding": "EQT said it will aggressively rebid oilfield services and that AI tools will help pinpoint those efforts. It expects low-single-digit cost reductions on the procurement side, and said it is focused on 'the procurement and the backside' as well as the field.", "Metric": "Low-single-digit % procurement savings targeted", "Source_IDs": "R056", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Procurement analytics (spend categorization, rebid targeting) is a quick back-office AI win with measurable savings.", "Partners": "", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "EQTD-095", "Initiative_ID": "AI-095", "Title": "Issuer transcript follow-up", "Date": "2026-02-18", "Finding": "Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return.", "Metric": "Prospective low-single-digit procurement cost reduction; unit and spend denominator unspecified", "Source_IDs": "EQTD01", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "29 September 2026 follow-up"}], "Related_IDs": [], "Reference_IDs": ["REF-056"], "Provenance": ["User-supplied reference", "Issuer transcript follow-up, 29 September 2026"], "Latest_disclosure": "2026-02-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["north_america-009"], "Research_search": "{\"key\": \"eqt-procurement-ai\", \"existing_ids\": [\"AI-095\"], \"company\": \"EQT\", \"title\": \"AI-assisted service rebids and procurement targeting\", \"department\": \"Supply chain & logistics\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Tools discussed for prospective rebid targeting; deployment extent undisclosed\", \"practical_relevance\": \"Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return.\", \"business_problem\": \"Identify service contract rebids and procurement savings opportunities.\", \"workflow\": \"Management identified AI tools that could help prioritize service rebids. Broader purchasing activity was discussed separately.\", \"technology_data\": \"No tools, supplier-spend data or process design disclosed.\", \"integration_human_role\": \"No review/approval workflow disclosed.\", \"deployment_scope\": \"No deployment scope or implementation date specified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Expected procurement cost reduction\", \"value\": \"low-single-digit\", \"period\": \"management expectation at Q4 2025 call\", \"attribution\": \"Management expectation in issuer-hosted transcript; not AI-only\", \"source_ids\": [\"EQTD01\"]}], \"economics\": \"Prospective procurement opportunity; the call does not specify a unit or spend denominator, achieved savings or an AI-only effect.\", \"timeline\": [{\"date\": \"2026-02-18\", \"event\": \"Management discussed AI-supported rebid targeting and procurement focus.\", \"source_ids\": [\"R056\"]}], \"limitations\": [\"Low-single-digit procurement cost reduction is an opportunity estimate, not a realized saving or an isolated AI effect. Tool, vendor, scope and baseline remain undisclosed.\"], \"assessment\": \"The use case is specific, but the public statement provides no spend denominator, named system, implementation detail or achieved outcome.\", \"sources\": [{\"id\": \"EQTD01\", \"title\": \"EQT Q4 2025 earnings call — AI-assisted procurement targeting\", \"url\": \"https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf\", \"publisher\": \"EQT / FactSet CallStreet\", \"date\": \"2026-02-18\", \"type\": \"Issuer-hosted corrected earnings-call transcript\", \"locator\": \"Printed pp. 20–21; Toby Rice, service rebids and procurement\", \"review_status\": \"full text reviewed\", \"supports\": [\"Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return.\"], \"accessed\": \"2026-09-29\"}, {\"id\": \"R056\", \"title\": \"EQT Q4 2025 Earnings Call\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/\", \"publisher\": \"The Motley Fool\", \"date\": \"2026-02-18\", \"type\": \"secondary\", \"locator\": \"management remarks on service rebids/procurement\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Management said AI tools would help target service rebids and procurement, expecting low-single-digit procurement cost reductions; not isolated to AI and no method, baseline or achieved value disclosed.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-012\"}], \"field_sources\": {\"business_problem\": [\"EQTD01\"], \"workflow\": [\"EQTD01\"], \"technology_data\": [\"EQTD01\"], \"integration_human_role\": [\"EQTD01\"], \"deployment_scope\": [\"EQTD01\"], \"economics\": [\"EQTD01\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Issuer-hosted corrected transcript now reviewed. The prior secondary-only access limitation is resolved; implementation and economic attribution remain unknown.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-009\", \"source_ids\": [\"EQTD01\", \"R056\"], \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-095\"]}", "Research_status": "Expanded source review · updated 29 September 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-095", "company": "EQT", "outcome": "supported", "reviewer_outcome": "supported", "checks": ["issuer transcript", "dates", "stage", "attribution"], "source_urls_reviewed": ["https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf"], "findings": [], "residual_gaps": ["No disclosed implementation scope, realized savings or AI-only attribution."], "reviewed_at": "2026-09-29"}, "Audit_outcome": "supported", "Normalized_stage": "Unclear", "Stage_review": {"id": "AI-095", "stage": "Unclear", "stage_basis": "EQT CEO said during the 2026-02-18 earnings call that the company had AI tools that 'would allow' it to pinpoint service-rebid efforts, while describing aggressive future rebidding and a possible low-single-digit procurement opportunity. The quote establishes tool availability, but not that this particular targeting workflow is live or piloted.", "source_url": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "source_date": "2026-02-18", "evidence_origin": "Operator", "review_correction": "Classify as Unclear rather than Pilot or Operating pending direct disclosure that AI was applied to bids. EQT has tools; target workflow/use extent remain unconfirmed.", "limits": "No named tool, user group, data, rebid count, realized savings, or deployment status. The low-single-digit cost opportunity is prospective and not AI-attributed.", "source_id": "EQTD01"}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Issuer-hosted transcript reviewed on 29 September 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-096", "Company": "Matador Resources", "Company_primary": "Matador Resources", "Segment": "Independent E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "Cross-functional AI program for operations analytics, frac monitoring and geosteering", "Extracted_finding": "The Q1 2026 call names a cross-functional AI committee and practical operational applications: analytics alerts intended to reduce downtime, real-time frac monitoring and AI-assisted lateral targeting. The call supplies operating context and data scale but does not isolate AI-attributed performance.", "Context": "Matador describes using operational data and AI in drilling/completions and lateral targeting, with a cross-functional committee to coordinate adoption.", "Implementation": "Management described AI analytics that surface actionable alerts to reduce downtime; real-time frac monitoring of pressures, water and sand volumes and logistics as simultaneous/trimul-frac operations expand; and AI-assisted geosteering to remain in the target zone. The transcript cites about 40 million data points per day and more than 36 drilling records in Q1 2026, but neither figure is an AI outcome.", "Assessment": "This is more than a vendor-only strategy statement: management names operating workflows and coordination. Deployment counts, model details and measured AI-specific improvement remain undisclosed.", "Reported_metric": "About 40 million operational data points per day; 36+ drilling records in Q1 2026 (company-reported operating context, not an AI-attributed result)", "Stage_at_disclosure": "Q1 2026 call describes workflows in use/underway and a cross-functional AI committee; deployment scale and individual workflow maturity are not quantified.", "Interpretation_limit": "Transcript is a third-party rendering of management remarks. Data/record counts do not establish AI contribution or performance; no controlled result is reported.", "Source_IDs": "RS039; R057; R058", "Source_date": "2026-05-07", "Source_type": "earnings-call transcript (third-party transcription)", "Source_locator": "Q1 2026 call: AI committee, real-time operations data, frac monitoring and geosteering", "Source_URL": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "Evidence_status": "Management-reported transcript reviewed; workflows named, outcomes not independently validated", "Review_tier": "Reviewed source; third-party transcript", "Time_scope": "Q1 2026 management remarks; exact individual workflow start dates not disclosed.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": ["Drilling & completions", "Drilling & completions", "IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-058", "Initiative_ID": "AI-096", "Title": "AI committee and operating workflows in frac monitoring and geosteering", "Date": "2026-05-07", "Finding": "Management described a cross-functional committee for controlled AI adoption, analytics alerts intended to reduce downtime, real-time frac monitoring of pressure/water/sand/logistics, and AI-assisted lateral targeting. The call cited about 40 million data points per day and 36+ drilling records in Q1; those are context/operating-scale figures, not AI-attributed results.", "Metric": "About 40 million data points/day; more than 36 drilling records in Q1 2026 (company-reported, not AI-specific performance)", "Source_IDs": "R058", "Evidence": "Third-party transcript of company management remarks; reviewed", "Origin": "Q1 2026 earnings-call transcript", "Supplied_interpretation": "Operational applications are named, but contribution and realized performance are not isolated.", "Partners": "MaxCom (real-time operations center)", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-057", "REF-058"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-07", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-008"], "Research_search": " {\"key\": \"matador-ai-vendor-fast-follower\", \"existing_ids\": [\"AI-096\"], \"company\": \"Matador Resources\", \"title\": \"AI analytics in production, completions and drilling workflows\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"Drilling & completions\", \"Drilling & completions\", \"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"Management described active integration of AI-driven analytics across operations in Q1 2026; operating footprint and duration by workflow not quantified.\", \"practical_relevance\": \"Management describes real-time monitoring and decision support spanning control-room production data, hydraulic fracturing operations, and lateral placement. The transcript provides practical workflow detail while leaving model architecture, vendor, validation and AI-attributable operating results undisclosed.\", \"business_problem\": \"Matador seeks to make operational signals actionable quickly to reduce downtime and identify inefficiencies, while improving completion execution and well placement.\", \"workflow\": \"Management says the control room and field staff monitor more than 40 million data points per day in real time to make production signals actionable. For completions, the company monitors hydraulic-fracturing water and sand pressures and volumes plus logistics. For drilling, management says AI assists targeting the preferred portion of the lateral target. The CEO described committee oversight and a measured adoption pace.\", \"technology_data\": \"The call names AI-driven analytics and a daily stream of more than 40 million data points, but does not specify model types, vendors, source systems, data governance or validation methods.\", \"integration_human_role\": \"Matador describes a cross-functional committee overseeing AI use. Control-room personnel monitor data with field staff; AI is described as assisting targeting and analytics, not controlling drilling or completions autonomously.\", \"deployment_scope\": \"Management described use across production, completions and drilling, including real-time control-room/field monitoring and lateral targeting. Counts of wells, rigs, crews or assets covered were not disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-attributed savings, production uplift, cost reduction or return disclosed. The call discusses cycle-time and cost levers but does not attribute those results exclusively to AI.\", \"timeline\": [{\"date\": \"2026-05-07\", \"event\": \"Q1 2026 earnings call described AI-driven analytics in production monitoring, completion monitoring and drilling lateral targeting, with a committee-led measured approach.\", \"source_ids\": [\"R058\"]}], \"limitations\": [\"Earnings call transcript is third-party-hosted, although the statements are attributed to Matador management.\", \"More than 40 million daily data points is an operating scale statement, not an AI outcome.\", \"No named vendor, model, asset counts, independent validation or AI-specific economics disclosed.\", \"Natural-gas transport/sales to AI data centers is infrastructure and remains distinct from internal AI adoption.\"], \"assessment\": \"Analyst inference: this is more than strategy-only because management describes ongoing, workflow-specific AI use with human operators in the loop. Its practical value is plausible in surfacing anomalies, monitoring completions and improving target placement; deployment maturity and measurable causal impact remain unproven.\", \"sources\": [{\"id\": \"R058\", \"title\": \"Matador Q1 2026 Earnings Call Transcript\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/\", \"publisher\": \"The Motley Fool (transcript of Matador management statements)\", \"date\": \"2026-05-07\", \"type\": \"earnings call transcript (secondary transcription)\", \"locator\": \"Q&A, management response on AI adoption and operations\", \"review_status\": \"full text reviewed\", \"supports\": [\"Matador executives say a committee is overseeing measured AI adoption and that AI-driven analytics are being integrated across operations.\", \"The CEO and operations executive describe control-room and field review of >40 million data points per day, real-time monitoring of frac pressures/volumes/logistics, and AI-assisted lateral targeting.\", \"Management does not provide an AI-specific performance or economic result.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"R058\"}], \"field_sources\": {\"business_problem\": [\"R058\"], \"workflow\": [\"R058\"], \"technology_data\": [\"R058\"], \"integration_human_role\": [\"R058\"], \"deployment_scope\": [\"R058\"], \"economics\": [\"R058\"]}, \"verification\": \"Management statements reviewed in secondary-hosted earnings transcript\", \"corrections_to_existing\": \"The prior strategy-only packet said no workflow was named. Q1 2026 management remarks identify production control-room monitoring, completions monitoring, and lateral targeting workflows; no AI-attributable results are established.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-008\", \"source_ids\": [\"R058\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-096\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "Management statements reviewed in secondary-hosted earnings transcript", "Audit_flags": [{"claim": "Named vendor-led technical AI deployment.", "reason": "Official release only identifies AI integration as a focus area; CEO transcript is third-party-hosted and identifies no vendors or use case.", "source_url": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Q1 transcript is a third-party rendering; operational outcomes and AI-attributed metrics remain management-reported and unvalidated.", "source_url": ""}], "Audit_review": {"id": "AI-096", "company": "Matador Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence as a third-party rendering of management’s earnings-call remarks. Preserve speaker attribution and distinguish pilot/workflow disclosure from independently validated results.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/"}, {"severity": "medium", "field": "workflow / stage", "issue": "The Q1 2026 call names operational uses: alerts intended to reduce downtime, real-time frac monitoring of pressures/water/sand/logistics, and AI-assisted lateral targeting. The source describes use but does not independently validate performance or establish a measured AI uplift.", "correction": "Remove the stale strategy-only characterization; describe the call-specific workflows while keeping scale, validation and attribution limits explicit.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}, {"severity": "low", "field": "Supplemental source review", "issue": "Q1 2026 transcript gives more operational detail than the inherited “vendor-led cautious integration strategy” title: daily real-time data monitored by control-room/field teams, frac water/sand pressures/volumes and logistics, and AI-assisted preferred lateral targeting. Reported 40M data points/day is company-stated scale, not an outcome metric.", "correction": "", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "Q1 transcript is a third-party rendering; operational outcomes and AI-attributed metrics remain management-reported and unvalidated."], "partition": 3, "reviewer_outcome": "partially supported; operational workflows named, AI-attributed results not established"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Management statements reviewed in secondary-hosted earnings transcript", "claim_status": "Management-reported transcript reviewed; workflows named, outcomes not independently validated", "independent_validation": "Not established by this collection"}}, {"ID": "AI-097", "Company": "Chord Energy", "Company_primary": "Chord Energy", "Segment": "Upstream E&P", "HQ": "US", "Theme": "Operational AI", "Initiative": "AI-driven rod-lift optimization", "Extracted_finding": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "Context": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "Implementation": "Model inputs, vendor, architecture and control/approval mode are undisclosed.", "Assessment": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "Reported_metric": "Operator reports early approximately 25% improvement in rod-pump run times. The 99% denominator is rod-lift wells, not the entire well fleet.", "Stage_at_disclosure": "Scaled. Issuer FY2025 results dated 2026-02-25 say AI/ML reached about 99% of Chord’s rod-lift wells during 2025; early rod-pump run-time improvement was about 25%.", "Interpretation_limit": "Company-reported early result without a disclosed comparison method. Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.", "Source_IDs": "R059; R060; SW7-005; SW7-006; SW7-007; SW7-008; OCT14C-CHORD-LIFT", "Source_date": "2026-02-25", "Source_type": "issuer earnings release", "Source_locator": "Operations Update, lines 210–216; specifically line 216 (AI deployment, run time metric, workover rig days).", "Source_URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "Evidence_status": "Full source opened and read.", "Review_tier": "Reviewed source", "Time_scope": "2025 deployment reported 2026-02-25; May deck supplies mixed-program economics; August release confirms continuing use.", "Coverage_bucket": "Seventh targeted gap review", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Internal adoption", "Disclosures": [{"ID": "REF-059", "Initiative_ID": "AI-097", "Title": "AI rod-pump optimization arrests base decline", "Date": "2026-05", "Finding": "Chord adjusted its artificial-intelligence rod-pump controls to maximize productivity from older Bakken wells, alongside lowering pumps deeper and adding workover rigs. Management showed a positive impact on base production and a halt in decline for the targeted well group.", "Metric": "No quantified metric supplied", "Source_IDs": "R059", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Retuning AI lift controllers for legacy wells is a cheap, short-cycle barrel source when prices call for it.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-060", "Initiative_ID": "AI-097", "Title": "AI rod-pump program deployed field-wide", "Date": "2026-08", "Finding": "Chord said AI optimization of its rod-pump program is now 'pretty broadly implemented' across its Williston field, where nearly every well ends up on rod pump. The system sets pump loading to reduce wear and improve production, and Chord is looking for other places to apply similar computer-driven optimization.", "Metric": "No quantified metric supplied", "Source_IDs": "R060", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardized lift type (rod pump) across a field makes one AI optimizer broadly deployable.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "SW7-D-AI-097", "Initiative_ID": "AI-097", "Title": "AI-driven rod-lift optimization", "Date": "2026-02-25", "Finding": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "Metric": "Operator reports early approximately 25% improvement in rod-pump run times. The 99% denominator is rod-lift wells, not the entire well fleet.", "Source_IDs": "SW7-005; SW7-006; SW7-007; SW7-008", "Evidence": "Full source opened and read.", "Origin": "Seventh research swarm"}, {"ID": "OCT14C-AI-097", "Initiative_ID": "AI-097", "Title": "Rod-lift AI sits inside a broader artificial-lift improvement program", "Date": "2026-02-25", "Finding": "Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register. The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.", "Metric": "The broader program reports over 50% improvement in ESP replacement cycle times, over 25% improvement in failure rates since early 2025, and about 1,200 fewer workover rig days in 2025. These are not separately isolated ML effects.", "Source_IDs": "OCT14C-CHORD-LIFT", "Source_URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted."}], "Related_IDs": [], "Reference_IDs": ["REF-059", "REF-060"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-08-05", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["swarm7-ai-097", "north_america-013", "oct14c-ai-097"], "Research_search": "{\"packet_id\": \"oct14c-ai-097\", \"key\": \"oct14c-ai-097\", \"record_ids\": [\"AI-097\"], \"company\": \"Chord Energy\", \"title\": \"Rod-lift AI sits inside a broader artificial-lift improvement program\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Existing scaled AI workflow; broader operating results kept separate\", \"business_problem\": \"Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register.\", \"workflow\": \"The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.\", \"technology_data\": \"The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.\", \"deployment_scope\": \"2025 results, disclosed 25 February 2026. ESP comparisons start at the beginning of 2025; an exact matched AI evaluation period is not supplied.\", \"integration_human_role\": \"Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.\", \"economics\": \"The broader program reports over 50% improvement in ESP replacement cycle times, over 25% improvement in failure rates since early 2025, and about 1,200 fewer workover rig days in 2025. These are not separately isolated ML effects.\", \"practical_relevance\": \"What part of fewer failures and workovers comes from ML rather than equipment changes, maintenance and field practices?\", \"assessment\": \"What part of fewer failures and workovers comes from ML rather than equipment changes, maintenance and field practices?\", \"limitations\": [\"Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.\"], \"source_ids\": [\"OCT14C-CHORD-LIFT\"], \"sources\": [{\"id\": \"OCT14C-CHORD-LIFT\", \"title\": \"Chord fourth-quarter and full-year 2025 results\", \"url\": \"https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend\", \"date\": \"2026-02-25\", \"type\": \"Operator earnings release\", \"publisher\": \"Chord Energy\", \"locator\": \"Operations Update / Production and LOE; two consecutive artificial-lift paragraphs\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-02-25\", \"event\": \"Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register.\", \"source_ids\": [\"OCT14C-CHORD-LIFT\"]}], \"field_sources\": {\"business_problem\": [\"OCT14C-CHORD-LIFT\"], \"workflow\": [\"OCT14C-CHORD-LIFT\"], \"technology_data\": [\"OCT14C-CHORD-LIFT\"], \"deployment_scope\": [\"OCT14C-CHORD-LIFT\"], \"economics\": [\"OCT14C-CHORD-LIFT\"]}}", "Research_status": "Clarified 14 October 2026 · third pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "AI rod-pump optimization halted base decline.", "reason": "Third-party transcript combines AI controls with pump depth and workovers; AI-only causal effect is not established.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-097", "company": "Chord Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Asset_location": "Not disclosed", "Followup_review": {"date": "2026-10-01", "decision": "Enriched", "note": "Issuer evidence supersedes prior secondary-only status. Original pending disclosure retained as history."}, "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-097", "stage": "Scaled", "stage_basis": "Issuer FY2025 results dated 2026-02-25 say AI/ML reached about 99% of Chord’s rod-lift wells during 2025; early rod-pump run-time improvement was about 25%.", "source_url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "source_date": "2026-02-25", "evidence_origin": "Operator", "limits": "99% denominator is rod-lift wells, not total company wells. The ~25% is an early improvement, not fleet-wide average. Mixed corporate/production cost-savings numbers are not AI-attributed.", "source_id": "SW7-005"}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "~99% of rod-lift wells", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Company-reported early result without a disclosed comparison method. Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Existing scaled AI workflow; broader operating results kept separate", "independent_validation": "Not established"}, "Current_source_check": "2026-10-14", "Research_origin": ["research/gap-swarm-oct14c/README.md"]}, {"ID": "AI-098", "Company": "Canadian Natural Resources", "Company_primary": "Canadian Natural Resources", "Segment": "Independent E&P", "HQ": "Canada", "Theme": "Enterprise AI", "Initiative": "Cost-justified AI with production-critical use gated", "Extracted_finding": "This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.", "Context": "Use AI for employee knowledge, data insights and efficiency while controlling cyber, IP and operational risk.", "Implementation": "AIF describes incremental, cost-justified adoption and bars reliance on AI in time-sensitive or production-critical processes until controls meet standards.", "Assessment": "This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Policy in place; individual deployments not described", "Interpretation_limit": "Policy statement does not verify any named AI product or production deployment.", "Source_IDs": "RS037; RS038; R061", "Source_date": "2026-03-25", "Source_type": "annual report", "Source_locator": "Artificial Intelligence section, p.21", "Source_URL": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-03-25; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-061", "Initiative_ID": "AI-098", "Title": "Formal AI policy: no AI in production-critical processes", "Date": "2026-03", "Finding": "CNRL's 2025 Annual Information Form says it uses AI 'responsibly' to widen employee access to knowledge, deliver data-driven insights and improve efficiency through 'incremental, cost-justified deployments'. It will not rely on AI for time-sensitive or production-critical processes until controls meet necessary standards. Employees must stay involved in and accountable for AI outputs, and AI risk is reported to board committees.", "Metric": "No quantified metric supplied", "Source_IDs": "R061", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A conservative, human-in-the-loop policy disclosed in securities filings shows how cost-focused operators govern GenAI.", "Partners": "", "Original_date": "2026-03", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-061"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-03-25", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-007"], "Research_search": " {\"key\": \"cnrl-governed-enterprise-ai\", \"existing_ids\": [\"AI-098\"], \"company\": \"Canadian Natural Resources\", \"title\": \"Cost-justified AI with production-critical use gated\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Strategy only\", \"stage\": \"Policy in place; individual deployments not described\", \"practical_relevance\": \"This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.\", \"business_problem\": \"Use AI for employee knowledge, data insights and efficiency while controlling cyber, IP and operational risk.\", \"workflow\": \"AIF describes incremental, cost-justified adoption and bars reliance on AI in time-sensitive or production-critical processes until controls meet standards.\", \"technology_data\": \"No specific AI tools or datasets named.\", \"integration_human_role\": \"Employees must participate in creation/review/decisions and remain accountable; AI risk enters ERM and board committee reporting.\", \"deployment_scope\": \"Enterprise-wide policy; no application count or operational footprint disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific spend or return disclosed.\", \"timeline\": [{\"date\": \"2025-03-25\", \"event\": \"AIF described responsible AI program and risk policy.\", \"source_ids\": [\"RS037\"]}, {\"date\": \"2026\", \"event\": \"Company governance page summarized knowledge, data insight and cost-efficiency uses.\", \"source_ids\": [\"RS038\"]}], \"limitations\": [\"Policy statement does not verify any named AI product or production deployment.\"], \"assessment\": \"This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.\", \"sources\": [{\"id\": \"RS037\", \"title\": \"Canadian Natural 2025 Annual Information Form\", \"url\": \"https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf\", \"publisher\": \"Canadian Natural Resources\", \"date\": \"2026-03-25\", \"type\": \"annual report\", \"locator\": \"Artificial Intelligence section, p.21\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"AI is used to expand employee knowledge access, deliver data-driven insight and improve efficiency via incremental, cost-justified deployment; company will not rely on it for time-sensitive/production-critical processes until controls meet standards; employees remain involved/accountable; board risk oversight described.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"na-009\"}, {\"id\": \"RS038\", \"title\": \"Governance and Risk Management\", \"url\": \"https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/\", \"publisher\": \"Canadian Natural Resources\", \"date\": null, \"type\": \"company announcement\", \"locator\": \"Use of AI in operations\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Corroborates program-level AI scope; refers readers to AIF for policy specifics.\"], \"accessed\": \"2026-09-27\", \"publication_date_status\": \"undated page; accessed 2026-09-27\", \"original_id\": \"na-013\"}], \"field_sources\": {\"business_problem\": [\"RS037\", \"RS038\"], \"workflow\": [\"RS037\", \"RS038\"], \"technology_data\": [\"RS037\", \"RS038\"], \"integration_human_role\": [\"RS037\", \"RS038\"], \"deployment_scope\": [\"RS037\", \"RS038\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"AI-098 is supported by primary AIF; wording “no AI in production-critical processes” is conditional pending control validation.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-007\", \"source_ids\": [\"RS037\", \"RS038\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-098\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-098", "company": "Canadian Natural Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-099", "Company": "Suncor Energy", "Company_primary": "Suncor Energy", "Segment": "Independent E&P", "HQ": "Canada", "Theme": "Automation context", "Initiative": "Autonomous haulage fleet expansion at Base Plant (AI not established)", "Extracted_finding": "Suncor describes a large operating autonomous haul fleet and control-room oversight. The source supports autonomy/automation but does not establish that AI is used.", "Context": "Suncor moves oil-sands material across Base Plant mine routes using an Autonomous Haulage System.", "Implementation": "Suncor says the driverless haul trucks use GPS, wireless/LTE communications and perceptive technologies. AHS software gives control-room operators real-time visibility so they can manage traffic flow and adjust to conditions. The page does not identify an AI model or say AI optimizes driving conditions.", "Assessment": "The page supports a substantial autonomy deployment and human oversight, not an AI adoption claim. Nearly 120 is the reported snapshot; 140 is a future target whose year cannot be established because the page is undated.", "Reported_metric": "Autonomous haul trucks: almost 120 across two integrated mines at the undated page snapshot; 140 equipped trucks planned by an unspecified year-end (Company-reported)", "Stage_at_disclosure": "Autonomous haulage operating; page snapshot is undated; 140 is prospective", "Interpretation_limit": "The page does not mention AI. It gives no publication date, so the snapshot and prospective target cannot be assigned to 2025 or 2026. Do not treat the target as deployed. Keep dynamic dispatch as a separate record.", "Source_IDs": "RS036; R062", "Source_date": "Undated; accessed 2026-09-27", "Source_type": "undated company feature", "Source_locator": "Base Plant fleet, command centre and end-of-page target statement", "Source_URL": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "Evidence_status": "Verified primary for autonomous automation; AI not established", "Review_tier": "Reviewed source", "Time_scope": "Undated page accessed 2026-09-27; an in-window publication date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Automation or digital workflow; explicit AI basis not established", "Disclosures": [{"ID": "REF-062", "Initiative_ID": "AI-099", "Title": "Autonomous haulage fleet snapshot and prospective target", "Date": null, "Finding": "Suncor describes an autonomous haul fleet operating at Base Plant, with almost 120 trucks across two integrated mines in the current page snapshot and a plan for 140 equipped trucks by an unspecified year-end. It identifies GPS, wireless/LTE and perceptive technologies and central control-room oversight. The page does not establish an AI component or support an AI-specific optimization claim.", "Metric": "Almost 120 operating autonomous haul trucks in the undated page snapshot; 140 planned by an unspecified year-end", "Source_IDs": "RS036", "Evidence": "Primary source reviewed; automation supported, AI not established", "Origin": "User-supplied reference, corrected against source", "Supplied_interpretation": "Autonomous fleet scale and workforce implications require separate evidence; this page does not quantify job impacts.", "Partners": "Komatsu trucks; source does not establish a named AHS software vendor", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-062"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["north_america-006"], "Research_search": " {\"key\": \"suncor-ai-mine-operations\", \"existing_ids\": [\"AI-099\"], \"company\": \"Suncor Energy\", \"title\": \"Autonomous haulage fleet expansion at Base Plant (AI not established)\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Automation / AI not established\", \"stage\": \"Autonomous haulage operating; undated page snapshot; 140 trucks is a prospective target\", \"practical_relevance\": \"Suncor reports an operating autonomous haulage fleet with control-room oversight. The source does not establish AI use; almost 120 is a snapshot and 140 is a prospective target.\", \"business_problem\": \"Move oil-sands material safely and consistently across Base Plant mine routes.\", \"workflow\": \"Driverless haul trucks move material using GPS, wireless/LTE communications and perceptive technologies. AHS software provides central control-room operators with real-time location and movement visibility so they can manage traffic flow and adapt to driving conditions.\", \"technology_data\": \"Suncor names GPS, wireless/LTE, perceptive technologies, AHS software and collision/obstacle detection. It does not identify an AI model, optimization algorithm or input-data architecture.\", \"integration_human_role\": \"Central control-room operators monitor real-time truck movement and manage flow; field deployment teams, operators and dispatchers contributed. Human oversight is explicit.\", \"deployment_scope\": \"Suncor says all ore at Base Plant is moved by AHS trucks and reports almost 120 trucks across two integrated mines in an undated page snapshot. A target of 140 equipped trucks is prospective; its year is not stated on the page.\", \"partners\": [\"Komatsu (haul trucks; software vendor not specified in the reviewed page)\"], \"disclosed_results\": [{\"metric\": \"Autonomous haul trucks\", \"value\": \"Almost 120 across two integrated mines in the undated page snapshot; 140 planned\", \"period\": \"Undated page snapshot; target year not stated\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS036\"]}], \"economics\": \"Suncor describes safety, reliability, efficiency and maintenance benefits of the AHS fleet but gives no quantified economic result in this page. These are automation benefits, not AI-attributed returns.\", \"timeline\": [{\"date\": \"2018\", \"event\": \"Suncor says Base Plant operations began using autonomous haul trucks.\", \"source_ids\": [\"RS036\"]}, {\"date\": \"undated\", \"event\": \"The page snapshot reports almost 120 trucks across two integrated mines and a prospective target of 140 by an unspecified year-end.\", \"source_ids\": [\"RS036\"]}], \"limitations\": [\"The page does not mention AI; classify as autonomous automation unless another technical source establishes an AI component.\", \"The source page is undated; the almost-120 snapshot and 140 target cannot be assigned to a publication year.\", \"The 140 count is planned, not deployed; the target year is not stated in the reviewed page.\", \"No quantified economics or workforce/job-impact result is disclosed in this source.\", \"Suncor's dynamic dispatch workflow is separate and is not included here.\"], \"assessment\": \"Analyst inference: the disclosure supports a large-scale autonomous mine-transport system with a human control-room response loop. It does not support inclusion in an AI-adoption count without evidence of an AI component.\", \"sources\": [{\"id\": \"RS036\", \"title\": \"The Future of Canadian Mining is Autonomous — and Suncor is Leading the Way\", \"url\": \"https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous\", \"publisher\": \"Suncor\", \"date\": null, \"type\": \"company feature\", \"locator\": \"Base Plant fleet, command centre and end-of-page target statement\", \"review_status\": \"full text reviewed\", \"supports\": [\"Suncor describes driverless AHS trucks using GPS, wireless/LTE communications and perceptive technologies; its page reports an almost-120 fleet snapshot across two mines, control-room oversight, and a prospective 140-truck target.\", \"The reviewed page contains no AI/artificial-intelligence claim and supplies no publication date; the target year therefore remains unspecified.\"], \"accessed\": \"2026-09-27\", \"publication_date_status\": \"undated page; accessed 2026-09-27\", \"original_id\": \"na-007\"}], \"field_sources\": {\"business_problem\": [\"RS036\"], \"workflow\": [\"RS036\"], \"technology_data\": [\"RS036\"], \"integration_human_role\": [\"RS036\"], \"deployment_scope\": [\"RS036\"], \"economics\": [\"RS036\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Remove the unsupported AI-optimization sentence and avoid inferring a page publication year or dated 140-truck target.\", \"related_existing_ids\": [], \"cohort\": \"north_america\", \"packet_id\": \"north_america-006\", \"source_ids\": [\"RS036\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-099\"]}", "Research_status": "Expanded source review", "Research_category": "Automation / AI not established", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Autonomous haul fleet “nears 140 trucks” in 2025.", "reason": "Official Suncor feature says nearly 120 were operating, with 140 planned by year end 2026; 140 is a target.", "source_url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"}], "Audit_review": {"id": "AI-099", "company": "Suncor Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "The reopened Suncor source describes GPS, wireless/LTE, perceptive technologies and an Autonomous Haulage System; it contains no AI/artificial-intelligence reference. It confirms automation deployment but does not substantiate classification as AI.", "correction": "Classify as conventional autonomy/automation context or exclude from AI-adoption counts unless an additional technical source documents an AI component. Keep autonomous fleet scale and human control-room oversight as supported facts.", "source_url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"}, {"severity": "low", "field": "Supplemental source review", "issue": "The International Mining article supports almost 120 current autonomous trucks at Base Plant and 2025 expansion, but neither AI/ML basis nor the 140-truck planned target is stated there.", "correction": "Keep autonomous haulage distinct from AI. If retaining 140 as a plan, source it separately; this secondary article supports near-120 current fleet, not that target.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}, {"severity": "high", "field": "Supplemental source review", "issue": "The Suncor page describes autonomous trucks using GPS, wireless/LTE and perception technologies, with control-room oversight. It contains no AI/artificial-intelligence reference. The page is undated, so neither a 2025 nor 2026 publication date can be inferred; 140 trucks is a future target with no dated year on the page.", "correction": "Replace the AI framing with 'Automation / AI not established'; retain the operating haulage evidence, nearly-120 snapshot, human control-room role and future 140 target without assigning a publication/year to the page. Remove the claim that AI optimizes driving conditions and unverified workforce job-loss claims from the disclosure.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "AI classification is not supported by the primary page; workflow and fleet facts are supported as autonomous automation."], "partition": 3, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Verified primary for autonomous automation; AI not established", "independent_validation": "Not established by this collection"}}, {"ID": "AI-100", "Company": "Phillips 66", "Company_primary": "Phillips 66", "Segment": "Refining", "HQ": "US", "Theme": "Operational AI", "Initiative": "Machine learning in refinery turnaround planning", "Extracted_finding": "Supports disciplined turnaround work planning so more maintenance can be prepared outside the outage window, potentially reducing duration and lost throughput.", "Context": "CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact. Management referenced lowering turnaround forecast mid-year and beating it, but did not attribute the forecast outcome to ML.", "Implementation": "CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact", "Assessment": "This is practical industrial decision support with human planners in the loop, but public evidence supports the use case only at a high level; no causal performance claim should be made.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "In use; launch date not disclosed", "Interpretation_limit": "Company-wide turnaround financial performance cannot be attributed specifically to ML. No model details, rollout date, refinery coverage or measured before/after result disclosed.", "Source_IDs": "R063", "Source_date": "2026-02-04", "Source_type": "earnings call transcript", "Source_locator": null, "Source_URL": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-02-04; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-063", "Initiative_ID": "AI-100", "Title": "Machine learning shortens refinery turnarounds", "Date": "2026-02", "Finding": "CEO Mark Lashier said Phillips 66 is taking work out of turnarounds and using tools 'like machine learning' to plan and execute them. This shortens duration and financial impact, and the company cut its turnaround spend forecast substantially mid-2025 and still beat it.", "Metric": "Turnaround spend forecast cut substantially mid-2025, then beaten", "Source_IDs": "R063", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "ML-driven turnaround scoping (moving work outside the outage window) is a direct availability and cost lever for refiners.", "Partners": "", "Original_date": "2026-02", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-063"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["midstream-006"], "Research_search": " {\"key\": \"p66-ml-turnaround-planning\", \"existing_ids\": [\"AI-100\"], \"related_existing_ids\": [], \"company\": \"Phillips 66\", \"title\": \"Machine learning in refinery turnaround planning\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"In use; launch date not disclosed\", \"practical_relevance\": \"Supports disciplined turnaround work planning so more maintenance can be prepared outside the outage window, potentially reducing duration and lost throughput.\", \"business_problem\": \"CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact. Management referenced lowering turnaround forecast mid-year and beating it, but did not attribute the forecast outcome to ML.\", \"workflow\": \"CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact\", \"technology_data\": \"Machine learning is named; training data, models, software, scope of work and integration not disclosed.\", \"integration_human_role\": \"Turnaround planners and operating/maintenance teams use tools in planning; no autonomous work approvals or workforce changes disclosed.\", \"deployment_scope\": \"Company turnaround program; no refinery count, units or adoption coverage disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No savings attributable to ML. Management cited lower turnaround forecast/spend and shorter-duration goals without a quantified AI-specific bridge.\", \"timeline\": [{\"date\": \"2026-02-04\", \"event\": \"CEO discussed ML use on Q4 2025 earnings call.\", \"source_ids\": [\"R063\"]}], \"limitations\": [\"Company-wide turnaround financial performance cannot be attributed specifically to ML.\", \"No model details, rollout date, refinery coverage or measured before/after result disclosed.\"], \"assessment\": \"This is practical industrial decision support with human planners in the loop, but public evidence supports the use case only at a high level; no causal performance claim should be made.\", \"sources\": [{\"id\": \"R063\", \"title\": \"Phillips 66 Q4 2025 earnings call transcript\", \"url\": \"https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf\", \"publisher\": \"Phillips 66\", \"date\": \"2026-02-04\", \"type\": \"earnings call transcript\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"CEO describes machine learning tools in turnaround planning, moving work outside turnaround to shorten duration/financial impact; company says forecast/spend improved, does not isolate AI causation or savings.\"], \"original_id\": \"p66q425\"}], \"field_sources\": {\"business_problem\": [\"R063\"], \"workflow\": [\"R063\"], \"technology_data\": [\"R063\"], \"integration_human_role\": [\"R063\"], \"deployment_scope\": [\"R063\"], \"economics\": [\"R063\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Original existing claim that tools including ML were used is confirmed by issuer transcript. The earnings/spend improvement is not proven to have been caused by AI.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-006\", \"source_ids\": [\"R063\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-100\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-100", "company": "Phillips 66", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-101", "Company": "Phillips 66", "Company_primary": "Phillips 66", "Segment": "Refining", "HQ": "US", "Theme": "Enterprise AI", "Initiative": "Company AI program and frontline solution replication", "Extracted_finding": "Creates a repeatable path for frontline engineers to identify data-heavy operating problems and adapt solutions across functions rather than relying only on isolated pilots.", "Context": "CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years. Refining leadership describes finding opportunities in a large set of cost-out work.", "Implementation": "CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years", "Assessment": "Enterprise enablement and frontline participation are meaningful deployment signals, but absent case-level evidence the program itself is an adoption architecture, not proof of realized value.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Broad program described as active; use cases scaling", "Interpretation_limit": "Broad program description, without use-case inventory or adoption metric. “Days and weeks” is a qualitative deployment-speed comparison, not independently measured performance. Cost-out portfolio performance is not attributed to AI.", "Source_IDs": "R064", "Source_date": "2026-08", "Source_type": "earnings call transcript", "Source_locator": null, "Source_URL": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-08; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-064", "Initiative_ID": "AI-101", "Title": "Company AI program 'normalizing' AI for front line", "Date": "2026-08", "Finding": "Lashier said Phillips 66 has implemented an AI program that is 'normalizing the use of AI inside the company', with dedicated staff finding use cases and extending them across the organization. Front-line engineers use AI to extract data and deploy fixes in 'days and weeks' instead of months or years. Refining leadership said AI helps it find opportunities hidden in 'the fog of too much data' across about 200 cost-out projects.", "Metric": "About 200 cost-reduction projects in flight; refining cost cut by more than $1/bbl, heading to $1.50/bbl (AI one of several contributors)", "Source_IDs": "R064", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A central use-case team that replicates front-line wins across sites, described as 'human-centered', is a scalable refinery AI operating model.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-064"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-08", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["midstream-007"], "Research_search": " {\"key\": \"p66-company-ai-program\", \"existing_ids\": [\"AI-101\"], \"related_existing_ids\": [], \"company\": \"Phillips 66\", \"title\": \"Company AI program and frontline solution replication\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Broad program described as active; use cases scaling\", \"practical_relevance\": \"Creates a repeatable path for frontline engineers to identify data-heavy operating problems and adapt solutions across functions rather than relying only on isolated pilots.\", \"business_problem\": \"CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years. Refining leadership describes finding opportunities in a large set of cost-out work.\", \"workflow\": \"CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years\", \"technology_data\": \"AI program includes AI and ML; specific LLM/model/vendor/data architecture not stated. Company notes ML has long been used for maintenance/turnarounds.\", \"integration_human_role\": \"Frontline engineers are central to discovering, building and deploying solutions; program staff support use-case reuse. Human-centered framing; no autonomous operating authority disclosed.\", \"deployment_scope\": \"Company-wide aspiration and multiple functions described, but no count of use cases, facilities, users or deployed models disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No quantified AI-specific savings; references to cost-out projects are not equivalent to AI-attributed savings.\", \"timeline\": [{\"date\": \"2026-08\", \"event\": \"CEO and refining leadership described AI normalization and frontline adoption on Q2 2026 earnings call.\", \"source_ids\": [\"R064\"]}], \"limitations\": [\"Broad program description, without use-case inventory or adoption metric.\", \"“Days and weeks” is a qualitative deployment-speed comparison, not independently measured performance.\", \"Cost-out portfolio performance is not attributed to AI.\"], \"assessment\": \"Enterprise enablement and frontline participation are meaningful deployment signals, but absent case-level evidence the program itself is an adoption architecture, not proof of realized value.\", \"sources\": [{\"id\": \"R064\", \"title\": \"Phillips 66 Q2 2026 earnings call transcript\", \"url\": \"https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf\", \"publisher\": \"Phillips 66\", \"date\": \"2026-08\", \"type\": \"earnings call transcript\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"CEO describes broad AI program, copying use cases, frontline engineers extracting data and deploying solutions more quickly; leadership says ML long used for maintenance/turnarounds.\"], \"original_id\": \"p66q226\"}], \"field_sources\": {\"business_problem\": [\"R064\"], \"workflow\": [\"R064\"], \"technology_data\": [\"R064\"], \"integration_human_role\": [\"R064\"], \"deployment_scope\": [\"R064\"], \"economics\": [\"R064\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-007\", \"source_ids\": [\"R064\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-101\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-101", "company": "Phillips 66", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-102", "Company": "Valero Energy", "Company_primary": "Valero Energy", "Segment": "Refining", "HQ": "US", "Theme": "Operational AI", "Initiative": "Early AI/ML applications and operational-data foundation", "Extracted_finding": "A data foundation and a few selectively deployed applications could support operational availability, but public details do not identify the specific decisions or refinery systems affected.", "Context": "Valero's refining SVP Greg Bram said Valero is evaluating several AI and ML applications and has deployed some to improve availability. He stressed that success depends on high-quality operating data, which Valero began standardizing 10-15 years ago. Valero also uses robotics for tank cleaning, drones for inspections, and control-room consolidation.", "Implementation": "Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "A decade of consistent operating-data practices is the prerequisite for refinery AI; start with availability and inspection use cases.", "Reported_metric": "No quantified metric supplied", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Transcript is third-party hosted, not a company-primary transcript; attributed wording is secondary-only. Official webcast page confirms event, not the reported AI claims. Existing details about availability, model scope and data history have not been verified against official transcript/audio.", "Source_IDs": "R065; RS101", "Source_date": "2025-10-23", "Source_type": "secondary transcript", "Source_locator": null, "Source_URL": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2025-10-23; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-065", "Initiative_ID": "AI-102", "Title": "'Cautiously optimistic' on AI; data quality first", "Date": "2025-10", "Finding": "Valero's refining SVP Greg Bram said Valero is evaluating several AI and ML applications and has deployed some to improve availability. He stressed that success depends on high-quality operating data, which Valero began standardizing 10-15 years ago. Valero also uses robotics for tank cleaning, drones for inspections, and control-room consolidation.", "Metric": "No quantified metric supplied", "Source_IDs": "R065", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A decade of consistent operating-data practices is the prerequisite for refinery AI; start with availability and inspection use cases.", "Partners": "", "Original_date": "2025-10", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-065"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-10-23", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["midstream-008"], "Research_search": " {\"key\": \"valero-early-ml-applications\", \"existing_ids\": [\"AI-102\"], \"related_existing_ids\": [], \"company\": \"Valero Energy\", \"title\": \"Early AI/ML applications and operational-data foundation\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"A few applications reported; scale unclear\", \"practical_relevance\": \"A data foundation and a few selectively deployed applications could support operational availability, but public details do not identify the specific decisions or refinery systems affected.\", \"business_problem\": \"On the Q3 2025 call, the refining SVP reportedly said the company was evaluating uses for tangible value and had deployed techniques in a few applications; he emphasized high-quality standardized operating data as a prerequisite.\", \"workflow\": \"On the Q3 2025 call, the refining SVP reportedly said the company was evaluating uses for tangible value and had deployed techniques in a few applications; he emphasized high-quality standardized operating data as a prerequisite\", \"technology_data\": \"Machine learning/artificial intelligence named; data standards are described as foundational, with no architecture or model detail.\", \"integration_human_role\": \"Operational teams are implied users; transcript does not clarify roles, control boundaries or human review.\", \"deployment_scope\": \"A few unspecified applications; no site or business-unit scope disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No outcome or economics quantified.\", \"timeline\": [{\"date\": \"2025-10-23\", \"event\": \"Q3 earnings call discussion captured in a third-party transcript; issuer event page confirms call date.\", \"source_ids\": [\"R065\", \"RS101\"]}], \"limitations\": [\"Transcript is third-party hosted, not a company-primary transcript; attributed wording is secondary-only.\", \"Official webcast page confirms event, not the reported AI claims.\", \"Existing details about availability, model scope and data history have not been verified against official transcript/audio.\"], \"assessment\": \"The company appears cautious and data-readiness-focused, but this profile remains a secondary-transcript lead until the original webcast/audio is checked; do not elevate the reported “few applications” into a verified deployment footprint.\", \"sources\": [{\"id\": \"R065\", \"title\": \"Valero Q3 2025 earnings call transcript (The Motley Fool)\", \"url\": \"https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/\", \"publisher\": \"The Motley Fool\", \"date\": \"2025-10-23\", \"type\": \"secondary transcript\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"SVP says a few AI/ML applications deployed or evaluated for tangible value; stresses operational-data quality. Not issuer-hosted; primary webcast page separately confirms call date.\"], \"original_id\": \"vlo\"}, {\"id\": \"RS101\", \"title\": \"Valero Q3 2025 earnings call event\", \"url\": \"https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx\", \"publisher\": \"Valero\", \"date\": \"2025-10-23\", \"type\": \"investor relations event page\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Confirms official call event, not transcript content.\"], \"original_id\": \"vloir\"}], \"field_sources\": {\"business_problem\": [\"R065\"], \"workflow\": [\"R065\"], \"technology_data\": [\"R065\"], \"integration_human_role\": [\"R065\"], \"deployment_scope\": [\"R065\"], \"economics\": [\"R065\"]}, \"verification\": \"secondary only\", \"corrections_to_existing\": \"Do not treat AI-102 as primary verified. Earnings transcript source is third party; issuer event page only confirms call date.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-008\", \"source_ids\": [\"R065\", \"RS101\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-102\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": {"id": "AI-102", "company": "Valero Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "source quality", "issue": "Valero Q3 issuer event page only verifies that a call took place; the detailed AI-readiness statements are from a secondary transcript.", "correction": "Attribute detailed comments to secondary transcription until verified against an issuer recording/transcript.", "source_url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-103", "Company": "Williams", "Company_primary": "Williams", "Segment": "Midstream", "HQ": "US", "Theme": "Operational AI", "Initiative": "Sequent AI-assisted gas marketing test", "Extracted_finding": "A narrowly described human-versus-model test in physical gas marketing illustrates a measurable decision-support experiment. Its practical value is possible improvement in how pipeline/storage capacity and local gas positions are marketed, while the disclosure does not say this test became a live execution workflow.", "Context": "Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs. The model score was compared with an experienced Sequent gas trader using a hindsight “perfect execution” scoring frame.", "Implementation": "Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs", "Assessment": "Concrete test with a crisp comparison, but small scope and hindsight benchmark limit inference; meaningful signal of commercial AI exploration, not proof of a scaled trading system.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Pilot / one-month test disclosed", "Interpretation_limit": "The test is described as occurring in “last year” relative to August 2025; it may fall inside or outside the 2024-09-26 start of the review window. Exact test date is not disclosed. This is secondary? The transcript is issuer-hosted and primary, so the claim is company-reported; score benchmark is not independently validated. No evidence that trading decisions became automated or that result generalized to other markets.", "Source_IDs": "RS091; RS092; R067", "Source_date": "2025-08-05", "Source_type": "earnings call transcript", "Source_locator": null, "Source_URL": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-08-05; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Trading & marketing", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-067", "Initiative_ID": "AI-103", "Title": "New CIO; AI model out-trades 10-year gas trader", "Date": "2025-08", "Finding": "CEO Chad Zamarin said Williams hired a new chief information officer with 'a lot of passion around AI' to drive cost efficiency and business transformation, including AI in project execution and analytics in market intelligence. In a month-long internal test marketing gas in Dallas-Fort Worth, young data-analytics staff building AI models beat a 10-year Sequent trader. Williams is also applying data to maintenance strategy.", "Metric": "AI modeler scored 96% of perfect-foresight execution vs. 94% for the veteran trader", "Source_IDs": "R067", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Head-to-head pilots against top human performers are a persuasive way to build the internal case for AI in trading.", "Partners": "", "Original_date": "2025-08", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-067"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-08-05", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["midstream-001"], "Research_search": " {\"key\": \"williams-sequent-gas-marketing-ai\", \"existing_ids\": [\"AI-103\"], \"related_existing_ids\": [], \"company\": \"Williams\", \"title\": \"Sequent AI-assisted gas marketing test\", \"department\": \"Trading & marketing\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Pilot / one-month test disclosed\", \"practical_relevance\": \"A narrowly described human-versus-model test in physical gas marketing illustrates a measurable decision-support experiment. Its practical value is possible improvement in how pipeline/storage capacity and local gas positions are marketed, while the disclosure does not say this test became a live execution workflow.\", \"business_problem\": \"Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs. The model score was compared with an experienced Sequent gas trader using a hindsight “perfect execution” scoring frame.\", \"workflow\": \"Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs\", \"technology_data\": \"AI models built by internal data analytics staff; specific model type, features, training data and infrastructure not disclosed.\", \"integration_human_role\": \"Experienced trader remained the benchmark and human execution authority is not explicitly described; the account is a retrospective anecdote from the CEO.\", \"deployment_scope\": \"One unique DFW asset, one month; company said it was a prior-year experiment. Broader deployment is not established.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"Not disclosed; no revenue uplift, reduced cost or customer outcome quantified.\", \"timeline\": [{\"date\": \"2025-08-05\", \"event\": \"CEO described prior-year one-month DFW test during Q2 2025 earnings call.\", \"source_ids\": [\"RS091\"]}], \"limitations\": [\"The test is described as occurring in “last year” relative to August 2025; it may fall inside or outside the 2024-09-26 start of the review window. Exact test date is not disclosed.\", \"This is secondary? The transcript is issuer-hosted and primary, so the claim is company-reported; score benchmark is not independently validated.\", \"No evidence that trading decisions became automated or that result generalized to other markets.\"], \"assessment\": \"Concrete test with a crisp comparison, but small scope and hindsight benchmark limit inference; meaningful signal of commercial AI exploration, not proof of a scaled trading system.\", \"sources\": [{\"id\": \"RS091\", \"title\": \"Williams Q2 2025 earnings call transcript\", \"url\": \"https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787\", \"publisher\": \"Williams\", \"date\": \"2025-08-05\", \"type\": \"earnings call transcript\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Sequent gas-marketing AI anecdote; CEO describes one-month DFW test, model score 96% vs trader 94% against a 100% hindsight benchmark; narrow test only.\"], \"original_id\": \"wmb25call\"}, {\"id\": \"RS092\", \"title\": \"Williams Q2 2025 earnings event page\", \"url\": \"https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025\", \"publisher\": \"Williams\", \"date\": \"2025-08-05\", \"type\": \"investor relations event page\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Confirms issuer-hosted earnings materials including transcript and presentation.\"], \"original_id\": \"wmb25ir\"}], \"field_sources\": {\"business_problem\": [\"RS091\"], \"workflow\": [\"RS091\"], \"technology_data\": [\"RS091\"], \"integration_human_role\": [\"RS091\"], \"deployment_scope\": [\"RS091\"], \"economics\": [\"RS091\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Corrects existing secondary transcript attribution: Williams itself hosts the earnings transcript and reports one-month AI model score 96% versus trader 94% in a single DFW test; does not substantiate broad scaled trading or realized financial value.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-001\", \"source_ids\": [\"RS091\", \"RS092\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-103\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-103", "company": "Williams", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-104", "Company": "Kinder Morgan", "Company_primary": "Kinder Morgan", "Segment": "Midstream", "HQ": "US", "Theme": "Operational AI", "Initiative": "Strategic investment in Flyscan AI-assisted aerial inspection", "Extracted_finding": "The investment backs development of remote-sensing inspection intended to detect leaks and right-of-way encroachment from routine aircraft patrols. It does not establish that KMI has deployed the system across its own pipeline network.", "Context": "Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas. KMI joined the strategic investor group in January 2025 and described the technology positively.", "Implementation": "Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas", "Assessment": "A concrete adjacent technology investment, but not evidence of operational deployment by KMI or quantified leak-detection improvements.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Strategic investor; operational KMI rollout not established", "Interpretation_limit": "Source is Flyscan’s own announcement; primary evidence for the partnership but vendor claims about product capability are not independent. KM’s Q1 2025 earnings release reviewed without further AI rollout disclosure. Do not conflate Flyscan’s overall fleet/operator metrics with KMI usage.", "Source_IDs": "RS102; RS103; R068", "Source_date": "2025-01-14", "Source_type": "partner announcement", "Source_locator": null, "Source_URL": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-01-14; strategic investment announced, KMI operating rollout not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-068", "Initiative_ID": "AI-104", "Title": "Strategic investment in Flyscan AI leak detection", "Date": "2025-01", "Finding": "Kinder Morgan joined Flyscan Systems' strategic investor group alongside Marathon Petroleum, Enbridge and Hatch. Flyscan uses remote sensing and AI to detect liquid hydrocarbon leaks and automate right-of-way inspection from patrol aircraft, and is expanding into natural-gas detection. KM COO James Holland called the technology 'state-of-the-art'.", "Metric": "No quantified metric supplied", "Source_IDs": "R068", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Midstream operators are using corporate venture stakes to shape AI inspection tools for their own networks.", "Partners": "Flyscan Systems", "Original_date": "2025-01", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-068"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-04-16", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["midstream-009"], "Research_search": " {\"key\": \"kmi-flyscan-investment\", \"existing_ids\": [\"AI-104\"], \"related_existing_ids\": [], \"company\": \"Kinder Morgan\", \"title\": \"Strategic investment in Flyscan AI-assisted aerial inspection\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Strategic investor; operational KMI rollout not established\", \"practical_relevance\": \"The investment backs development of remote-sensing inspection intended to detect leaks and right-of-way encroachment from routine aircraft patrols. It does not establish that KMI has deployed the system across its own pipeline network.\", \"business_problem\": \"Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas. KMI joined the strategic investor group in January 2025 and described the technology positively.\", \"workflow\": \"Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas\", \"technology_data\": \"Vendor describes remote sensing and AI image analysis; sensors/model specifics and independent validation not disclosed.\", \"integration_human_role\": \"Inspection outputs would support follow-up investigation; KMI’s exact review or response workflow is not disclosed.\", \"deployment_scope\": \"Partnership/investment is confirmed; locations, miles, flights or active KMI use are not disclosed.\", \"partners\": [\"Flyscan Systems\", \"Marathon Petroleum\", \"Enbridge\", \"Hatch\"], \"disclosed_results\": [], \"economics\": \"Investment amount and economic return not disclosed.\", \"timeline\": [{\"date\": \"2025-01-14\", \"event\": \"Flyscan announced Kinder Morgan joined its strategic investor group.\", \"source_ids\": [\"RS102\"]}], \"limitations\": [\"Source is Flyscan’s own announcement; primary evidence for the partnership but vendor claims about product capability are not independent.\", \"KM’s Q1 2025 earnings release reviewed without further AI rollout disclosure.\", \"Do not conflate Flyscan’s overall fleet/operator metrics with KMI usage.\"], \"assessment\": \"A concrete adjacent technology investment, but not evidence of operational deployment by KMI or quantified leak-detection improvements.\", \"sources\": [{\"id\": \"RS102\", \"title\": \"Flyscan announces Kinder Morgan as strategic investor\", \"url\": \"https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/\", \"publisher\": \"Flyscan Systems\", \"date\": \"2025-01-14\", \"type\": \"partner announcement\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"KMI joins strategic investor group; vendor describes remote sensing and AI visual patrol/leak detection. This confirms investment/partnership, not KMI rollout.\"], \"original_id\": \"fly\"}, {\"id\": \"RS103\", \"title\": \"Kinder Morgan Q1 2025 earnings release\", \"url\": \"https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx\", \"publisher\": \"Kinder Morgan\", \"date\": \"2025-04-16\", \"type\": \"earnings release\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Reviewed as in-window operating/financial source; no Flyscan or internal AI deployment disclosed in this release.\"], \"original_id\": \"kmq125\"}], \"field_sources\": {\"business_problem\": [\"RS102\"], \"workflow\": [\"RS102\"], \"technology_data\": [\"RS102\"], \"integration_human_role\": [\"RS102\"], \"deployment_scope\": [\"RS102\"], \"economics\": [\"RS102\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"Existing record’s investment/partnership claim is supported; operational adoption and natural-gas rollout remain unverified.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-009\", \"source_ids\": [\"RS102\", \"RS103\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-104\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one supporting long document was reviewed at relevant excerpts only.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-104", "company": "Kinder Morgan", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx"], "findings": [{"severity": "medium", "field": "Supplemental source review", "issue": "The profile's cited Flyscan URL is dated 2025-01-14, while Source_date is 2025-04-16, the date of the separate KMI Q1 earnings release.", "correction": "Set the initiative Source_date to 2025-01-14 and keep the profile's caveat that investment/partnering does not establish KMI deployment.", "source_url": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-105", "Company": "SLB", "Company_primary": "SLB", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Enterprise AI", "Initiative": "SLB–NVIDIA AI Factory for Energy and modular AI infrastructure", "Extracted_finding": "Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.", "Context": "Provide energy companies with infrastructure and model environments to scale AI across operational data and workflows.", "Implementation": "Three tracks: modular design partnership for NVIDIA DSX AI factories; develop AI Factory for Energy reference environment; optimize processing across SLB digital platforms.", "Assessment": "Analyst inference: Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Collaboration expanded March 2026; design and development announced", "Interpretation_limit": "The data-center modular design work is AI-demand infrastructure; separate from operator AI adoption. AI Factory for Energy is announced development, not established deployment.", "Source_IDs": "RS068; R071", "Source_date": "2026-03-25", "Source_type": "company announcement", "Source_locator": "Three strategic elements", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-03-25; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-071", "Initiative_ID": "AI-105", "Title": "SLB and NVIDIA build 'AI Factory for Energy'", "Date": "2026-03", "Finding": "SLB expanded its NVIDIA collaboration into three areas: modular DSX AI-factory data centers, an 'AI Factory for Energy' reference environment with domain generative models and industrial-scale agentic AI on SLB platforms, and accelerated computing across SLB software. SLB will make modular data-center components at its Louisiana technology center.", "Metric": "3.1 million sq ft Louisiana manufacturing center; partnership dates to 2008", "Source_IDs": "R071", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Oilfield services companies are bundling compute infrastructure with domain AI, so pure-software AI vendors should partner with them for delivery rather than compete on infrastructure.", "Partners": "NVIDIA, Lumi, Delfi", "Original_date": "2026-03", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-071"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-03-25", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-032"], "Research_search": " {\"key\": \"slb-nvidia-ai-factory\", \"existing_ids\": [\"AI-105\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"SLB–NVIDIA AI Factory for Energy and modular AI infrastructure\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Collaboration expanded March 2026; design and development announced\", \"practical_relevance\": \"Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.\", \"business_problem\": \"Provide energy companies with infrastructure and model environments to scale AI across operational data and workflows.\", \"workflow\": \"Three tracks: modular design partnership for NVIDIA DSX AI factories; develop AI Factory for Energy reference environment; optimize processing across SLB digital platforms.\", \"technology_data\": \"NVIDIA AI infrastructure, Omniverse libraries and Nemotron open models; SLB Delfi/Lumi digital platforms and domain data/models.\", \"integration_human_role\": \"No operating workflow or human control details disclosed; reference environment remains in development.\", \"deployment_scope\": \"Commercial infrastructure design and AI reference environment ambitions; no completed customer deployment named in announcement.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No deal value, capex, customer economics or production outcome disclosed.\", \"timeline\": [{\"date\": \"2026-03-25\", \"event\": \"SLB and NVIDIA announce expanded collaboration in three workstreams.\", \"source_ids\": [\"RS068\"]}], \"limitations\": [\"The data-center modular design work is AI-demand infrastructure; separate from operator AI adoption.\", \"AI Factory for Energy is announced development, not established deployment.\"], \"assessment\": \"Analyst inference: Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.\", \"sources\": [{\"id\": \"RS068\", \"title\": \"SLB Industrializes AI for the Energy Industry with NVIDIA\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/\", \"publisher\": \"SLB\", \"date\": \"2026-03-25\", \"type\": \"company announcement\", \"locator\": \"Three strategic elements\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB/NVIDIA announced modular DSX data-center design, an energy AI reference environment using generative and agentic AI, and accelerated computing for SLB digital platforms.\"], \"original_id\": \"slb_nvidia\"}], \"field_sources\": {\"business_problem\": [\"RS068\"], \"workflow\": [\"RS068\"], \"technology_data\": [\"RS068\"], \"integration_human_role\": [\"RS068\"], \"deployment_scope\": [\"RS068\"], \"economics\": [\"RS068\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-032\", \"source_ids\": [\"RS068\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-105\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-105", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/"], "findings": [{"severity": "medium", "field": "stage", "issue": "SLB/NVIDIA source describes industrialization/partnership and product infrastructure, not measured upstream operator outcomes.", "correction": "Present as commercial product/infrastructure collaboration; no AI adoption result inferred.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-106", "Company": "SLB / Tachyus", "Company_primary": "SLB", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Agreement to acquire Tachyus reservoir modeling and optimization software", "Extracted_finding": "A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.", "Context": "Enable faster tactical reservoir management in response to changing operations and improve waterflood, production and recovery decisions.", "Implementation": "Aqueon evaluates thousands of reservoir scenarios in minutes for waterflood management, pressure forecasting, production optimization, saltwater disposal and EOR; post-close SLB plans Delfi/Lumi integration for closed-loop reservoir/production workflows.", "Assessment": "Analyst inference: A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.", "Reported_metric": "Wells using Aqueon: >7,500 (Vendor-reported)", "Stage_at_disclosure": "Acquisition completed in Q2 2026; operational integration into Delfi/Lumi not yet specifically verified.", "Interpretation_limit": "SLB plans Delfi/Lumi integration; acquisition close is verified, but completed technical integration and customer deployment are not. 7,500+ wells is vendor-reported installed scope, not independent outcome validation.", "Source_IDs": "RS069; RS070; R073", "Source_date": "2026-07-24", "Source_type": "company announcement", "Source_locator": "Aqueon workflows, deployment and planned integration", "Source_URL": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-073", "Initiative_ID": "AI-106", "Title": "SLB acquires Tachyus physics-ML reservoir startup", "Date": "2026-05", "Finding": "SLB agreed to acquire Houston-based Tachyus, whose Aqueon platform combines machine learning with reservoir physics for waterflood management, pressure forecasting and production optimization. SLB plans to fold it into Delfi and Lumi. Terms were not disclosed.", "Metric": "Aqueon deployed on 7,500+ wells; customers include ConocoPhillips, Shell, Equinor, PDO", "Source_IDs": "R073", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Startups with proven physics-plus-ML products and named major customers are being bought by oilfield services companies, which makes an exit to SLB, Halliburton or Baker Hughes a realistic plan.", "Partners": "Tachyus Aqueon, Delfi, Lumi, ConocoPhillips, Shell, Equinor, Petroleum Development Oman", "Original_date": "2026-05", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-073"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-033"], "Research_search": " {\"key\": \"slb-tachyus\", \"existing_ids\": [\"AI-106\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"Agreement to acquire Tachyus reservoir modeling and optimization software\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Acquisition completed in Q2 2026; operational integration into Delfi/Lumi not yet specifically verified.\", \"practical_relevance\": \"A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.\", \"business_problem\": \"Enable faster tactical reservoir management in response to changing operations and improve waterflood, production and recovery decisions.\", \"workflow\": \"Aqueon evaluates thousands of reservoir scenarios in minutes for waterflood management, pressure forecasting, production optimization, saltwater disposal and EOR; post-close SLB plans Delfi/Lumi integration for closed-loop reservoir/production workflows.\", \"technology_data\": \"Machine learning combined with reservoir physics; Aqueon models reservoir behavior under operating conditions.\", \"integration_human_role\": \"Reservoir and production teams adjust field strategy; post-acquisition decision autonomy is not specified.\", \"deployment_scope\": \"SLB announced a proposed software acquisition; Aqueon vendor says deployed across 7,500+ wells globally.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Wells using Aqueon\", \"value\": \">7,500\", \"period\": \"At May 2026 announcement\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS069\"]}], \"economics\": \"Terms not disclosed; no AI-specific customer ROI reported.\", \"timeline\": [{\"date\": \"2026-05-29\", \"event\": \"SLB announces agreement to acquire Tachyus.\", \"source_ids\": [\"RS069\"]}, {\"date\": \"2026-07-24\", \"event\": \"SLB Q2 results confirm the Tachyus acquisition completed during the quarter.\", \"source_ids\": [\"RS070\"]}, {\"date\": \"2026-07-24\", \"event\": \"SLB Q2 results confirm the Tachyus acquisition completed during the quarter.\", \"source_ids\": [\"RS070\"]}], \"limitations\": [\"SLB plans Delfi/Lumi integration; acquisition close is verified, but completed technical integration and customer deployment are not.\", \"7,500+ wells is vendor-reported installed scope, not independent outcome validation.\"], \"assessment\": \"Analyst inference: A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.\", \"sources\": [{\"id\": \"RS069\", \"title\": \"SLB to Acquire Tachyus\", \"url\": \"https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update\", \"publisher\": \"SLB\", \"date\": \"2026-05-29\", \"type\": \"company announcement\", \"locator\": \"Aqueon workflows, deployment and planned integration\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB describes Aqueon as combining machine learning and reservoir physics for operational reservoir-management tasks and reports deployment across 7,500+ wells.\", \"SLB plans post-close integration into Delfi and Lumi; close remained subject to customary conditions at announcement.\"], \"original_id\": \"slb_tachyus\"}, {\"id\": \"RS070\", \"title\": \"SLB Announces Second-Quarter 2026 Results\", \"url\": \"https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"company announcement\", \"locator\": \"Acquisition completion\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB reported that it completed the Tachyus acquisition during Q2 2026.\"], \"original_id\": \"slb_tachyus_completed\"}], \"field_sources\": {\"business_problem\": [\"RS069\"], \"workflow\": [\"RS069\"], \"technology_data\": [\"RS069\"], \"integration_human_role\": [\"RS069\"], \"deployment_scope\": [\"RS069\"], \"economics\": [\"RS069\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-033\", \"source_ids\": [\"RS069\", \"RS070\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-106\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-106", "company": "SLB / Tachyus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update"], "findings": [{"severity": "medium", "field": "stage", "issue": "Tachyus is described in an acquisition agreement; announcement does not disclose consideration or establish realized SLB integration outcomes.", "correction": "Keep as announced acquisition, distinguish subsequent closing if separately sourced.", "source_url": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-107", "Company": "SLB / S&P Global Energy", "Company_primary": "SLB", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Agreement to acquire S&P Global Energy upstream software portfolio and co-develop AI models", "Extracted_finding": "The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.", "Context": "Extend SLB digital subsurface coverage into widely used U.S. onshore/unconventional software workflows and connect planning tools with AI.", "Implementation": "Portfolio includes Kingdom, Petra, Harmony Enterprise and other engineering tools; separate collaboration plans AI models using S&P upstream data with Lumi and Tela.", "Assessment": "Analyst inference: The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Acquisition completed September 1, 2026; AI model and platform integration remains prospective.", "Interpretation_limit": "Acquisition is complete, but the announced AI model collaboration and integration work remain prospective in the reviewed evidence. Do not frame the acquired software installed base as SLB AI adoption.", "Source_IDs": "RS071; RS072; R074", "Source_date": "2026-09-01", "Source_type": "company announcement", "Source_locator": "Transaction scope and planned AI collaboration", "Source_URL": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-09-01; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-074", "Initiative_ID": "AI-107", "Title": "SLB buys Kingdom, Petra upstream software for agentic AI", "Date": "2026-04", "Finding": "SLB agreed to acquire S&P Global Energy's geoscience and petroleum-engineering software (Kingdom, Petra, Harmony Enterprise), which is widely used by US onshore operators. The two companies will co-develop AI models on S&P's upstream data using Lumi and Tela. Closing is expected in H2 2026 or early 2027.", "Metric": "~110,000 users across ~4,000 organizations (reported); data spans 113 countries", "Source_IDs": "R074", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "SLB is buying the desktop tools independents already use in order to push agentic AI to US onshore operators, the same mid-market segment many AI startups target.", "Partners": "Kingdom, Petra, Harmony Enterprise, Lumi, Tela, S&P Global Energy", "Original_date": "2026-04", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-074"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-09-01", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-034"], "Research_search": " {\"key\": \"slb-sp-global-upstream\", \"existing_ids\": [\"AI-107\"], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"Agreement to acquire S&P Global Energy upstream software portfolio and co-develop AI models\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Acquisition completed September 1, 2026; AI model and platform integration remains prospective.\", \"practical_relevance\": \"The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.\", \"business_problem\": \"Extend SLB digital subsurface coverage into widely used U.S. onshore/unconventional software workflows and connect planning tools with AI.\", \"workflow\": \"Portfolio includes Kingdom, Petra, Harmony Enterprise and other engineering tools; separate collaboration plans AI models using S&P upstream data with Lumi and Tela.\", \"technology_data\": \"Existing geoscience/petroleum engineering software and upstream data; proposed integration with SLB Lumi and Tela.\", \"integration_human_role\": \"Existing user workflows to be maintained; AI decision authority or workflow design not yet disclosed.\", \"deployment_scope\": \"Agreement covers software portfolio; transaction not yet closed as of stated expected schedule.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Terms not disclosed; no AI model or customer economics disclosed.\", \"timeline\": [{\"date\": \"2026-04-24\", \"event\": \"SLB announces definitive acquisition agreement and separate AI model collaboration.\", \"source_ids\": [\"RS071\"]}, {\"date\": \"2026-09-01\", \"event\": \"S&P Global confirms divestiture completion to SLB.\", \"source_ids\": [\"RS072\"]}], \"limitations\": [\"Acquisition is complete, but the announced AI model collaboration and integration work remain prospective in the reviewed evidence.\", \"Do not frame the acquired software installed base as SLB AI adoption.\"], \"assessment\": \"Analyst inference: The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.\", \"sources\": [{\"id\": \"RS071\", \"title\": \"SLB Advances Digital Strategy with Acquisition of S&P Global Energy’s Upstream Software Portfolio\", \"url\": \"https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition\", \"publisher\": \"SLB\", \"date\": \"2026-04-24\", \"type\": \"company announcement\", \"locator\": \"Transaction scope and planned AI collaboration\", \"review_status\": \"full text reviewed\", \"supports\": [\"SLB signed a definitive agreement to acquire S&P Global Energy upstream geoscience/petroleum software; separately, the parties plan to build AI models using Lumi/Tela and S&P data.\", \"The announced close was expected H2 2026 or early 2027, subject to approvals.\"], \"original_id\": \"slb_spglobal\"}, {\"id\": \"RS072\", \"title\": \"S&P Global Completes Divestiture of Upstream Energy Software Portfolio to SLB\", \"url\": \"https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB\", \"publisher\": \"S&P Global\", \"date\": \"2026-09-01\", \"type\": \"company announcement\", \"locator\": \"Acquisition completion\", \"review_status\": \"full text reviewed\", \"supports\": [\"S&P Global announced completion of the upstream geoscience and petroleum engineering software portfolio divestiture to SLB on September 1, 2026.\", \"S&P Global said it would continue distributing its proprietary data through the divested workflow tools; financial terms were not disclosed.\"], \"original_id\": \"slb_spglobal_completed\"}], \"field_sources\": {\"business_problem\": [\"RS071\"], \"workflow\": [\"RS071\"], \"technology_data\": [\"RS071\"], \"integration_human_role\": [\"RS071\"], \"deployment_scope\": [\"RS071\"], \"economics\": [\"RS071\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-034\", \"source_ids\": [\"RS071\", \"RS072\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-107\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-107", "company": "SLB / S&P Global Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition"], "findings": [{"severity": "medium", "field": "stage", "issue": "The April acquisition announcement and September completion are separate events; AI-model co-development is a stated ambition.", "correction": "Distinguish acquired portfolio from future joint AI-model work.", "source_url": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-108", "Company": "Microsoft", "Company_primary": "Microsoft", "Segment": "Big tech", "HQ": "United States", "Theme": "Enterprise AI", "Initiative": "ADNOC–Microsoft co-development and deployment of operational AI agents", "Extracted_finding": "The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.", "Context": "Extend ADNOC’s existing enterprise AI program into operationally focused agents and autonomous workflows across the value chain.", "Implementation": "ADNOC and Microsoft plan to co-develop and deploy AI agents; Microsoft will also provide advanced AI tools and upskilling. The same agreement includes separate Masdar/XRG energy-supply work for Microsoft data-center growth.", "Assessment": "Analyst inference: The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Strategic agreement announced Nov. 2, 2025; co-development/deployment scope announced, no named agent deployment in the release", "Interpretation_limit": "Strategic agreement and plans, not evidence of agents already deployed. Separate energy supply for Microsoft data centers is infrastructure supporting AI demand, not an ADNOC operational AI workflow. Do not merge this partnership with ENERGYai (AI-040) or employee-training record (AI-071).", "Source_IDs": "RS030; R075", "Source_date": "2025-11-02", "Source_type": "company announcement", "Source_locator": "Operational AI agents and separate energy-for-AI component", "Source_URL": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-11-02; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-075", "Initiative_ID": "AI-108", "Title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "Date": "2025-11", "Finding": "ADNOC, Masdar, XRG and Microsoft signed an agreement to co-develop and deploy AI agents across ADNOC's value chain and to supply energy for Microsoft's data centers. The deal builds on ADNOC's enterprise-wide Microsoft Copilot rollout, which began in November 2023.", "Metric": "40,000+ employees AI-trained; Copilot utilization >90%; 70,000+ hours/month productivity gained", "Source_IDs": "R075", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Hyperscalers trade power-supply deals for AI footprint inside NOCs, and ADNOC's Copilot adoption numbers are a useful benchmark for workforce-enablement pitches.", "Partners": "Microsoft Copilot, Azure, ADNOC, Masdar, XRG", "Original_date": "2025-11", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-075"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-11-02", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["nocs-020"], "Research_search": " {\"key\": \"adnoc-microsoft-ai-agents\", \"existing_ids\": [\"AI-108\"], \"related_existing_ids\": [], \"company\": \"ADNOC\", \"title\": \"ADNOC–Microsoft co-development and deployment of operational AI agents\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Strategy only\", \"stage\": \"Strategic agreement announced Nov. 2, 2025; co-development/deployment scope announced, no named agent deployment in the release\", \"practical_relevance\": \"The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.\", \"business_problem\": \"Extend ADNOC’s existing enterprise AI program into operationally focused agents and autonomous workflows across the value chain.\", \"workflow\": \"ADNOC and Microsoft plan to co-develop and deploy AI agents; Microsoft will also provide advanced AI tools and upskilling. The same agreement includes separate Masdar/XRG energy-supply work for Microsoft data-center growth.\", \"technology_data\": \"Microsoft AI tools and ADNOC operational data/workflows; detailed agent architecture, integrations and model governance not disclosed.\", \"integration_human_role\": \"Autonomous operations are an ambition; human approvals, safety boundaries and production status for the agents are unspecified.\", \"deployment_scope\": \"ADNOC value chain-wide agreement; data-center energy work involves Masdar and XRG and is not itself AI adoption.\", \"partners\": [\"Microsoft\", \"Masdar\", \"XRG\"], \"disclosed_results\": [], \"economics\": \"No deal value, implementation budget, deployed-agent count or AI-specific results disclosed.\", \"timeline\": [{\"date\": \"2025-11-02\", \"event\": \"ADNOC and Microsoft announced AI-agent collaboration; Masdar/XRG separately joined the energy-for-AI component.\", \"source_ids\": [\"RS030\"]}], \"limitations\": [\"Strategic agreement and plans, not evidence of agents already deployed.\", \"Separate energy supply for Microsoft data centers is infrastructure supporting AI demand, not an ADNOC operational AI workflow.\", \"Do not merge this partnership with ENERGYai (AI-040) or employee-training record (AI-071).\"], \"assessment\": \"Analyst inference: The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.\", \"sources\": [{\"id\": \"RS030\", \"title\": \"ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI\", \"url\": \"https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai\", \"publisher\": \"ADNOC\", \"date\": \"2025-11-02\", \"type\": \"company announcement\", \"locator\": \"Operational AI agents and separate energy-for-AI component\", \"review_status\": \"full text reviewed\", \"supports\": [\"ADNOC and Microsoft announced plans to co-develop and deploy AI agents across ADNOC’s value chain; no particular agent or deployment was named.\", \"The wider release separately includes Masdar/XRG energy and infrastructure solutions for Microsoft AI/data-center growth.\"], \"original_id\": \"adnoc_msft_ai\"}], \"field_sources\": {\"business_problem\": [\"RS030\"], \"workflow\": [\"RS030\"], \"technology_data\": [\"RS030\"], \"integration_human_role\": [\"RS030\"], \"deployment_scope\": [\"RS030\"], \"economics\": [\"RS030\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"nocs\", \"packet_id\": \"nocs-020\", \"source_ids\": [\"RS030\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-108\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-108", "company": "Microsoft", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "strategy", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-109", "Company": "Halliburton", "Company_primary": "Halliburton", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Automation context", "Initiative": "Octiv Auto Frac autonomous stage execution with Coterra", "Extracted_finding": "A named customer rolled a real automated completion workflow beyond an initial pilot. This should be categorized as automation, not AI, because the source does not establish an AI algorithm.", "Context": "Execute customer-designed hydraulic-fracturing stages more consistently and reduce manual variability during pumping.", "Implementation": "Coterra configures design and job controls; Octiv Auto Frac automatically executes stage delivery via ZEUS, using Sensori for monitoring. Initial rollout expanded to remaining Halliburton-run Coterra Permian completion programs.", "Assessment": "Analyst inference: This is an operationally relevant autonomous frac workflow with a reported stage-efficiency result. The reviewed sources do not establish AI/ML, so it is classified as automation.", "Reported_metric": "stage efficiency increase: 17% (Vendor-reported)", "Stage_at_disclosure": "Initial launch Jan 2025; rollout across remaining Halliburton-executed Permian programs", "Interpretation_limit": "Stage-efficiency baseline/definition not given; all metrics are Halliburton-reported. The cited sources identify autonomous/intelligent automation but do not disclose an AI/ML method; treat as automation, not AI.", "Source_IDs": "RS048; R076", "Source_date": "2025-01-06", "Source_type": "company announcement", "Source_locator": "Deployment and results paragraphs", "Source_URL": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "Evidence_status": "verified partner", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-01-06; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Automation or digital workflow; explicit AI basis not established", "Disclosures": [{"ID": "REF-076", "Initiative_ID": "AI-109", "Title": "Octiv Auto Frac: first fully automated frac with Coterra", "Date": "2025-01", "Finding": "Halliburton and Coterra launched Octiv Auto Frac on the ZEUS electric frac platform. It automates stage execution, making thousands of pumping decisions from the operator's design without a human at the controls, and Coterra extended it to its remaining Permian completion programs.", "Metric": "17% increase in stage efficiency on initial rollout", "Source_IDs": "R076", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Autonomy in completions reaches the field through service companies that own the equipment and control loop, which leaves software-only players to compete on design, analytics and planning.", "Partners": "Coterra Energy, ZEUS, Sensori", "Original_date": "2025-01", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-076"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-01-06", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-011"], "Research_search": " {\"key\": \"halliburton-coterra-autofrac\", \"existing_ids\": [\"AI-109\"], \"related_existing_ids\": [], \"company\": \"Halliburton\", \"title\": \"Octiv Auto Frac autonomous stage execution with Coterra\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Automation / AI not established\", \"stage\": \"Initial launch Jan 2025; rollout across remaining Halliburton-executed Permian programs\", \"practical_relevance\": \"A named customer rolled a real automated completion workflow beyond an initial pilot. This should be categorized as automation, not AI, because the source does not establish an AI algorithm.\", \"business_problem\": \"Execute customer-designed hydraulic-fracturing stages more consistently and reduce manual variability during pumping.\", \"workflow\": \"Coterra configures design and job controls; Octiv Auto Frac automatically executes stage delivery via ZEUS, using Sensori for monitoring. Initial rollout expanded to remaining Halliburton-run Coterra Permian completion programs.\", \"technology_data\": \"Automated control system, equipment telemetry and fracture monitoring; autonomy is process automation, not a claim of AI.\", \"integration_human_role\": \"Operator retains design configuration/control. System executes stage-by-stage after customer setup; human oversight specifics not disclosed.\", \"deployment_scope\": \"Coterra’s Halliburton-executed Permian completions.\", \"partners\": [\"Coterra Energy\"], \"disclosed_results\": [{\"metric\": \"stage efficiency increase\", \"value\": \"17%\", \"period\": \"initial deployment, announced Jan 2025\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS048\"]}], \"economics\": \"Halliburton reports 17% higher stage efficiency in initial deployment. No absolute throughput/cost definition or independent validation.\", \"timeline\": [{\"date\": \"2025-01-06\", \"event\": \"Coterra launch and rollout to remaining Halliburton-executed Permian programs announced\", \"source_ids\": [\"RS048\"]}], \"limitations\": [\"Stage-efficiency baseline/definition not given; all metrics are Halliburton-reported.\", \"The cited sources identify autonomous/intelligent automation but do not disclose an AI/ML method; treat as automation, not AI.\"], \"assessment\": \"Analyst inference: This is an operationally relevant autonomous frac workflow with a reported stage-efficiency result. The reviewed sources do not establish AI/ML, so it is classified as automation.\", \"sources\": [{\"id\": \"RS048\", \"title\": \"Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program\", \"url\": \"https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program\", \"publisher\": \"Halliburton\", \"date\": \"2025-01-06\", \"type\": \"company announcement\", \"locator\": \"Deployment and results paragraphs\", \"review_status\": \"full text reviewed\", \"supports\": [\"Octiv Auto Frac automates stage delivery against customer-configured designs, using ZEUS and Sensori.\", \"Coterra was first operator to automate and control its frac design/execution; Halliburton reports 17% higher stage efficiency in initial rollout and rollout to remaining Halliburton-executed Permian completion programs.\"], \"original_id\": \"RS048\"}], \"field_sources\": {\"business_problem\": [\"RS048\"], \"workflow\": [\"RS048\"], \"technology_data\": [\"RS048\"], \"integration_human_role\": [\"RS048\"], \"deployment_scope\": [\"RS048\"], \"economics\": [\"RS048\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-011\", \"source_ids\": [\"RS048\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-109\"]}", "Research_status": "Expanded source review", "Research_category": "Automation / AI not established", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-109", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton primary release and World Oil account call the work automated/autonomous and explicitly describe intelligent automation; neither establishes an AI/ML method.", "correction": "Classify as automation evidence, not operational AI, unless another supporting source explicitly identifies AI/ML. Retain the 17% stage-efficiency result as vendor-reported automation result.", "source_url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program"}], "residual_gaps": ["No explicit AI method was found in the supporting primary or secondary sources checked."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-110", "Company": "Halliburton", "Company_primary": "Halliburton", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Automation context", "Initiative": "Closed-loop hydraulic fracturing with Chevron", "Extracted_finding": "A significant move from preset automation to feedback-driven closed-loop completions. Since the partner calls the capability intelligent fracturing but does not establish AI, do not classify as AI.", "Context": "Adapt frac delivery to local subsurface feedback while treatment is underway.", "Implementation": "ZEUS IQ couples automated OCTIV stage execution with Sensori fracture feedback; Chevron’s subsurface knowledge is embedded in an algorithm to adapt operations in real time.", "Assessment": "Analyst inference: Feedback-driven completions are a meaningful control capability; on reviewed evidence this is closed-loop automation rather than established AI.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Closed-loop autonomous operation achieved in Colorado by Q1 2025; platform described Apr 2025", "Interpretation_limit": "Deployment evidence comes from Halliburton/partner release; no measured performance result. Automation, not verified AI. The reviewed Halliburton source describes feedback-driven automation and a domain algorithm, but does not identify an AI/ML method.", "Source_IDs": "RS049; R077", "Source_date": "2025-04-12", "Source_type": "company announcement", "Source_locator": "Closed-loop intelligent fracturing description", "Source_URL": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "Evidence_status": "verified partner", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-04-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Automation or digital workflow; explicit AI basis not established", "Disclosures": [{"ID": "REF-077", "Initiative_ID": "AI-110", "Title": "ZEUS IQ closed-loop intelligent frac with Chevron", "Date": "2025-06", "Finding": "Halliburton debuted ZEUS IQ in May 2025, describing it as the industry's first closed-loop fracturing platform, combining Octiv Auto Frac with Sensori subsurface monitoring. With Chevron it ran feedback-driven completions in Colorado's DJ Basin, where the job plan was adjusted in real time without human intervention.", "Metric": "No quantified metric supplied", "Source_IDs": "R077", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The frontier has moved from alerts to closed-loop control, so AI builders need live subsurface feedback and actuation partners to compete.", "Partners": "Chevron, Octiv, Sensori", "Original_date": "2025-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-077"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-04-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-012"], "Research_search": " {\"key\": \"halliburton-chevron-closed-loop\", \"existing_ids\": [\"AI-110\"], \"related_existing_ids\": [], \"company\": \"Halliburton\", \"title\": \"Closed-loop hydraulic fracturing with Chevron\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Automation / AI not established\", \"stage\": \"Closed-loop autonomous operation achieved in Colorado by Q1 2025; platform described Apr 2025\", \"practical_relevance\": \"A significant move from preset automation to feedback-driven closed-loop completions. Since the partner calls the capability intelligent fracturing but does not establish AI, do not classify as AI.\", \"business_problem\": \"Adapt frac delivery to local subsurface feedback while treatment is underway.\", \"workflow\": \"ZEUS IQ couples automated OCTIV stage execution with Sensori fracture feedback; Chevron’s subsurface knowledge is embedded in an algorithm to adapt operations in real time.\", \"technology_data\": \"Wellsite-to-cloud loop, subsurface measurements and automated stage controls; AI not expressly claimed in source, so classify as digital feedback automation.\", \"integration_human_role\": \"No-human-intervention operation is reported; human design/oversight details and safety stop behavior are not described.\", \"deployment_scope\": \"Chevron Colorado completions; reported operation count not disclosed.\", \"partners\": [\"Chevron\"], \"disclosed_results\": [], \"economics\": \"No quantified productivity or recovery metric disclosed.\", \"timeline\": [{\"date\": \"2025-04-22\", \"event\": \"Halliburton says first closed-loop autonomous frac operation achieved during Q1 2025\", \"source_ids\": [\"RS049\"]}, {\"date\": \"2025-04-12\", \"event\": \"Chevron/Halliburton workflow described publicly\", \"source_ids\": [\"RS049\"]}], \"limitations\": [\"Deployment evidence comes from Halliburton/partner release; no measured performance result. Automation, not verified AI.\", \"The reviewed Halliburton source describes feedback-driven automation and a domain algorithm, but does not identify an AI/ML method.\"], \"assessment\": \"Analyst inference: Feedback-driven completions are a meaningful control capability; on reviewed evidence this is closed-loop automation rather than established AI.\", \"sources\": [{\"id\": \"RS049\", \"title\": \"Chevron and Halliburton Enable Intelligent Hydraulic Fracturing\", \"url\": \"https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing\", \"publisher\": \"Halliburton\", \"date\": \"2025-04-12\", \"type\": \"company announcement\", \"locator\": \"Closed-loop intelligent fracturing description\", \"review_status\": \"full text reviewed\", \"supports\": [\"ZEUS IQ combines OCTIV Auto Frac execution, Sensori fracture monitoring and Chevron subsurface knowledge embedded in an algorithm.\", \"Closed-loop system uses real-time feedback to adapt delivery to local formation conditions.\"], \"original_id\": \"RS049\"}], \"field_sources\": {\"business_problem\": [\"RS049\"], \"workflow\": [\"RS049\"], \"technology_data\": [\"RS049\"], \"integration_human_role\": [\"RS049\"], \"deployment_scope\": [\"RS049\"], \"economics\": [\"RS049\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-012\", \"source_ids\": [\"RS049\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-110\"]}", "Research_status": "Expanded source review", "Research_category": "Automation / AI not established", "Research_verification": "verified partner", "Audit_flags": [{"claim": "Chevron intelligent fracturing is AI", "reason": "Partner release confirms closed-loop, feedback-driven automation but does not establish that the system uses AI. Categorize as automation unless additional evidence confirms AI.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"}], "Audit_review": {"id": "AI-110", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton’s release describes closed-loop, feedback-driven autonomous completion control and a domain algorithm; no source checked calls it AI or machine learning.", "correction": "Classify as closed-loop automation, not AI, unless other explicit AI evidence is found.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"}], "residual_gaps": ["Chevron’s own announcement URL returned 403; Halliburton primary source was reviewed and does not identify an AI/ML method."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-111", "Company": "Halliburton / Sekal", "Company_primary": "Halliburton", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Halliburton acquisition of Sekal drilling automation", "Extracted_finding": "Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.", "Context": "Expand drilling automation portfolio and combine directional drilling, wellbore hydraulics and rig equipment control.", "Implementation": "Acquisition combines LOGIX remote operations and automation with Sekal DrillTronics autonomy platform and services. The companies had earlier deployed an automated on-bottom system for Equinor on the Norwegian Continental Shelf.", "Assessment": "Analyst inference: Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Acquired April 1, 2026; LOGIX/Sekal product integration announced", "Interpretation_limit": "Acquisition terms not disclosed. Do not imply announced acquisition itself proves full portfolio rollout; Equinor deployment predates acquisition.", "Source_IDs": "RS073; RS074; R078", "Source_date": "2026-04-01", "Source_type": "company announcement", "Source_locator": "Acquisition and portfolio combination", "Source_URL": "https://ir.halliburton.com/node/36266/pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-04-01; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-078", "Initiative_ID": "AI-111", "Title": "Halliburton acquires Sekal drilling-automation software", "Date": "2026-04", "Finding": "Halliburton bought Sekal from Sumitomo Corporation, combining Sekal's DrillTronics with its LOGIX automation and remote operations. In 2025 the two had delivered the first automated on-bottom drilling system, for Equinor on the Norwegian Continental Shelf.", "Metric": "DrillTronics used on 1,300+ wells; up to 25% reduction in well delivery time (reported)", "Source_IDs": "R078", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Oilfield services companies are buying automation IP they first proved through partnerships, so a joint pilot with a major service company is a common route to acquisition.", "Partners": "Sekal DrillTronics, LOGIX, Equinor, Sumitomo", "Original_date": "2026-04", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-078"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-04-01", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-035"], "Research_search": " {\"key\": \"hall-sekal-acquisition\", \"existing_ids\": [\"AI-111\"], \"related_existing_ids\": [\"AI-016\"], \"company\": \"Halliburton\", \"title\": \"Halliburton acquisition of Sekal drilling automation\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Acquired April 1, 2026; LOGIX/Sekal product integration announced\", \"practical_relevance\": \"Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.\", \"business_problem\": \"Expand drilling automation portfolio and combine directional drilling, wellbore hydraulics and rig equipment control.\", \"workflow\": \"Acquisition combines LOGIX remote operations and automation with Sekal DrillTronics autonomy platform and services. The companies had earlier deployed an automated on-bottom system for Equinor on the Norwegian Continental Shelf.\", \"technology_data\": \"LOGIX, DrillTronics and rig automation control system; model/control details described in pre-acquisition Equinor pilot source.\", \"integration_human_role\": \"System optimized within safe limits and enabled single-button parameter optimization in pilot; precise human override/control boundary is not fully specified.\", \"deployment_scope\": \"Acquisition completed; workflow shown in 2025 Equinor well deployment.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Terms not disclosed in reviewed announcement.\", \"timeline\": [{\"date\": \"2025-02-26\", \"event\": \"Halliburton and Sekal reported integrated automated on-bottom drilling system for Equinor.\", \"source_ids\": [\"RS074\"]}, {\"date\": \"2026-04-01\", \"event\": \"Halliburton completed acquisition of Sekal from Sumitomo.\", \"source_ids\": [\"RS073\"]}], \"limitations\": [\"Acquisition terms not disclosed.\", \"Do not imply announced acquisition itself proves full portfolio rollout; Equinor deployment predates acquisition.\"], \"assessment\": \"Analyst inference: Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.\", \"sources\": [{\"id\": \"RS073\", \"title\": \"Halliburton Acquires Sekal AS From Sumitomo Corporation to Accelerate Drilling Automation Services\", \"url\": \"https://ir.halliburton.com/node/36266/pdf\", \"publisher\": \"Halliburton\", \"date\": \"2026-04-01\", \"type\": \"company announcement\", \"locator\": \"Acquisition and portfolio combination\", \"review_status\": \"full text reviewed\", \"supports\": [\"Halliburton acquired Sekal from Sumitomo and said it will combine LOGIX automation/remote operations with DrillTronics automation.\"], \"original_id\": \"hall_sekal\"}, {\"id\": \"RS074\", \"title\": \"Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea\", \"url\": \"https://ir.halliburton.com/node/35181/pdf\", \"publisher\": \"Halliburton\", \"date\": \"2025-02-26\", \"type\": \"partner case study\", \"locator\": \"Integrated system and Equinor deployment\", \"review_status\": \"full text reviewed\", \"supports\": [\"Halliburton, Sekal and rig automation delivered an integrated closed-loop automated on-bottom drilling system for an Equinor well in Norway.\"], \"original_id\": \"hall_equinor\"}], \"field_sources\": {\"business_problem\": [\"RS073\", \"RS074\"], \"workflow\": [\"RS073\", \"RS074\"], \"technology_data\": [\"RS073\", \"RS074\"], \"integration_human_role\": [\"RS073\", \"RS074\"], \"deployment_scope\": [\"RS073\", \"RS074\"], \"economics\": [\"RS073\", \"RS074\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Acquisition completed on 2026-04-01 per Halliburton primary source; the Equinor automated drilling deployment occurred in 2025 before the acquisition.\", \"cohort\": \"services\", \"packet_id\": \"services-035\", \"source_ids\": [\"RS073\", \"RS074\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-111\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient.", "source_url": ""}], "Audit_review": {"id": "AI-111", "company": "Halliburton / Sekal", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/node/35181/pdf", "https://ir.halliburton.com/node/36266/pdf"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "Sekal acquisition source supports drilling-automation capability, not necessarily an AI model.", "correction": "Classify as drilling automation; do not infer AI from autonomy.", "source_url": "https://ir.halliburton.com/node/36266/pdf"}], "residual_gaps": ["The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-112", "Company": "Baker Hughes", "Company_primary": "Baker Hughes", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Leucipa GenAI assistant and lift optimizer for Repsol", "Extracted_finding": "A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.", "Context": "Improve access to production insights and reduce manual surveillance burden on plunger-lift well engineers.", "Implementation": "Leucipa virtual assistant offers conversational data access and real-time/predictive insight. Repsol was piloting Lift Optimizer for plunger-lift wells to help monitor, control and maintain wells.", "Assessment": "Analyst inference: A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Agreement and early pilots announced June 2025", "Interpretation_limit": "Company announcement describes early-stage deployment/pilot as of June 2025.", "Source_IDs": "R081", "Source_date": "2025-06-10", "Source_type": "company announcement", "Source_locator": "Leucipa assistant and lift optimizer", "Source_URL": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-06-10; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-081", "Initiative_ID": "AI-112", "Title": "Leucipa adds genAI assistant, AI lift optimizer for Repsol", "Date": "2025-06", "Finding": "Baker Hughes and Repsol agreed to deploy new Leucipa capabilities, including a generative AI virtual assistant. Repsol is piloting the AI-powered Leucipa Lift Optimizer on plunger-lift wells to reduce the time production engineers spend on individual wells, following an October 2024 MOU on AI field automation.", "Metric": "No quantified metric supplied", "Source_IDs": "R081", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Production-engineer time on lift optimization is a proven, fundable use case that majors now co-develop with oilfield services companies.", "Partners": "Repsol, Leucipa", "Original_date": "2025-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-081"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-06-10", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-014"], "Research_search": " {\"key\": \"baker-repsol-leucipa\", \"existing_ids\": [\"AI-112\"], \"related_existing_ids\": [], \"company\": \"Baker Hughes\", \"title\": \"Leucipa GenAI assistant and lift optimizer for Repsol\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Agreement and early pilots announced June 2025\", \"practical_relevance\": \"A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.\", \"business_problem\": \"Improve access to production insights and reduce manual surveillance burden on plunger-lift well engineers.\", \"workflow\": \"Leucipa virtual assistant offers conversational data access and real-time/predictive insight. Repsol was piloting Lift Optimizer for plunger-lift wells to help monitor, control and maintain wells.\", \"technology_data\": \"GenAI chat capability and production/lift optimization analytics within automated field production solution; models and source systems not described.\", \"integration_human_role\": \"Supports production engineers; optimizer pilot not stated to autonomously control wells.\", \"deployment_scope\": \"Repsol plunger-lift wells; number/field footprint not disclosed.\", \"partners\": [\"Repsol\"], \"disclosed_results\": [], \"economics\": \"No measured outcomes, return or deployment scale disclosed.\", \"timeline\": [{\"date\": \"2025-06-10\", \"event\": \"Leucipa GenAI and Lift Optimizer announced; Repsol early-stage pilot\", \"source_ids\": [\"R081\"]}], \"limitations\": [\"Company announcement describes early-stage deployment/pilot as of June 2025.\"], \"assessment\": \"Analyst inference: A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.\", \"sources\": [{\"id\": \"R081\", \"title\": \"Baker Hughes, Repsol Launching New AI-Powered Functionality in Leucipa\", \"url\": \"https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm\", \"publisher\": \"Baker Hughes\", \"date\": \"2025-06-10\", \"type\": \"company announcement\", \"locator\": \"Leucipa assistant and lift optimizer\", \"review_status\": \"full text reviewed\", \"supports\": [\"Agreement adds GenAI virtual assistant for conversational access to data and real-time insights and predictive analytics.\", \"Repsol was piloting Leucipa Lift Optimizer for plunger-lift wells to reduce time engineers spend monitoring, controlling and maintaining individual wells.\"], \"original_id\": \"baker_repsol\"}], \"field_sources\": {\"business_problem\": [\"R081\"], \"workflow\": [\"R081\"], \"technology_data\": [\"R081\"], \"integration_human_role\": [\"R081\"], \"deployment_scope\": [\"R081\"], \"economics\": [\"R081\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-014\", \"source_ids\": [\"R081\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-112\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-112", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-113", "Company": "Weatherford", "Company_primary": "Weatherford", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Industrial Intelligence portfolio for AI and edge operations", "Extracted_finding": "The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.", "Context": "Connect heterogeneous industrial data and deliver analytics/control and AI tools within customers’ existing operations systems.", "Implementation": "Portfolio includes AI/physics-based MetaFlow virtual flow metering, Engineer’s Briefcase document/transcription/image/language tools, and StrataCore production model, with unified data and edge/software layers.", "Assessment": "Analyst inference: The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Portfolio launched Oct 2025", "Interpretation_limit": "Vendor portfolio description should not be mistaken for rollout. Separate internal AI agents reported in sustainability report from customer product capability.", "Source_IDs": "RS051; RS052; R082", "Source_date": "2026-04-01", "Source_type": "company announcement", "Source_locator": "Product portfolio descriptions", "Source_URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-04-01; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-082", "Initiative_ID": "AI-113", "Title": "Weatherford Industrial Intelligence portfolio launched", "Date": "2025-10", "Finding": "At FWRD 2025, Weatherford combined ForeSite, Centro, PetroVisor and new tools into 'Industrial Intelligence'. New pieces include the StrataCore production foundation model, the MetaFlow AI virtual flow meter, Canvas no-code data science, and the WFRD Software Launchpad SaaS hub.", "Metric": "Operates in ~75 countries", "Source_IDs": "R082", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Even mid-sized oilfield services companies now pitch production foundation models and no-code tools, so 'foundation model for production' is no longer a differentiator.", "Partners": "ForeSite, Centro, PetroVisor, StrataCore, Canvas", "Original_date": "2025-10", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-082"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-04-01", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-015"], "Research_search": " {\"key\": \"weatherford-intelligence\", \"existing_ids\": [\"AI-113\"], \"related_existing_ids\": [], \"company\": \"Weatherford\", \"title\": \"Industrial Intelligence portfolio for AI and edge operations\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Portfolio launched Oct 2025\", \"practical_relevance\": \"The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.\", \"business_problem\": \"Connect heterogeneous industrial data and deliver analytics/control and AI tools within customers’ existing operations systems.\", \"workflow\": \"Portfolio includes AI/physics-based MetaFlow virtual flow metering, Engineer’s Briefcase document/transcription/image/language tools, and StrataCore production model, with unified data and edge/software layers.\", \"technology_data\": \"AI and physics models, edge systems, unified Synsera data model and software Launchpad; interoperable with customer systems/hardware.\", \"integration_human_role\": \"Domain engineers, field/surveillance engineers and executives are named user groups; specific autonomy/human controls vary by tool.\", \"deployment_scope\": \"Portfolio breadth advertised; specific customer deployments or performance evidence not disclosed in launch release.\", \"partners\": [\"Weatherford software partners and customers\"], \"disclosed_results\": [], \"economics\": \"No consolidated revenue or customer productivity metric disclosed.\", \"timeline\": [{\"date\": \"2025-10-01\", \"event\": \"Industrial Intelligence portfolio launched\", \"source_ids\": [\"RS051\"]}, {\"date\": \"2025\", \"event\": \"Internal agents embedded in company workflows, per 2025 sustainability report\", \"source_ids\": [\"RS052\"]}], \"limitations\": [\"Vendor portfolio description should not be mistaken for rollout. Separate internal AI agents reported in sustainability report from customer product capability.\"], \"assessment\": \"Analyst inference: The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.\", \"sources\": [{\"id\": \"RS051\", \"title\": \"Weatherford Launches Industrial Intelligence at FWRD 2025\", \"url\": \"https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126\", \"publisher\": \"Weatherford\", \"date\": \"2025-10-01\", \"type\": \"company announcement\", \"locator\": \"Product portfolio descriptions\", \"review_status\": \"full text reviewed\", \"supports\": [\"Portfolio includes MetaFlow AI/physics-based virtual flow meter, Engineer’s Briefcase AI document transcription/recognition/translation tools and StrataCore production foundational model.\", \"It also includes unified data model, edge and SaaS software; portfolio launch does not itself verify customer deployment.\"], \"original_id\": \"weatherford_2025\"}, {\"id\": \"RS052\", \"title\": \"Weatherford 2025 Sustainability Report: Artificial Intelligence\", \"url\": \"https://www.weatherford.com/documents/sustainability/sustainability-report/\", \"publisher\": \"Weatherford\", \"date\": \"2026-04-01\", \"type\": \"sustainability report\", \"locator\": \"Artificial Intelligence section\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Weatherford reports an AI Impact Hub and agents embedded in internal workflows in 2025.\", \"Named agents review material test reports, cross-reference parts/inventory, classify spend and simplify bidding; training and human involvement stated.\"], \"original_id\": \"weatherford_internal\"}], \"field_sources\": {\"business_problem\": [\"RS051\", \"RS052\"], \"workflow\": [\"RS051\", \"RS052\"], \"technology_data\": [\"RS051\", \"RS052\"], \"integration_human_role\": [\"RS051\", \"RS052\"], \"deployment_scope\": [\"RS051\", \"RS052\"], \"economics\": [\"RS051\", \"RS052\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-015\", \"source_ids\": [\"RS051\", \"RS052\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-113\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "At least one supporting long document was reviewed at relevant excerpts only.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-113", "company": "Weatherford", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "https://www.weatherford.com/documents/sustainability/sustainability-report/"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The user-supplied GlobeNewswire link failed, but Weatherford's issuer-hosted copy of the October 1, 2025 release was reviewed and supports the portfolio-launch details.", "correction": "Retain launch facts with the Weatherford investor-relations copy as primary evidence; describe components as offered capabilities, not customer deployment or outcomes.", "source_url": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-114", "Company": "NOV", "Company_primary": "NOV", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "AIQ collaboration for edge-enabled drilling AI pilots", "Extracted_finding": "Strategic agreement is an AI roadmap and test program, not a field deployment yet.", "Context": "Develop AI workflows for live drilling operations with secure access to equipment and operational data.", "Implementation": "NOV and AIQ plan targeted PoCs and pilots; Max edge platform captures data, processes it at edge and integrates AI workflows.", "Assessment": "Analyst inference: Strategic agreement is an AI roadmap and test program, not a field deployment yet.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Collaboration announced Jan 2026; pilot planning stage", "Interpretation_limit": "Announcement specifies future PoC/pilot, not completed deployment.", "Source_IDs": "R083", "Source_date": "2026-01-12", "Source_type": "company announcement", "Source_locator": "Collaboration scope", "Source_URL": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-01-12; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-083", "Initiative_ID": "AI-114", "Title": "NOV and AIQ pilot AI drilling on Max platform", "Date": "2026-01", "Finding": "NOV and AIQ agreed to combine NOV's Max industrial IoT and edge-computing platform with AIQ's AI to design and test AI-driven drilling optimization. Work will run through proof-of-concept and pilot deployments in live operations.", "Metric": "No quantified metric supplied", "Source_IDs": "R083", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Equipment makers with edge platforms are looking for AI partners, so an AI builder can reach the rig without owning hardware by plugging into their edge layer.", "Partners": "AIQ, NOV Max", "Original_date": "2026-01", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-083"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-01-12", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-017"], "Research_search": " {\"key\": \"nov-aiq-pilot\", \"existing_ids\": [\"AI-114\"], \"related_existing_ids\": [], \"company\": \"NOV\", \"title\": \"AIQ collaboration for edge-enabled drilling AI pilots\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Collaboration announced Jan 2026; pilot planning stage\", \"practical_relevance\": \"Strategic agreement is an AI roadmap and test program, not a field deployment yet.\", \"business_problem\": \"Develop AI workflows for live drilling operations with secure access to equipment and operational data.\", \"workflow\": \"NOV and AIQ plan targeted PoCs and pilots; Max edge platform captures data, processes it at edge and integrates AI workflows.\", \"technology_data\": \"NOV drilling equipment/Max IIoT with AIQ industrial AI; no models or use-case performance specified.\", \"integration_human_role\": \"Future real-time operational decision support; details on crew interaction not disclosed.\", \"deployment_scope\": \"Targeted live-setting pilots, no customer or rig named.\", \"partners\": [\"AIQ\"], \"disclosed_results\": [], \"economics\": \"No contract or economic terms disclosed.\", \"timeline\": [{\"date\": \"2026-01-12\", \"event\": \"NOV-AIQ collaboration announced\", \"source_ids\": [\"R083\"]}], \"limitations\": [\"Announcement specifies future PoC/pilot, not completed deployment.\"], \"assessment\": \"Analyst inference: Strategic agreement is an AI roadmap and test program, not a field deployment yet.\", \"sources\": [{\"id\": \"R083\", \"title\": \"NOV and AIQ to Advance AI-Enabled Drilling Optimization\", \"url\": \"https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization\", \"publisher\": \"NOV\", \"date\": \"2026-01-12\", \"type\": \"company announcement\", \"locator\": \"Collaboration scope\", \"review_status\": \"full text reviewed\", \"supports\": [\"NOV and AIQ plan proof-of-concept/pilot AI drilling solutions in live settings.\", \"NOV Max platform supports secure data capture, edge processing and workflow integration; announcement is plan, not completed deployment.\"], \"original_id\": \"nov_aiq\"}], \"field_sources\": {\"business_problem\": [\"R083\"], \"workflow\": [\"R083\"], \"technology_data\": [\"R083\"], \"integration_human_role\": [\"R083\"], \"deployment_scope\": [\"R083\"], \"economics\": [\"R083\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-017\", \"source_ids\": [\"R083\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-114\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-114", "company": "NOV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization"], "findings": [{"severity": "medium", "field": "stage", "issue": "NOV/AIQ source describes planned or pilot edge-AI drilling collaboration; no scaled commercial deployment or results shown.", "correction": "Retain pilot stage and avoid productized/deployed language.", "source_url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-115", "Company": "Nabors / Corva", "Company_primary": "Nabors", "Segment": "Oilfield services", "HQ": "Bermuda / United States", "Theme": "Operational AI", "Initiative": "RigCLOUD Powered by Corva drilling intelligence", "Extracted_finding": "Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.", "Context": "Give drilling crews and remote centers real-time analytics and predictive insight across connected rig fleets.", "Implementation": "RigCLOUD’s edge/cloud infrastructure connects Corva AI applications; optional SmartROS integration adds rig automation and real-time rigsite applications.", "Assessment": "Analyst inference: Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Expanded strategic alliance announced Apr 2025", "Interpretation_limit": "Expanded partnership in window; cited product performance page does not date its case example, and Nabors first disclosed metrics in 2023.", "Source_IDs": "RS054; RS055; R084", "Source_date": "2025-04-10", "Source_type": "company announcement", "Source_locator": "RigCLOUD Powered by Corva scope", "Source_URL": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-04-10; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-084", "Initiative_ID": "AI-115", "Title": "RigCLOUD Powered by Corva drilling intelligence", "Date": "2025-04", "Finding": "Nabors expanded its alliance with Corva by integrating Corva's AI analytics into its RigCLOUD edge and cloud platform, sold as 'RigCLOUD Powered by Corva'. Corva also offers Corva Copilot, an AI assistant for drilling and completions.", "Metric": "No quantified metric supplied", "Source_IDs": "R084", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Drilling contractors white-label startup analytics rather than building their own, which gives startups distribution across a whole rig fleet.", "Partners": "Corva, Corva Copilot, RigCLOUD", "Original_date": "2025-04", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-084"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-04-10", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-018"], "Research_search": " {\"key\": \"nabors-corva-predictive\", \"existing_ids\": [\"AI-115\"], \"related_existing_ids\": [], \"company\": \"Nabors\", \"title\": \"RigCLOUD Powered by Corva drilling intelligence\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Expanded strategic alliance announced Apr 2025\", \"practical_relevance\": \"Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.\", \"business_problem\": \"Give drilling crews and remote centers real-time analytics and predictive insight across connected rig fleets.\", \"workflow\": \"RigCLOUD’s edge/cloud infrastructure connects Corva AI applications; optional SmartROS integration adds rig automation and real-time rigsite applications.\", \"technology_data\": \"Corva AI analytics and apps, edge-cloud data, SmartROS automation optional.\", \"integration_human_role\": \"Drillers/site managers/directional drillers access insights at rig and remote operations; automation depends on rig controls and user setup.\", \"deployment_scope\": \"Marketed across Nabors and third-party contractor rigs; no new customer rollout named in alliance expansion.\", \"partners\": [\"Corva AI\"], \"disclosed_results\": [], \"economics\": \"No new 2025 deployment economics disclosed. Older 2023 Predictive Drilling case reported 36% ROP gain and 9.7% vibration reduction; those results are outside window and are not treated as current-window deployment.\", \"timeline\": [{\"date\": \"2025-04-10\", \"event\": \"RigCLOUD-Corva collaboration expanded\", \"source_ids\": [\"RS054\"]}], \"limitations\": [\"Expanded partnership in window; cited product performance page does not date its case example, and Nabors first disclosed metrics in 2023.\"], \"assessment\": \"Analyst inference: Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.\", \"sources\": [{\"id\": \"RS054\", \"title\": \"Nabors and Corva Expand Strategic Alliance to Accelerate Digital Innovation\", \"url\": \"https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/\", \"publisher\": \"Nabors\", \"date\": \"2025-04-10\", \"type\": \"company announcement\", \"locator\": \"RigCLOUD Powered by Corva scope\", \"review_status\": \"full text reviewed\", \"supports\": [\"Expanded alliance integrates RigCLOUD edge/cloud platform with Corva AI analytics and apps for real-time drilling data, predictive insights and optimization.\", \"Operators and contractors can use the tools at rigsite and remotely; SmartROS controls are optional. Announcement describes solution capability, not specific named customer rollout.\"], \"original_id\": \"nabors_corva\"}, {\"id\": \"RS055\", \"title\": \"Predictive Drilling - Nabors\", \"url\": \"https://www.nabors.com/drilling-automation/predictive-drilling/\", \"publisher\": \"Nabors\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company product documentation\", \"locator\": \"Success in Action and availability\", \"review_status\": \"full text reviewed\", \"supports\": [\"Nabors reports the Predictive Drilling ROP optimizer combines AI/ML with SmartROS/RigCLOUD for remote autodriller setpoint control.\", \"First deployment reported 36% higher average ROP, 9.7% lower vibration and 80% lateral footage utilization; historic case source (2023) is outside window.\"], \"original_id\": \"nabors_trial\"}], \"field_sources\": {\"business_problem\": [\"RS054\", \"RS055\"], \"workflow\": [\"RS054\", \"RS055\"], \"technology_data\": [\"RS054\", \"RS055\"], \"integration_human_role\": [\"RS054\", \"RS055\"], \"deployment_scope\": [\"RS054\", \"RS055\"], \"economics\": [\"RS054\", \"RS055\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-018\", \"source_ids\": [\"RS054\", \"RS055\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-115\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Predictive Drilling 36% ROP improvement and 9.7% vibration reduction are current-window results", "reason": "Company product page reports these results, but Nabors primary case page dates first deployment and metrics to 2023; keep them as historical baseline only.", "source_url": "https://www.nabors.com/predictive-drilling-results-delaware/"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-115", "company": "Nabors / Corva", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "https://www.nabors.com/drilling-automation/predictive-drilling/"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The Apr 11, 2025 article describes intended integration of Nabors RigCLOUD edge/cloud/rig instrumentation with Corva AI analytics. It does not document deployed customer count, basin/rig scope, or realized gains; retain as partner/vendor offering/rollout stage.", "correction": "", "source_url": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-116", "Company": "Helmerich & Payne", "Company_primary": "Helmerich & Payne", "Segment": "Oilfield services", "HQ": "United States", "Theme": "AI products", "Initiative": "FlexRobotics rig-floor automation with an AI-enabled monitoring add-on", "Extracted_finding": "H&P reports a customer deployment of FlexRobotics rig-floor automation. H&P separately offers AI-enabled zone monitoring as an add-on camera feature; the reviewed sources do not confirm that the deployed customer rig used that AI add-on.", "Context": "H&P said it deployed its FlexRobotics rig-floor automation system on a rig for a supermajor in the Permian Basin. The system combines robotic arms, automation software and safety controls to keep people out of high-risk zones during drilling and tripping.", "Implementation": "Three robotic arms automate pipe tailing, cleaning/dope/alignment and racking. The separate product page lists AI-enabled zone monitoring as an add-on camera capability; deployment of that feature on the reported customer rig is not established.", "Assessment": "The deployment evidence is for industrial robotics/automation. The AI claim belongs to an ancillary commercial camera add-on and must remain separate from verified customer deployment.", "Reported_metric": "No quantified customer result disclosed. Product page lists 1:40 slip-to-slip connection time as a product capability, not a verified result for the reported deployment.", "Stage_at_disclosure": "Base FlexRobotics automation deployed on a Permian rig; AI-enabled zone-monitoring add-on is described in product documentation, but customer deployment is unconfirmed.", "Interpretation_limit": "Do not attribute AI to the deployed pipe-handling workflow or imply the AI camera feature was installed at the reported customer rig. The documented customer deployment is automation; AI is an available ancillary product capability.", "Source_IDs": "RS056; RS057; R085", "Source_date": "2026-02-04", "Source_type": "earnings release", "Source_locator": "FlexRobotics deployment", "Source_URL": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Dated source: 2026-02-04; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-085", "Initiative_ID": "AI-116", "Title": "FlexRobotics deployment and separate AI-enabled monitoring add-on", "Date": "2026-02-04", "Finding": "H&P reported FlexRobotics rig-floor automation deployed on a Permian rig for a supermajor across multiple pads. Separately, H&P product documentation lists AI-enabled zone monitoring as an add-on camera feature; it does not establish that this add-on was used on the reported customer deployment.", "Metric": "No quantified customer result disclosed; 1:40 connection time is a product-page capability statement, not deployment-specific", "Source_IDs": "RS056; RS057; R085", "Evidence": "Primary issuer earnings release and product documentation reviewed; AI add-on deployment unconfirmed", "Origin": "Issuer earnings release and H&P product page", "Partners": "NOV; Yaskawa", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-085"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-019"], "Research_search": " {\"key\": \"hp-flexrobotics\", \"existing_ids\": [\"AI-116\"], \"related_existing_ids\": [], \"company\": \"Helmerich & Payne\", \"title\": \"FlexRobotics rig-floor automation; AI-enabled monitoring add-on deployment unconfirmed\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Base automation deployed on a Permian rig; AI-enabled zone-monitoring add-on available in product documentation, customer deployment not confirmed.\", \"practical_relevance\": \"A concrete rig-floor robotics deployment reduces exposure to tubular handling work; AI evidence is limited to an optional monitoring capability whose use at the deployed rig is unconfirmed.\", \"business_problem\": \"Reduce personnel exposure and standardize tubular handling and connection work on rig floor.\", \"workflow\": \"Three robotic arms automate pipe tailing, cleaning/dope/alignment and racking. H&P separately lists AI-enabled zone monitoring as an add-on camera feature; it does not link that feature to the reported customer deployment.\", \"technology_data\": \"FlexRobotics uses robotic arms and automation software; product documentation describes AI-enabled zone monitoring on an add-on camera. AI methods, sensor data and use at the customer site are not disclosed.\", \"integration_human_role\": \"Automation performs pipe handling to keep personnel away from rig-floor red zones. The sources do not describe operator approvals/interventions or AI alert review; use of the add-on at the reported rig is unconfirmed.\", \"deployment_scope\": \"H&P reports one Permian rig deployment for an unnamed supermajor across multiple pads. The product page does not establish AI-add-on deployment at that site.\", \"partners\": [\"NOV\", \"Yaskawa\"], \"disclosed_results\": [], \"economics\": \"No quantified customer result disclosed. Product-page connection-time statement is a product capability, not a verified customer deployment outcome.\", \"timeline\": [{\"date\": \"2026-02-04\", \"event\": \"H&P reports first FlexRobotics system deployment for a supermajor\", \"source_ids\": [\"RS056\"]}, {\"date\": \"2026-05-06\", \"event\": \"H&P earnings say deployment is expanding\", \"source_ids\": [\"RS056\"]}], \"limitations\": [\"Earnings source confirms base robotics/automation deployment, not AI use in the deployed robot.\", \"AI-enabled zone monitoring is an ancillary add-on in product documentation; customer deployment of that feature is not established.\", \"Operator identity and deployment-specific performance/economics are not disclosed.\"], \"assessment\": \"Analyst inference: include the add-on as commercial AI product capability, while classifying the proven field deployment as robotics/automation. The available sources do not establish that the AI add-on was installed at the customer rig.\", \"sources\": [{\"id\": \"RS056\", \"title\": \"Helmerich & Payne, Inc. Announces Fiscal First Quarter Results\", \"url\": \"https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx\", \"publisher\": \"H&P\", \"date\": \"2026-02-04\", \"type\": \"earnings release\", \"locator\": \"FlexRobotics deployment\", \"review_status\": \"full text reviewed\", \"supports\": [\"H&P reports FlexRobotics deployed on a Permian rig for a supermajor across multiple pads; growing customer interest.\", \"Rig-floor system is collaborative robotics for tubular handling. H&P product page says AI-enabled zone monitoring is an ancillary feature; core handling automation is not AI.\"], \"original_id\": \"hp_q1\"}, {\"id\": \"RS057\", \"title\": \"FlexRobotics Hands-Free Rig Floor\", \"url\": \"https://www.hpinc.com/technologies/flexrobotics\", \"publisher\": \"H&P\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company product documentation\", \"locator\": \"Description, operation and AI-enabled zone monitoring\", \"review_status\": \"full text reviewed\", \"supports\": [\"Three robot arms handle pipe tailing, cleaning/alignment and racking. H&P lists 1:40 slip-to-slip connection times and red-zone personnel exposure reduction.\", \"System was developed with NOV and uses Yaskawa robot manipulators. Cameras support AI-enabled zone monitoring.\"], \"accessed_date\": \"2026-09-27\", \"original_id\": \"hp_product\"}], \"field_sources\": {\"business_problem\": [\"RS056\", \"RS057\"], \"workflow\": [\"RS056\", \"RS057\"], \"technology_data\": [\"RS056\", \"RS057\"], \"integration_human_role\": [\"RS056\", \"RS057\"], \"deployment_scope\": [\"RS056\", \"RS057\"], \"economics\": [\"RS056\", \"RS057\"]}, \"verification\": \"verified primary for base automation deployment; AI add-on is product capability only\", \"corrections_to_existing\": \"Separate customer deployment of base robotics/automation from the separately listed AI-enabled camera add-on; deployment of the add-on is unconfirmed.\", \"record_ids\": [\"AI-116\"], \"cohort\": \"services\", \"packet_id\": \"services-019\", \"source_ids\": [\"RS056\", \"RS057\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\"}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary for base automation deployment; AI add-on is product capability only", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-116", "company": "Helmerich & Payne", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "https://www.hpinc.com/technologies/flexrobotics"], "findings": [{"severity": "medium", "field": "classification / scope", "issue": "The earnings release confirms a FlexRobotics rig-floor deployment, but does not say AI was deployed with it. The separate H&P product page calls the base product automation and lists AI-enabled zone monitoring as an add-on camera feature.", "correction": "Classify the base workflow as robotics/automation; use Commercial AI product for the add-on capability only and state that the customer deployment of the AI add-on is unconfirmed.", "source_url": "https://www.hpinc.com/technologies/flexrobotics"}, {"severity": "low", "field": "Supplemental source review", "issue": "Issuer earnings release confirms FlexRobotics deployment on a Permian rig for a supermajor, but does not establish AI in the deployed robot. H&P’s separate product page calls FlexRobotics automation and lists AI zone monitoring as an add-on camera feature.", "correction": "Keep AI only for the explicitly described camera-based monitoring capability and qualify it as add-on/product capability; do not imply it was part of the customer deployment absent evidence. The deployed base robotics workflow is automation.", "source_url": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported; deployment is automation, AI add-on deployment unconfirmed"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary for base automation deployment; AI add-on is product capability only", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-117", "Company": "Patterson-UTI", "Company_primary": "Patterson-UTI", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Cortex machine-learning autodriller and REX rig alerts", "Extracted_finding": "This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.", "Context": "Improve repeatability of rig drilling performance and detect operating issues early.", "Implementation": "Cortex advanced ML autodriller applies machine learning to drilling parameters; REX cloud early-alert monitoring gives field operations insight. By year-end, nearly all rigs had Cortex automation applications.", "Assessment": "Analyst inference: This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Nearly all active rigs equipped with Cortex apps by Q4 2025", "Interpretation_limit": "Company’s broad “automation” rig penetration is not AI-specific; machine-learning claim applies to Cortex autodriller.", "Source_IDs": "R086; RS058", "Source_date": "2026-02-04", "Source_type": "earnings release", "Source_locator": "Drilling services remarks", "Source_URL": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-02-04; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-086", "Initiative_ID": "AI-117", "Title": "Cortex ML autodriller and REX monitoring drive rig margins", "Date": "2025-07", "Finding": "Patterson-UTI credits its Cortex automation platform (machine-learning autodriller), REX cloud early-alert monitoring and its Digital Performance Center with differentiated rig performance. On completions it uses Vertex frac automation and FleetStream real-time data feeds.", "Metric": "No quantified metric supplied", "Source_IDs": "R086", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Contractors market their digital and ML stack as a margin driver, so AI vendors should frame their value in the contractor's terms, such as revenue per rig and performance pricing.", "Partners": "Cortex, REX, Vertex, FleetStream, EOS Completions Platform", "Original_date": "2025-07", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-086"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-02-04", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-020"], "Research_search": " {\"key\": \"patterson-cortex-rex\", \"existing_ids\": [\"AI-117\"], \"related_existing_ids\": [], \"company\": \"Patterson-UTI\", \"title\": \"Cortex machine-learning autodriller and REX rig alerts\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Nearly all active rigs equipped with Cortex apps by Q4 2025\", \"practical_relevance\": \"This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.\", \"business_problem\": \"Improve repeatability of rig drilling performance and detect operating issues early.\", \"workflow\": \"Cortex advanced ML autodriller applies machine learning to drilling parameters; REX cloud early-alert monitoring gives field operations insight. By year-end, nearly all rigs had Cortex automation applications.\", \"technology_data\": \"Machine-learning autodriller and cloud alerting platform; data features/model detail and exact rig count not stated.\", \"integration_human_role\": \"Drillers receive operating setpoints/alerts; source does not detail control mode, override or alert actioning.\", \"deployment_scope\": \"Patterson-UTI U.S. rig fleet; nearly all rigs for broader Cortex automation suite.\", \"partners\": [\"Internal teams\"], \"disclosed_results\": [], \"economics\": \"No AI-specific performance or economic result; Q2 2025 management cited growing customer demand/revenue for drilling automation products.\", \"timeline\": [{\"date\": \"2025-07-30\", \"event\": \"Q2 results detail Cortex ML autodriller and REX alerts\", \"source_ids\": [\"R086\"]}, {\"date\": \"2026-02-04\", \"event\": \"Q4 results report Cortex automation applications on nearly all rigs\", \"source_ids\": [\"RS058\"]}], \"limitations\": [\"Company’s broad “automation” rig penetration is not AI-specific; machine-learning claim applies to Cortex autodriller.\"], \"assessment\": \"Analyst inference: This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.\", \"sources\": [{\"id\": \"R086\", \"title\": \"Patterson-UTI Energy Reports Financial Results for the Quarter Ended June 30, 2025\", \"url\": \"https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm\", \"publisher\": \"Patterson-UTI\", \"date\": \"2025-07-30\", \"type\": \"earnings release\", \"locator\": \"Drilling services remarks\", \"review_status\": \"full text reviewed\", \"supports\": [\"Cortex platform enables a machine-learning autodriller; REX is a cloud-based early-alert field monitoring system.\", \"Company cites increased demand and revenue from drilling automation technologies, but no separate AI return metric.\"], \"original_id\": \"pten\"}, {\"id\": \"RS058\", \"title\": \"Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025\", \"url\": \"https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm\", \"publisher\": \"Patterson-UTI\", \"date\": \"2026-02-04\", \"type\": \"earnings release\", \"locator\": \"Drilling Services\", \"review_status\": \"full text reviewed\", \"supports\": [\"Patterson-UTI said nearly all rigs were equipped with proprietary Cortex automation applications by year-end 2025.\", \"Management cited continued development of automation software and demand for performance-enhancing products; this is scale of automation suite, not AI-specific model usage.\"], \"original_id\": \"pten_q4\"}], \"field_sources\": {\"business_problem\": [\"R086\", \"RS058\"], \"workflow\": [\"R086\", \"RS058\"], \"technology_data\": [\"R086\", \"RS058\"], \"integration_human_role\": [\"R086\", \"RS058\"], \"deployment_scope\": [\"R086\", \"RS058\"], \"economics\": [\"R086\", \"RS058\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-020\", \"source_ids\": [\"R086\", \"RS058\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-117\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-117", "company": "Patterson-UTI", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "scaled", "Scale_note": "nearly all active rigs", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-118", "Company": "Liberty Energy", "Company_primary": "Liberty Energy", "Segment": "Oilfield services", "HQ": "United States", "Theme": "Operational AI", "Initiative": "Forge LLM orchestration for StimCommander frac control", "Extracted_finding": "A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.", "Context": "Coordinate rate and pressure settings across frac assets and improve speed/efficiency of hydraulic stimulation.", "Implementation": "StimCommander uses AI-driven automated rate and pressure control; Forge LLM adds intelligent asset orchestration and continuous AI optimization.", "Assessment": "Analyst inference: A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.", "Reported_metric": "time to desired fluid injection rate: 65% improvement (Company-reported; StimCommander automation, not Forge alone); hydraulic efficiency: 5%-10% improvement (Company-reported; StimCommander automation, not Forge alone)", "Stage_at_disclosure": "Forge launched in Q3 2025; StimCommander use already operating", "Interpretation_limit": "Metrics are company-reported and are not attributed to Forge specifically; keep digiPrime pump savings separate.", "Source_IDs": "R087", "Source_date": "2025-10-16", "Source_type": "earnings release", "Source_locator": "Forge and StimCommander paragraphs", "Source_URL": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-10-16; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-087", "Initiative_ID": "AI-118", "Title": "Liberty launches Forge LLM for frac asset orchestration", "Date": "2025-10", "Finding": "Liberty launched Forge, its own large language model for intelligent asset orchestration, which works with StimCommander, its AI-driven automated rate and pressure control software. The company reported fleet-level automation gains in its Q3 2025 results.", "Metric": "StimCommander: 65% faster time to target injection rate; 5-10% hydraulic efficiency gain; digiPrime maintenance savings >30%", "Source_IDs": "R087", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Frac companies are building proprietary LLMs to orchestrate their own equipment fleets, which leaves little room for outside vendors in pressure pumping itself.", "Partners": "Forge, StimCommander, digiPrime", "Original_date": "2025-10", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-087"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-10-16", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["services-021"], "Research_search": " {\"key\": \"liberty-forge-stimcommander\", \"existing_ids\": [\"AI-118\"], \"related_existing_ids\": [], \"company\": \"Liberty Energy\", \"title\": \"Forge LLM orchestration for StimCommander frac control\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Forge launched in Q3 2025; StimCommander use already operating\", \"practical_relevance\": \"A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.\", \"business_problem\": \"Coordinate rate and pressure settings across frac assets and improve speed/efficiency of hydraulic stimulation.\", \"workflow\": \"StimCommander uses AI-driven automated rate and pressure control; Forge LLM adds intelligent asset orchestration and continuous AI optimization.\", \"technology_data\": \"Cloud LLM plus automated controls and frac fleet telemetry; detailed architecture and data scope not disclosed.\", \"integration_human_role\": \"System automates setpoint/time-to-rate control; role of frac engineers and safety constraints not described in release.\", \"deployment_scope\": \"Liberty frac fleets; specific customer/site count not disclosed.\", \"partners\": [\"Internal Liberty teams\"], \"disclosed_results\": [{\"metric\": \"time to desired fluid injection rate\", \"value\": \"65% improvement\", \"period\": \"Q3 2025 company report\", \"attribution\": \"Company-reported; StimCommander automation, not Forge alone\", \"source_ids\": [\"R087\"]}, {\"metric\": \"hydraulic efficiency\", \"value\": \"5%-10% improvement\", \"period\": \"Q3 2025 company report\", \"attribution\": \"Company-reported; StimCommander automation, not Forge alone\", \"source_ids\": [\"R087\"]}], \"economics\": \"Liberty reports 65% faster time to desired injection rate and 5-10% better hydraulic efficiency. These relate to fleet automation/StimCommander; no causal contribution from Forge alone is quantified. DigiPrime >30% maintenance savings is hardware result, not AI.\", \"timeline\": [{\"date\": \"2025-10-16\", \"event\": \"Forge launched and described as augmenting StimCommander\", \"source_ids\": [\"R087\"]}], \"limitations\": [\"Metrics are company-reported and are not attributed to Forge specifically; keep digiPrime pump savings separate.\"], \"assessment\": \"Analyst inference: A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.\", \"sources\": [{\"id\": \"R087\", \"title\": \"Liberty Energy Inc. Announces Third Quarter 2025 Financial and Operational Results\", \"url\": \"https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289\", \"publisher\": \"Liberty Energy\", \"date\": \"2025-10-16\", \"type\": \"earnings release\", \"locator\": \"Forge and StimCommander paragraphs\", \"review_status\": \"full text reviewed\", \"supports\": [\"Liberty launched Forge LLM for intelligent asset orchestration integrated with StimCommander rate/pressure control.\", \"Company reports automated control improved time to target fluid injection rate by 65% and hydraulic efficiency by 5-10%; maintenance savings >30% on digiPrime are pump technology results, not AI.\"], \"original_id\": \"liberty\"}], \"field_sources\": {\"business_problem\": [\"R087\"], \"workflow\": [\"R087\"], \"technology_data\": [\"R087\"], \"integration_human_role\": [\"R087\"], \"deployment_scope\": [\"R087\"], \"economics\": [\"R087\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-021\", \"source_ids\": [\"R087\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-118\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-118", "company": "Liberty Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289"], "findings": [{"severity": "medium", "field": "integration", "issue": "Liberty states an LLM interface/orchestration layer for StimCommander; evidence does not show the model directly bypassing safety checks or independently controlling pumps.", "correction": "Describe natural-language command/configuration workflow with human/control-system governance; do not claim autonomous pump control by the LLM.", "source_url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-119", "Company": "Snowflake", "Company_primary": "Snowflake", "Segment": "Big tech", "HQ": "United States", "Theme": "Enterprise AI", "Initiative": "Snowflake Energy Solutions data and partner offering", "Extracted_finding": "The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.", "Context": "Energy organizations must combine IT, operational technology and IoT data while managing specialized planning and analytics tasks.", "Implementation": "Energy Solutions packages Snowflake data platform with more than 30 partner solutions; partners provide specialized workloads such as SAP operational/financial data links and spatial/grid analytics. This is a platform ecosystem; release does not specify a named oil/gas AI workflow deployed by an operator.", "Assessment": "The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Launched 2026-01-27", "Interpretation_limit": "No specific named oil/gas workflow, production deployment or ROI in reviewed announcement.", "Source_IDs": "R090", "Source_date": "2026-01-27", "Source_type": "company announcement", "Source_locator": "Energy Solutions, partner ecosystem, SAP and spatial data", "Source_URL": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-01-27; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-090", "Initiative_ID": "AI-119", "Title": "Snowflake launches Energy Solutions with 30+ partner apps", "Date": "2026-01", "Finding": "Snowflake launched Energy Solutions for its AI Data Cloud, bundling governance, 30+ partner-built apps, industry datasets, Snowflake Intelligence and Cortex AI. Named energy customers include ExxonMobil and Expand Energy.", "Metric": "30+ partner apps; 13,900+ Snowflake customers overall", "Source_IDs": "R090", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Data platforms are becoming the distribution layer for O&G AI apps, so listing as a partner app on Snowflake or Databricks puts a vendor next to data the operator already governs.", "Partners": "Snowflake Cortex AI, Snowflake Intelligence, ExxonMobil, Expand Energy, CARTO, Itron", "Original_date": "2026-01", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-090"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-01-27", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-016"], "Research_search": " {\"key\": \"snowflake-energy\", \"existing_ids\": [\"AI-119\"], \"related_existing_ids\": [], \"company\": \"Snowflake\", \"title\": \"Snowflake Energy Solutions data and partner offering\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched 2026-01-27\", \"practical_relevance\": \"The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.\", \"business_problem\": \"Energy organizations must combine IT, operational technology and IoT data while managing specialized planning and analytics tasks.\", \"workflow\": \"Energy Solutions packages Snowflake data platform with more than 30 partner solutions; partners provide specialized workloads such as SAP operational/financial data links and spatial/grid analytics. This is a platform ecosystem; release does not specify a named oil/gas AI workflow deployed by an operator.\", \"technology_data\": \"Snowflake AI Data Cloud/data sharing and partner apps; release describes cross-domain data foundation, not a novel named O&G model.\", \"integration_human_role\": \"Customer organizations control access and partner applications; detailed AI oversight workflows not disclosed.\", \"deployment_scope\": \"Energy sector offering with named major customers, but no customer-specific production AI implementation described.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No quantified customer economics from the launch.\", \"timeline\": [{\"date\": \"2026-01-27\", \"event\": \"Snowflake announced Energy Solutions.\", \"source_ids\": [\"R090\"]}], \"limitations\": [\"No specific named oil/gas workflow, production deployment or ROI in reviewed announcement.\"], \"assessment\": \"The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.\", \"sources\": [{\"id\": \"R090\", \"title\": \"Snowflake launches Energy Solutions\", \"url\": \"https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/\", \"publisher\": \"Snowflake\", \"date\": \"2026-01-27\", \"type\": \"company announcement\", \"locator\": \"Energy Solutions, partner ecosystem, SAP and spatial data\", \"review_status\": \"full text reviewed\", \"supports\": [\"Energy Solutions is an industry offering combining Snowflake data platform and partner solutions to unify IT/OT/IoT data.\", \"Announcement names customers but does not substantiate a named oil and gas AI production deployment by them.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-018\"}], \"field_sources\": {\"business_problem\": [\"R090\"], \"workflow\": [\"R090\"], \"technology_data\": [\"R090\"], \"integration_human_role\": [\"R090\"], \"deployment_scope\": [\"R090\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-016\", \"source_ids\": [\"R090\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-119\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-119", "company": "Snowflake", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Snowflake Energy Solutions is a data/product ecosystem launch, not a named operator AI deployment.", "correction": "Separate platform offer and partners from actual adoption.", "source_url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-120", "Company": "Cognite", "Company_primary": "Cognite", "Segment": "Startup / scaleup", "HQ": "Norway", "Theme": "AI products", "Initiative": "Atlas AI industrial agent workbench", "Extracted_finding": "Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.", "Context": "Industrial staff search siloed operational/engineering information and need faster cause analysis, work preparation and handover across assets.", "Implementation": "Low-code agents use templates for root cause investigation, work package generation, cross-source research, handover summaries and work-order classification.", "Assessment": "Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Major release 2025-09-09", "Interpretation_limit": "No customer-specific rollout metrics in reviewed release.", "Source_IDs": "RS082; R093", "Source_date": "2025-09-09", "Source_type": "company announcement", "Source_locator": "Agent workbench, templates, governance", "Source_URL": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-09-09; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-093", "Initiative_ID": "AI-120", "Title": "Cognite Atlas AI major release, low-code agent workbench", "Date": "2025-09", "Finding": "Cognite shipped a major Atlas AI release, which it describes as a low-code industrial agent workbench on real-time OT, IT and engineering data. Customers named include Aker BP, ADNOC Offshore, Hess and HMH, with SLB among its partners, and the company says it adds new customers weekly.", "Metric": "New customers onboarded weekly (company claim)", "Source_IDs": "R093", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A low-code agent builder where operators assemble their own agents is the competing model to custom agent services.", "Partners": "Aker BP, ADNOC Offshore, Hess, HMH, Celanese, Idemitsu, SLB, Radix", "Original_date": "2025-09", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-093"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-09-09", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-012"], "Research_search": " {\"key\": \"cognite-atlas\", \"existing_ids\": [\"AI-120\"], \"related_existing_ids\": [\"AI-049\"], \"company\": \"Cognite\", \"title\": \"Atlas AI industrial agent workbench\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Major release 2025-09-09\", \"practical_relevance\": \"Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.\", \"business_problem\": \"Industrial staff search siloed operational/engineering information and need faster cause analysis, work preparation and handover across assets.\", \"workflow\": \"Low-code agents use templates for root cause investigation, work package generation, cross-source research, handover summaries and work-order classification.\", \"technology_data\": \"Connects live OT/IT/engineering data using industrial knowledge graph with data governance and granular access; announcement describes architecture generally.\", \"integration_human_role\": \"Industrial users configure agents; role/access governance described, but customer-side approval policy per workflow not detailed.\", \"deployment_scope\": \"Product available across industrial customers; Cognite names Aker BP and others as customers but release does not tie each named customer to each template.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No quantified benefit metrics in the announcement.\", \"timeline\": [{\"date\": \"2025-09-09\", \"event\": \"Cognite announced Atlas AI major release.\", \"source_ids\": [\"RS082\"]}], \"limitations\": [\"No customer-specific rollout metrics in reviewed release.\"], \"assessment\": \"Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.\", \"sources\": [{\"id\": \"RS082\", \"title\": \"Cognite Atlas AI major release accelerates industrial AI\", \"url\": \"https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release\", \"publisher\": \"Cognite\", \"date\": \"2025-09-09\", \"type\": \"company announcement\", \"locator\": \"Agent workbench, templates, governance\", \"review_status\": \"full text reviewed\", \"supports\": [\"Atlas AI provides low-code agent workbench over industrial OT/IT/engineering data.\", \"Templates cover root-cause analysis, work package generation, cross-source research, handover summaries and work-order classification.\", \"Cognite names energy customers; announcement is product update not independent evaluation of customer results.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-015\"}], \"field_sources\": {\"business_problem\": [\"RS082\"], \"workflow\": [\"RS082\"], \"technology_data\": [\"RS082\"], \"integration_human_role\": [\"RS082\"], \"deployment_scope\": [\"RS082\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-012\", \"source_ids\": [\"RS082\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-120\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-120", "company": "Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release"], "findings": [{"severity": "medium", "field": "stage", "issue": "Atlas AI launch describes a commercial agent workbench; no company-specific deployment or outcome is inherent in the product release.", "correction": "Keep capability claims separate from customer proof.", "source_url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-121", "Company": "Cognite", "Company_primary": "Cognite", "Segment": "Industrial AI — chemicals (adjacent to oil and gas)", "HQ": "Norway", "Theme": "AI products", "Initiative": "NVIDIA-accelerated time-series anomaly detection with Celanese", "Extracted_finding": "Noisy continuous sensor streams can hide equipment anomalies and delay intervention.", "Context": "Industrial equipment produces continuous noisy readings and performance metrics with context scattered across operational information.", "Implementation": "Initial joint use case analyzes real-time time-series data to identify critical anomalies and support proactive intervention before disruption.", "Assessment": "Analyst inference: The named time-series anomaly workflow is technically concrete but is an adjacent chemicals example, not evidence of oil-and-gas operator deployment.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Integration announced 2025-10-14; Celanese use case aims at anomaly-detection deployment", "Interpretation_limit": "Initial anomaly use case is a joint aim/use case; release does not disclose production scale or measured accuracy. Customer is Celanese, a chemicals manufacturer; treat as adjacent process-industry evidence, not oil-and-gas operator adoption.", "Source_IDs": "R094", "Source_date": "2025-10-14", "Source_type": "company announcement", "Source_locator": "Celanese time-series anomaly detection", "Source_URL": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-10-14; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-094", "Initiative_ID": "AI-121", "Title": "Cognite integrates NVIDIA accelerated computing for agents", "Date": "2025-10", "Finding": "Cognite integrated its Atlas AI and Data Fusion platform with NVIDIA accelerated computing and software. The first use case is real-time anomaly detection in time-series data with Celanese, and Cognite later integrated NVIDIA NV-Tesseract for forecasting.", "Metric": "No quantified metric supplied", "Source_IDs": "R094", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-series anomaly detection at scale is the core agent workload in operations, and partnering with a GPU vendor adds credibility and lowers compute cost.", "Partners": "NVIDIA, Celanese, NV-Tesseract", "Original_date": "2025-10", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-094"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-10-14", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-021"], "Research_search": " {\"key\": \"cognite-nvidia-celanese\", \"existing_ids\": [\"AI-121\"], \"related_existing_ids\": [\"AI-120\"], \"company\": \"Cognite\", \"title\": \"NVIDIA-accelerated time-series anomaly detection with Celanese\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Adjacent process industry (chemicals)\", \"stage\": \"Integration announced 2025-10-14; Celanese use case aims at anomaly-detection deployment\", \"practical_relevance\": \"Noisy continuous sensor streams can hide equipment anomalies and delay intervention.\", \"business_problem\": \"Industrial equipment produces continuous noisy readings and performance metrics with context scattered across operational information.\", \"workflow\": \"Initial joint use case analyzes real-time time-series data to identify critical anomalies and support proactive intervention before disruption.\", \"technology_data\": \"Cognite Atlas AI and Data Fusion integrated with NVIDIA accelerated computing and software; NV-Tesseract time-series foundation model family identified.\", \"integration_human_role\": \"Agents are described as producing traceable, trustworthy results in maintenance workflows; specific alert review and intervention authority are not disclosed.\", \"deployment_scope\": \"Named customer Celanese; specialty chemicals, not oil and gas. Cognite says Celanese has scaled other use cases to 50+ sites; release does not assign NVIDIA anomaly system to all sites.\", \"partners\": [\"NVIDIA\", \"Celanese\"], \"disclosed_results\": [], \"economics\": \"No customer-specific economics or false-positive metrics disclosed.\", \"timeline\": [{\"date\": \"2025-10-14\", \"event\": \"Cognite announced the accelerated integration and first Celanese time-series anomaly use case.\", \"source_ids\": [\"R094\"]}], \"limitations\": [\"Initial anomaly use case is a joint aim/use case; release does not disclose production scale or measured accuracy.\", \"Customer is Celanese, a chemicals manufacturer; treat as adjacent process-industry evidence, not oil-and-gas operator adoption.\"], \"assessment\": \"Analyst inference: The named time-series anomaly workflow is technically concrete but is an adjacent chemicals example, not evidence of oil-and-gas operator deployment.\", \"sources\": [{\"id\": \"R094\", \"title\": \"Cognite and NVIDIA partner to advance industrial AI performance\", \"url\": \"https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and\", \"publisher\": \"Cognite\", \"date\": \"2025-10-14\", \"type\": \"company announcement\", \"locator\": \"Celanese time-series anomaly detection\", \"review_status\": \"full text reviewed\", \"supports\": [\"Atlas AI/Data Fusion integrated with NVIDIA accelerated computing and software.\", \"Initial Celanese project targets real-time anomaly detection in industrial time-series data.\", \"Celanese reports over 50 global Cognite sites, but release does not say all sites use the NVIDIA model.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"R094\"}], \"field_sources\": {\"business_problem\": [\"R094\"], \"workflow\": [\"R094\"], \"technology_data\": [\"R094\"], \"integration_human_role\": [\"R094\"], \"deployment_scope\": [\"R094\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-021\", \"source_ids\": [\"R094\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-121\"]}", "Research_status": "Expanded source review", "Research_category": "Adjacent process industry (chemicals)", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "The release describes a product/customer collaboration and target use case; realized Celanese deployment/results are not established.", "source_url": ""}], "Audit_review": {"id": "AI-121", "company": "Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and"], "findings": [{"severity": "medium", "field": "classification / deployment scope", "issue": "The customer example is Celanese, a chemicals manufacturer; its time-series anomaly detection use case is process-industry adjacent to oil and gas.", "correction": "Keep the AI use case but label the customer and example as chemicals/adjacent process industry; exclude it from O&G-operator adoption comparisons.", "source_url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and"}], "residual_gaps": ["The release describes a product/customer collaboration and target use case; realized Celanese deployment/results are not established."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-122", "Company": "Novi Labs", "Company_primary": "Novi Labs", "Segment": "Startup / scaleup", "HQ": "United States", "Theme": "Investment AI", "Initiative": "Invictus $35m growth investment in Novi Labs", "Extracted_finding": "Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.", "Context": "Novi seeks capital to expand upstream energy intelligence, product work and acquisition strategy.", "Implementation": "Invictus investment funds product innovation and expansion of a machine-learning and reservoir-science platform used in upstream forecasting and capital decisions. This is an investor financing event, not itself a new operator deployment.", "Assessment": "Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.", "Reported_metric": "Growth investment: $35 million (Company-reported)", "Stage_at_disclosure": "Investment announced 2025-06-25", "Interpretation_limit": "Funding announcement gives no customer-level implementation details.", "Source_IDs": "R095", "Source_date": "2025-06-25", "Source_type": "company announcement", "Source_locator": "Funding and platform overview", "Source_URL": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-06-25; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": "Invictus Growth Partners investment in Novi Labs", "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-095", "Initiative_ID": "AI-122", "Title": "Invictus invests $35m in Novi Labs", "Date": "2025-06-25", "Finding": "Novi Labs announced $35 million from Invictus Growth Partners to expand its energy AI analytics business. ML forecasts support well design and upstream capital allocation.", "Metric": "$35m investment / company reports >$50bn annual capital decisions through platform", "Source_IDs": "R095", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Growth equity is backing vertical AI analytics for upstream investing, where accurate forecasts on proprietary data are the moat.", "Partners": "Invictus Growth Partners, Shell, ExxonMobil, Chevron, Devon (reported users)", "Original_date": "2025-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-095"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-06-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-004"], "Research_search": " {\"key\": \"novi-investment\", \"existing_ids\": [\"AI-122\"], \"related_existing_ids\": [], \"company\": \"Novi Labs\", \"title\": \"Invictus $35m growth investment in Novi Labs\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Investment announced 2025-06-25\", \"practical_relevance\": \"Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.\", \"business_problem\": \"Novi seeks capital to expand upstream energy intelligence, product work and acquisition strategy.\", \"workflow\": \"Invictus investment funds product innovation and expansion of a machine-learning and reservoir-science platform used in upstream forecasting and capital decisions. This is an investor financing event, not itself a new operator deployment.\", \"technology_data\": \"Novi says platform combines ML, proprietary data and reservoir science; architecture and incremental funded deliverables are not detailed.\", \"integration_human_role\": \"No operating customer workflow change or human review process disclosed in financing release.\", \"deployment_scope\": \"Company/vendor platform market-wide; no deployment counts tied to the financing.\", \"partners\": [\"Invictus Growth Partners\"], \"disclosed_results\": [{\"metric\": \"Growth investment\", \"value\": \"$35 million\", \"period\": \"2025-06-25 announcement\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R095\"]}], \"economics\": \"$35m financing disclosed; use of proceeds is high-level; no return or AI-specific spend breakdown.\", \"timeline\": [{\"date\": \"2025-06-25\", \"event\": \"Novi announced $35m investment from Invictus Growth Partners.\", \"source_ids\": [\"R095\"]}], \"limitations\": [\"Funding announcement gives no customer-level implementation details.\"], \"assessment\": \"Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.\", \"sources\": [{\"id\": \"R095\", \"title\": \"Novi Labs raises $35 million from Invictus Growth Partners\", \"url\": \"https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/\", \"publisher\": \"Novi Labs\", \"date\": \"2025-06-25\", \"type\": \"company announcement\", \"locator\": \"Funding and platform overview\", \"review_status\": \"full text reviewed\", \"supports\": [\"Invictus Growth Partners invested $35 million to support Novi product innovation, expansion and acquisitions.\", \"Novi describes platform combining machine learning, proprietary datasets and reservoir science for upstream capital decisions.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-005\"}], \"field_sources\": {\"business_problem\": [\"R095\"], \"workflow\": [\"R095\"], \"technology_data\": [\"R095\"], \"integration_human_role\": [], \"deployment_scope\": [\"R095\"], \"economics\": [\"R095\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-004\", \"source_ids\": [\"R095\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-122\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-122", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/"], "findings": [{"severity": "low", "field": "classification", "issue": "The $35m transaction is a growth investment in Novi Labs; it does not evidence operator deployment or customer impact.", "correction": "Keep as investment/acquisition event and separate from Novi product workflow dossiers.", "source_url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/"}], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-123", "Company": "Novi Labs", "Company_primary": "Novi Labs", "Segment": "Startup / scaleup", "HQ": "United States", "Theme": "Investment AI", "Initiative": "Acquisitions combine upstream AI forecasting with Canada and market intelligence", "Extracted_finding": "Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.", "Context": "Cross-border upstream evaluation often omits commodity pricing, transport capacity, basis and demand, making asset economics incomplete.", "Implementation": "Novi acquired Turing Analytics (EVA) to extend analytical coverage into Canadian basins, followed by RBN Energy to add commodity/midstream research. Combined thesis connects well forecasts and inventory quality with transport, commodity flows, basis and downstream demand for investment/commercial decisions.", "Assessment": "Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Turing announced 2026-04-29; RBN announced 2026-05-28", "Interpretation_limit": "Acquisition prices and integration status not disclosed. Turing and RBN are distinct acquisitions and do not evidence a single completed AI product workflow.", "Source_IDs": "RS078; R096", "Source_date": "2026-05-28", "Source_type": "company announcement", "Source_locator": "Acquisition rationale and geographic coverage", "Source_URL": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-05-28; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-096", "Initiative_ID": "AI-123", "Title": "Novi acquires Turing Analytics and RBN Energy", "Date": "2026-05", "Finding": "Novi Labs acquired Calgary-based Turing Analytics (EVA, an April 2026 deal) and Houston-based RBN Energy (a May 2026 deal). The acquisitions connect upstream analytics with Canadian coverage and midstream and commodity-market intelligence in one decision platform.", "Metric": "RBN reaches 120,000+ monthly users", "Source_IDs": "R096", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Well-funded vertical AI players are consolidating data and audience assets, so owning proprietary content becomes a competitive lever.", "Partners": "Turing Analytics (EVA), RBN Energy", "Original_date": "2026-05", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-096"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-28", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-006"], "Research_search": " {\"key\": \"novi-north-america-platform\", \"existing_ids\": [\"AI-123\"], \"related_existing_ids\": [], \"company\": \"Novi Labs\", \"title\": \"Acquisitions combine upstream AI forecasting with Canada and market intelligence\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Turing announced 2026-04-29; RBN announced 2026-05-28\", \"practical_relevance\": \"Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.\", \"business_problem\": \"Cross-border upstream evaluation often omits commodity pricing, transport capacity, basis and demand, making asset economics incomplete.\", \"workflow\": \"Novi acquired Turing Analytics (EVA) to extend analytical coverage into Canadian basins, followed by RBN Energy to add commodity/midstream research. Combined thesis connects well forecasts and inventory quality with transport, commodity flows, basis and downstream demand for investment/commercial decisions.\", \"technology_data\": \"Novi says its platform combines proprietary well-level data and ML production forecasting; acquisitions bring Canadian analytics and RBN market research/expertise. Integration and technical architecture across acquisitions not documented.\", \"integration_human_role\": \"Acquisition releases describe decision support for operators, investors and commercial teams; no automated approval or human governance details.\", \"deployment_scope\": \"North American energy market; Turing states client companies cover >50% Canadian production; RBN reach claims refer to research readership/users, not AI deployment.\", \"partners\": [\"Turing Analytics\", \"RBN Energy\"], \"disclosed_results\": [], \"economics\": \"Deal terms undisclosed. No evidence yet that post-acquisition integrated product has shipped or generated returns.\", \"timeline\": [{\"date\": \"2026-04-29\", \"event\": \"Novi announced acquisition of Turing Analytics/EVA.\", \"source_ids\": [\"RS078\"]}, {\"date\": \"2026-05-28\", \"event\": \"Novi announced acquisition of RBN Energy.\", \"source_ids\": [\"R096\"]}], \"limitations\": [\"Acquisition prices and integration status not disclosed.\", \"Turing and RBN are distinct acquisitions and do not evidence a single completed AI product workflow.\"], \"assessment\": \"Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.\", \"sources\": [{\"id\": \"RS078\", \"title\": \"Novi Labs acquires Turing Analytics\", \"url\": \"https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/\", \"publisher\": \"Novi Labs\", \"date\": \"2026-04-29\", \"type\": \"company announcement\", \"locator\": \"Acquisition rationale and geographic coverage\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Novi acquired Turing Analytics/EVA to combine US and Canadian upstream datasets and analytics for investment evaluation.\", \"Novi says Turing clients represent over 50% of Canadian oil and gas production; company-reported.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-007\"}, {\"id\": \"R096\", \"title\": \"Novi Labs acquires RBN Energy\", \"url\": \"https://novilabs.com/news/novi-labs-acquires-rbn-energy/\", \"publisher\": \"Novi Labs\", \"date\": \"2026-05-28\", \"type\": \"company announcement\", \"locator\": \"Combination and market workflows\", \"review_status\": \"full text reviewed\", \"supports\": [\"RBN acquisition combines Novi upstream AI forecasts with commodity, midstream and demand intelligence.\", \"Decision context includes basis, transport capacity, commodity flows and end-market pricing.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-008\"}], \"field_sources\": {\"business_problem\": [\"R096\"], \"workflow\": [\"RS078\", \"R096\"], \"technology_data\": [\"RS078\", \"R096\"], \"integration_human_role\": [\"R096\"], \"deployment_scope\": [\"RS078\", \"R096\"], \"economics\": [\"RS078\", \"R096\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-006\", \"source_ids\": [\"RS078\", \"R096\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-123\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Novi/Turing/RBN acquisitions already form one integrated product.", "reason": "Acquisition announcements describe strategy and intended unified platform; reviewed sources do not establish technical integration completion.", "source_url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/"}], "Audit_review": {"id": "AI-123", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "https://novilabs.com/news/novi-labs-acquires-rbn-energy/"], "findings": [{"severity": "low", "field": "classification / deployment", "issue": "Acquisitions and planned integration enlarge a vendor portfolio but do not alone establish an integrated AI workflow in production.", "correction": "Describe deal/integration stage explicitly; avoid implying realized customer outcomes.", "source_url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "investment", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-124", "Company": "Collide", "Company_primary": "Collide", "Segment": "Startup / scaleup", "HQ": "United States", "Theme": "Investment AI", "Initiative": "Collide $5m seed and energy-specific GenAI knowledge platform", "Extracted_finding": "Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.", "Context": "Petroleum engineers and energy teams need technical knowledge from authoritative papers, internal processes and practitioner Q&A without time-consuming manual search.", "Implementation": "Collide raised seed financing to scale its RAG/LLM platform, which retrieves and synthesizes technical content into energy-focused cited answers. The original product evolved from a community/knowledge hub into an enterprise AI product.", "Assessment": "Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.", "Reported_metric": "Seed funding: $5 million (Company-reported); Estimated petroleum-engineer time savings: 25%-33% (Company-reported); Community users: 6,000+ (not equivalent to AI enterprise users) (Company-reported)", "Stage_at_disclosure": "Seed announced 2025-04-15; platform rollout underway", "Interpretation_limit": "Funding announcement conflates community footprint and enterprise service; it does not establish 6,000 AI users. 25%-33% savings is vendor estimate with no study design disclosed.", "Source_IDs": "R101", "Source_date": "2025-04-15", "Source_type": "company announcement", "Source_locator": "Seed funding and energy knowledge platform", "Source_URL": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-04-15; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-101", "Initiative_ID": "AI-124", "Title": "Collide raises $5M seed for O&G genAI platform", "Date": "2025-04", "Finding": "Houston-based Collide closed a $5M seed round led by Mercury Fund, with operator angels Bryan Sheffield, Billy Quinn and David Albin. It serves public, private and PE-backed E&Ps and grew from an industry community platform.", "Metric": "$5M seed; 6,000+ community users growing 11% MoM (at raise)", "Source_IDs": "R101", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Building a practitioner community first gave Collide a distribution base before selling enterprise AI.", "Partners": "Mercury Fund", "Original_date": "2025-04", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-101"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-04-15", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-022"], "Research_search": " {\"key\": \"collide-seed-platform\", \"existing_ids\": [\"AI-124\"], \"related_existing_ids\": [\"AI-126\"], \"company\": \"Collide\", \"title\": \"Collide $5m seed and energy-specific GenAI knowledge platform\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Seed announced 2025-04-15; platform rollout underway\", \"practical_relevance\": \"Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.\", \"business_problem\": \"Petroleum engineers and energy teams need technical knowledge from authoritative papers, internal processes and practitioner Q&A without time-consuming manual search.\", \"workflow\": \"Collide raised seed financing to scale its RAG/LLM platform, which retrieves and synthesizes technical content into energy-focused cited answers. The original product evolved from a community/knowledge hub into an enterprise AI product.\", \"technology_data\": \"Retrieval-augmented generation and large language models over a curated knowledge base combining technical papers, internal processes and Q&A.\", \"integration_human_role\": \"Answers are citation-supported; announcement emphasizes engineer knowledge work, but no formal review gates or enterprise data controls specified.\", \"deployment_scope\": \"Company says it services public/private enterprise and PE-backed E&Ps; user metric combines community and platform users, so not enterprise deployments.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Seed funding\", \"value\": \"$5 million\", \"period\": \"2025-04-15\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R101\"]}, {\"metric\": \"Estimated petroleum-engineer time savings\", \"value\": \"25%-33%\", \"period\": \"Company estimate at announcement\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R101\"]}, {\"metric\": \"Community users\", \"value\": \"6,000+ (not equivalent to AI enterprise users)\", \"period\": \"2025-04-15 announcement\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R101\"]}], \"economics\": \"$5m seed led by Mercury Fund. Company estimates petroleum engineers save 25%-33% of time; not independent or customer-specific.\", \"timeline\": [{\"date\": \"2025-04-15\", \"event\": \"Collide $5m seed and energy-specific GenAI knowledge platform disclosed.\", \"source_ids\": [\"R101\"]}], \"limitations\": [\"Funding announcement conflates community footprint and enterprise service; it does not establish 6,000 AI users.\", \"25%-33% savings is vendor estimate with no study design disclosed.\"], \"assessment\": \"Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.\", \"sources\": [{\"id\": \"R101\", \"title\": \"Collide secures $5m to scale first GenAI platform for energy sector\", \"url\": \"https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector\", \"publisher\": \"Collide (via Business Wire)\", \"date\": \"2025-04-15\", \"type\": \"company announcement\", \"locator\": \"Seed funding and energy knowledge platform\", \"review_status\": \"full text reviewed\", \"supports\": [\"Collide announced a $5m seed led by Mercury Fund; proceeds support platform and engineering expansion.\", \"Platform uses retrieval augmented generation and LLMs over curated technical papers, internal processes and Q&A to provide cited, domain-specific energy answers.\", \"Collide reported 6,000+ community users and estimated 25%-33% time savings for petroleum engineers; these are company estimates and community users are not equivalent to deployed enterprise users.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-033\"}], \"field_sources\": {\"business_problem\": [\"R101\"], \"workflow\": [\"R101\"], \"technology_data\": [\"R101\"], \"integration_human_role\": [\"R101\"], \"deployment_scope\": [\"R101\"], \"economics\": [\"R101\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-022\", \"source_ids\": [\"R101\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-124\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-124", "company": "Collide", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-125", "Company": "Winn Resources / Collide", "Company_primary": "Winn Resources", "Segment": "Startup / scaleup", "HQ": "United States", "Theme": "AI products", "Initiative": "Winn Resources Texas RRC W-10/G-10 filing workflow", "Extracted_finding": "Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.", "Context": "Monthly regulatory filings require assembling well test and production data; missing values create delay and rework.", "Implementation": "Winn describes a protocol for investigating missing-data flags. The vendor release reports a roughly 50-well cycle and retains staff review; future revenue-statement extraction remains a plan.", "Assessment": "Concrete operator workflow with a clear manual process and reported result. Keep the RIGGS/Winn workflow distinct from Collide’s platform launch and Battalion investment.", "Reported_metric": "One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.", "Stage_at_disclosure": "W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.", "Interpretation_limit": "The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence.", "Source_IDs": "RS089; RS090; R102; M29-010", "Source_date": "2026-03-05", "Source_type": "Vendor-issued press release with named customer executive quotation", "Source_locator": "PRNewswire release sourced to Collide, 5 March 2026; 'Challenge', 'Solution', and 'Looking Ahead' sections. Specific case figures and approval process in paragraphs equivalent to 50-well cycle, 2-4 hour prior process, missing test flags, human review, and planned revenue extraction.", "Source_URL": "https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html", "Evidence_status": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-03-05; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-102", "Initiative_ID": "AI-125", "Title": "RIGGS energy LLM automates RRC filings", "Date": "2026-05", "Finding": "Collide rolled out RIGGS, an energy LLM that returns cited answers from authoritative sources. With Winn Resources it automated Texas Railroad Commission regulatory filings, and Collide says RIGGS beat general frontier models on a petroleum-engineering exam subset.", "Metric": "Filing time cut from hours to 20 min for 50 wells (>95%); 67.5% on 40-question PE exam subset vs Grok 4 62.5%, Claude Sonnet 4.5 52.5%", "Source_IDs": "R102", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Regulatory filings are a narrow, measurable back-office wedge for small operators, and domain benchmarks against frontier models are becoming a marketing tool.", "Partners": "Winn Resources", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "M29-D-AI-125", "Initiative_ID": "AI-125", "Title": "AI-supported Texas Railroad Commission monthly W-10/G-10 filings (detail enrichment; regulatory workflow with accounting adjacency)", "Date": "2026-03-05", "Finding": "Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.", "Metric": "One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.", "Source_IDs": "M29-010", "Evidence": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "Origin": "29 September 2026 gap review"}], "Related_IDs": [], "Reference_IDs": ["REF-102"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-03-05", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["missing-sept29-ai-125", "software_capital-023"], "Research_search": "{\"key\": \"missing-sept29-ai-125\", \"packet_id\": \"missing-sept29-ai-125\", \"record_ids\": [\"AI-125\"], \"existing_ids\": [\"AI-125\"], \"related_existing_ids\": [], \"company\": \"Winn Resources\", \"title\": \"AI-supported Texas Railroad Commission monthly W-10/G-10 filings (detail enrichment; regulatory workflow with accounting adjacency)\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.\", \"practical_relevance\": \"AI-125 already captures the headline time reduction but not the source’s missing-data exception trigger, customer-built follow-up protocol, full human approval gate, range of old cycle time, precise 50-well quote, or the prospective accounting use case. The new facts expand commercial/admin coverage without claiming AI use in live revenue accounting.\", \"business_problem\": \"\", \"workflow\": \"Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.\", \"technology_data\": \"Winn describes a protocol for investigating missing-data flags. The vendor release reports a roughly 50-well cycle and retains staff review; future revenue-statement extraction remains a plan.\", \"integration_human_role\": \"\", \"deployment_scope\": \"W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.\", \"period\": \"2026-03-05\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-010\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2026-03-05\", \"event\": \"W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.\", \"source_ids\": [\"M29-010\"]}], \"limitations\": [\"The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence.\"], \"assessment\": \"AI-125 already captures the headline time reduction but not the source’s missing-data exception trigger, customer-built follow-up protocol, full human approval gate, range of old cycle time, precise 50-well quote, or the prospective accounting use case. The new facts expand commercial/admin coverage without claiming AI use in live revenue accounting.\", \"sources\": [{\"id\": \"M29-010\", \"title\": \"Winn Resources Partners with Collide to Turn Operational Data Into a Competitive Advantage\", \"url\": \"https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html\", \"publisher\": \"Winn Resources\", \"date\": \"2026-03-05\", \"type\": \"Vendor-issued press release with named customer executive quotation\", \"locator\": \"PRNewswire release sourced to Collide, 5 March 2026; 'Challenge', 'Solution', and 'Looking Ahead' sections. Specific case figures and approval process in paragraphs equivalent to 50-well cycle, 2-4 hour prior process, missing test flags, human review, and planned revenue extraction.\", \"review_status\": \"Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.\", \"supports\": [\"Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.\", \"One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-010\"], \"field_sources\": {\"workflow\": [\"M29-010\"], \"technology_data\": [\"M29-010\"], \"deployment_scope\": [\"M29-010\"], \"practical_relevance\": [\"M29-010\"]}, \"verification\": \"verified partner\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Updated · 29 September 2026", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"reason": "The vendor’s >95% headline does not reconcile to its published time figures without another baseline."}], "Audit_review": {"id": "AI-125", "company": "Winn Resources", "outcome": "partially supported", "reviewer_outcome": "Corrected baseline arithmetic and source attribution", "checks": ["Source", "Metric arithmetic", "Company identity"], "source_urls_reviewed": ["https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html"], "findings": [{"severity": "medium", "field": "Reported result", "issue": "The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence.", "correction": "Use the observed cycle and label the unreconciled headline.", "source_url": "https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html"}], "residual_gaps": ["The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence."], "reviewed_at": "2026-09-29"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-126", "Company": "Collide", "Company_primary": "Collide", "Segment": "Startup / scaleup", "HQ": "United States", "Theme": "Investment AI", "Initiative": "Battalion strategic investment and planned Collide Operations System rollout", "Extracted_finding": "High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.", "Context": "Small operator has more than 100 TB across contracts, land records, production history and files; unstructured data impedes production and business analysis.", "Implementation": "Battalion says initial rollout will build a queryable company model across records, then target gas marketing reconciliation, field invoice review, power use/hedging and well failure diagnosis. Riggs is planned for production engineers at about day 90. Battalion retains review/control; system subscription and strategic equity investment disclosed.", "Assessment": "High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Strategic investment and rollout announced 2026-09-08", "Interpretation_limit": "Do not count Collide general product claims as verified Battalion implementation. Multiple apps are future plans; Riggs expected around day 90, after date of press release.", "Source_IDs": "R103; RS080", "Source_date": "2026-09-08", "Source_type": "company announcement", "Source_locator": "Strategic investment, intended deployment and Riggs", "Source_URL": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-09-08; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-103", "Initiative_ID": "AI-126", "Title": "Battalion Oil makes strategic investment in Collide", "Date": "2026-09", "Finding": "Public E&P Battalion Oil made an undisclosed strategic investment in Collide and is deploying its Riggs AI-native operations system across its upstream operations. Battalion plans to unify more than 100 TB of operational data, with Riggs reaching its production engineers around day 90.", "Metric": "100+ TB data; ~90-day time to production engineers; Collide cites $20B addressable market", "Source_IDs": "R103", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Operators investing in their AI vendors aligns incentives and creates reference customers, and a stated 90-day time-to-value target is a useful benchmark.", "Partners": "Battalion Oil, Riggs", "Original_date": "2026-09", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-103"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-09-08", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-010"], "Research_search": " {\"key\": \"collide-battalion\", \"existing_ids\": [\"AI-126\"], \"related_existing_ids\": [\"AI-124\", \"AI-125\"], \"company\": \"Battalion Oil\", \"title\": \"Battalion strategic investment and planned Collide Operations System rollout\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Investment or acquisition\", \"stage\": \"Strategic investment and rollout announced 2026-09-08\", \"practical_relevance\": \"High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.\", \"business_problem\": \"Small operator has more than 100 TB across contracts, land records, production history and files; unstructured data impedes production and business analysis.\", \"workflow\": \"Battalion says initial rollout will build a queryable company model across records, then target gas marketing reconciliation, field invoice review, power use/hedging and well failure diagnosis. Riggs is planned for production engineers at about day 90. Battalion retains review/control; system subscription and strategic equity investment disclosed.\", \"technology_data\": \"Collide says it uses a connected asset model and source-cited answers with deterministic petroleum calculations; Battalion release does not confirm precise model stack/data ingestion or availability dates for all planned apps.\", \"integration_human_role\": \"Company expressly describes human review/control of outputs. Operators and production engineers are intended users.\", \"deployment_scope\": \"Single named customer, Battalion; >100TB is source-company description of input corpus, not total production deployment scale.\", \"partners\": [\"Battalion Oil\"], \"disclosed_results\": [], \"economics\": \"Subscription cost/equity terms and realized economics not stated. Announcement anticipates reduced analysis time; no specific baseline result independently verified.\", \"timeline\": [{\"date\": \"2026-09-08\", \"event\": \"Battalion announced strategic investment in Collide and planned deployment.\", \"source_ids\": [\"R103\"]}], \"limitations\": [\"Do not count Collide general product claims as verified Battalion implementation.\", \"Multiple apps are future plans; Riggs expected around day 90, after date of press release.\"], \"assessment\": \"High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.\", \"sources\": [{\"id\": \"R103\", \"title\": \"Battalion Oil announces strategic investment in Collide\", \"url\": \"https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html\", \"publisher\": \"Battalion Oil Corporation\", \"date\": \"2026-09-08\", \"type\": \"company announcement\", \"locator\": \"Strategic investment, intended deployment and Riggs\", \"review_status\": \"full text reviewed\", \"supports\": [\"Battalion made strategic equity investment and will deploy Collide Operations System under subscription.\", \"Initial plan unites more than 100 TB of records, contracts and production history; planned use cases include gas reconciliation, field invoices, power, and failure diagnosis.\", \"Riggs is expected to be available to production engineers around day 90; human review/control retained.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-012\"}, {\"id\": \"RS080\", \"title\": \"Collide product: Operations System\", \"url\": \"https://collide.io/product\", \"publisher\": \"Collide\", \"date\": \"undated\", \"type\": \"company webpage\", \"locator\": \"Architecture/capability page\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Collide describes petroleum-specific connected asset model and grounded answers with source citations.\", \"Company describes deterministic tested calculations and deployment in customer cloud/data platform; page undated.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-013\"}], \"field_sources\": {\"business_problem\": [\"R103\"], \"workflow\": [\"R103\"], \"technology_data\": [\"R103\", \"RS080\"], \"integration_human_role\": [\"R103\", \"RS080\"], \"deployment_scope\": [\"R103\"], \"economics\": [\"R103\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-010\", \"source_ids\": [\"R103\", \"RS080\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-126\"]}", "Research_status": "Expanded source review", "Research_category": "Investment or acquisition", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-126", "company": "Collide", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "https://collide.io/product"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "investment", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-127", "Company": "Applied Computing", "Company_primary": "Applied Computing", "Segment": "Startup / scaleup", "HQ": "United Kingdom", "Theme": "AI products", "Initiative": "$20M Series A for Orbital O&G foundation model", "Extracted_finding": "London-based Applied Computing raised a $20M Series A led by KBR, with Databricks Ventures participating. Its Orbital foundation model combines time-series, physics and language modeling to predict plant state, find anomalies and simulate fixes.", "Context": "London-based Applied Computing raised a $20M Series A led by KBR, with Databricks Ventures participating. Its Orbital foundation model combines time-series, physics and language modeling to predict plant state, find anomalies and simulate fixes.", "Implementation": "Disclosed partners / technologies: KBR (INSITE 3.0), Databricks Ventures, Wipro. Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Strategic investors (an engineering firm and a data platform) give distribution, and a hybrid physics-time-series-language model is attracting funding.", "Reported_metric": "$20M Series A; 'double-digit millions' ARR within ~18 months; facilities use <8% of sensor data", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Imported claim and interpretation from the supplied reference. Source text, date, scope and outcome attribution require independent verification.", "Source_IDs": "R104", "Source_date": "2026-07", "Source_type": "TechCrunch", "Source_locator": "See source and disclosure timeline", "Source_URL": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Reference date within window; date not independently verified", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-104", "Initiative_ID": "AI-127", "Title": "$20M Series A for Orbital O&G foundation model", "Date": "2026-07", "Finding": "London-based Applied Computing raised a $20M Series A led by KBR, with Databricks Ventures participating. Its Orbital foundation model combines time-series, physics and language modeling to predict plant state, find anomalies and simulate fixes.", "Metric": "$20M Series A; 'double-digit millions' ARR within ~18 months; facilities use <8% of sensor data", "Source_IDs": "R104", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Strategic investors (an engineering firm and a data platform) give distribution, and a hybrid physics-time-series-language model is attracting funding.", "Partners": "KBR (INSITE 3.0), Databricks Ventures, Wipro", "Original_date": "2026-07", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-104"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-07", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Commercial AI product", "Audit_flags": [{"claim": "Applied Computing $20m Series A and claims about broad whole-plant model deployments.", "reason": "No primary company/investor funding release reviewed; only secondary reporting surfaced. The May 2025 primary source confirms a £9m seed and one unnamed refinery claim, not a later round or independently validated whole-plant performance.", "source_url": "https://appliedcomputing.com/press-release-28052025"}, {"claim": "$20m Series A led by KBR and Databricks Ventures; entire-plant model performance.", "reason": "No primary financing announcement located in bounded review; legacy evidence is TechCrunch secondary. Primary Applied Computing/KBR Jan 2026 release verifies INSITE 3.0 powered by Orbital for ammonia process optimization, but does not verify $20m round or the full-plant claims. Preserve as secondary-only and distinct from the £9m seed/unnamed refinery record.", "source_url": "https://appliedcomputing.com/press-release-21012026"}, {"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-127", "company": "Applied Computing", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/"], "findings": [{"severity": "medium", "field": "source / attribution", "issue": "The $20m Series A record relies on secondary TechCrunch in the legacy source ledger; an issuer/investor primary financing notice was not reopened in this audit.", "correction": "Keep financing amount and operational claims secondary-only until issuer/investor confirmation; separate from the earlier May 2025 seed raise.", "source_url": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "scale_only", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-128", "Company": "Seeq", "Company_primary": "Seeq", "Segment": "Software incumbent", "HQ": "United States", "Theme": "Enterprise AI", "Initiative": "Seeq Intelligence and Agent Q for operational decisions", "Extracted_finding": "Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.", "Context": "Industrial knowledge and event history are siloed, making it difficult for operations personnel to investigate deviations consistently and scale expert practice.", "Implementation": "Agent Q assembles current process data, historical events, previous analyses, documents and expert context into traceable investigations and prioritized recommendations. Users can build multistep custom agents, scheduled or triggered, and connect agents to external systems for context or action.", "Assessment": "Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Launched 2026-03-03; available now", "Interpretation_limit": "No named energy customer case, implementation architecture or results in launch disclosure.", "Source_IDs": "RS083; R105", "Source_date": "2026-03-03", "Source_type": "company announcement", "Source_locator": "Agent Q, build-your-own agents and integrations", "Source_URL": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-03-03; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-105", "Initiative_ID": "AI-128", "Title": "Seeq Intelligence with Agent Q and custom agents", "Date": "2026-03", "Finding": "Seeq launched Seeq Intelligence, adding Agent Q (a natural-language AI analyst), build-your-own agents for multi-step workflows, secure connections to customer systems and unstructured-document extraction. It is included in the Seeq Enterprise package, and ExxonMobil and Marathon Petroleum have presented at Seeq events.", "Metric": "No quantified metric supplied", "Source_IDs": "R105", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-series analytics incumbents are bundling agents into existing enterprise tiers rather than charging separately, which compresses what standalone agent vendors can charge.", "Partners": "ExxonMobil, Marathon Petroleum, PBF", "Original_date": "2026-03", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-105"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-03-03", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-013"], "Research_search": " {\"key\": \"seeq-intelligence\", \"existing_ids\": [\"AI-128\"], \"related_existing_ids\": [], \"company\": \"Seeq\", \"title\": \"Seeq Intelligence and Agent Q for operational decisions\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched 2026-03-03; available now\", \"practical_relevance\": \"Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.\", \"business_problem\": \"Industrial knowledge and event history are siloed, making it difficult for operations personnel to investigate deviations consistently and scale expert practice.\", \"workflow\": \"Agent Q assembles current process data, historical events, previous analyses, documents and expert context into traceable investigations and prioritized recommendations. Users can build multistep custom agents, scheduled or triggered, and connect agents to external systems for context or action.\", \"technology_data\": \"Seeq uses contextualized real-time operational data, documents and analytic steps; vendor describes secure extensibility and domain-aware AI but does not identify specific LLM or customer infrastructure in release.\", \"integration_human_role\": \"Designed to augment subject-matter experts; traceable output and integration enable review. Closed-loop action is technically supported in integrations but actual controls/customer approvals unspecified.\", \"deployment_scope\": \"Available Seeq Enterprise product; oil & gas among target sectors; no named O&G customer deployment for Agent Q in launch.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No customer-specific value quantified.\", \"timeline\": [{\"date\": \"2026-03-03\", \"event\": \"Seeq Intelligence unveiled and made available.\", \"source_ids\": [\"RS083\"]}], \"limitations\": [\"No named energy customer case, implementation architecture or results in launch disclosure.\"], \"assessment\": \"Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.\", \"sources\": [{\"id\": \"RS083\", \"title\": \"Introducing Seeq Intelligence\", \"url\": \"https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/\", \"publisher\": \"Seeq\", \"date\": \"2026-03-03\", \"type\": \"company announcement\", \"locator\": \"Agent Q, build-your-own agents and integrations\", \"review_status\": \"full text reviewed\", \"supports\": [\"Agent Q synthesizes real-time operational context, prior analyses, historical events, documents and domain expertise into investigations and recommended actions.\", \"Custom agents run multistep workflows on demand, scheduled or triggered; secure extensions can retrieve customer-system context and initiate workflows.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-016\"}], \"field_sources\": {\"business_problem\": [\"RS083\"], \"workflow\": [\"RS083\"], \"technology_data\": [\"RS083\"], \"integration_human_role\": [\"RS083\"], \"deployment_scope\": [\"RS083\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-013\", \"source_ids\": [\"RS083\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-128\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-128", "company": "Seeq", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Seeq Intelligence release establishes product capabilities, not deployment at a named O&G customer.", "correction": "Keep as commercial product launch absent customer evidence.", "source_url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-129", "Company": "AVEVA", "Company_primary": "AVEVA", "Segment": "Software incumbent", "HQ": "United Kingdom", "Theme": "Enterprise AI", "Initiative": "AVEVA industrial AI capabilities and CONNECT roadmap", "Extracted_finding": "Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.", "Context": "Industrial organizations face costly manual data mapping and fragmented OT/IT information that impede governed AI deployment.", "Implementation": "AVEVA announced data mapping assistance, industrial knowledge graph, CONNECT Flows and AI features in engineering/operations; AI can propose data mappings with lineage and governance. HMI/SCADA edge options discussed.", "Assessment": "Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Announced 2026-05-20; some releases are future roadmap", "Interpretation_limit": "No named O&G deployment or verified operating results. A major CONNECT AI release is planned for Q1 2027 and is out of window.", "Source_IDs": "R106", "Source_date": "2026-05-20", "Source_type": "company announcement", "Source_locator": "CONNECT, Flows and AI roadmap", "Source_URL": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2026-05-20; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-106", "Initiative_ID": "AI-129", "Title": "AVEVA embeds agentic AI, MCP across CONNECT", "Date": "2026-05", "Finding": "At AVEVA World 2026, AVEVA announced agentic twin-builder mapping, an AI-enriched industrial knowledge graph, and MCP integrations that let LLMs read operational data. It also announced a CONNECT major release planned for Q1 2027 and deeper Microsoft, Snowflake and ServiceNow integrations. Its Industrial AI Assistant runs on Azure OpenAI.", "Metric": "3,500+ delegates; 5,000 partners", "Source_IDs": "R106", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Industrial historians now expose data through MCP, so AI builders should plan to connect via MCP to PI and CONNECT rather than build custom connectors.", "Partners": "Microsoft Azure OpenAI, Snowflake, ServiceNow, Cognizant, Crosser", "Original_date": "2026-05", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-106"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2026-05-20", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["software_capital-014"], "Research_search": " {\"key\": \"aveva-ai-connect\", \"existing_ids\": [\"AI-129\"], \"related_existing_ids\": [], \"company\": \"AVEVA\", \"title\": \"AVEVA industrial AI capabilities and CONNECT roadmap\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Announced 2026-05-20; some releases are future roadmap\", \"practical_relevance\": \"Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.\", \"business_problem\": \"Industrial organizations face costly manual data mapping and fragmented OT/IT information that impede governed AI deployment.\", \"workflow\": \"AVEVA announced data mapping assistance, industrial knowledge graph, CONNECT Flows and AI features in engineering/operations; AI can propose data mappings with lineage and governance. HMI/SCADA edge options discussed.\", \"technology_data\": \"CONNECT, Twin Builder, knowledge graph and connectors are named. Major CONNECT release with additional AI functionality was roadmap for Q1 2027, beyond the window; 2026 Flows capabilities need release-level qualification.\", \"integration_human_role\": \"Data governance/lineage emphasized; specific review roles and operational authority are not detailed.\", \"deployment_scope\": \"Commercial industrial platform/roadmap; no oil/gas-specific customer deployment identified in release.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No deployment cost or customer economics disclosed.\", \"timeline\": [{\"date\": \"2026-05-20\", \"event\": \"AVEVA announced new AI capabilities at AVEVA World; major CONNECT release planned for Q1 2027.\", \"source_ids\": [\"R106\"]}], \"limitations\": [\"No named O&G deployment or verified operating results.\", \"A major CONNECT AI release is planned for Q1 2027 and is out of window.\"], \"assessment\": \"Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.\", \"sources\": [{\"id\": \"R106\", \"title\": \"AVEVA announces new capabilities to embed AI\", \"url\": \"https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/\", \"publisher\": \"AVEVA\", \"date\": \"2026-05-20\", \"type\": \"company announcement\", \"locator\": \"CONNECT, Flows and AI roadmap\", \"review_status\": \"full text reviewed\", \"supports\": [\"AVEVA announced AI data mapping with lineage/governance, industrial knowledge graph and Flows/connectors.\", \"Press release labels major CONNECT release for Q1 2027 as planned; roadmap is beyond research window and must not be counted as deployed in-window.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-017\"}], \"field_sources\": {\"business_problem\": [\"R106\"], \"workflow\": [\"R106\"], \"technology_data\": [\"R106\"], \"integration_human_role\": [\"R106\"], \"deployment_scope\": [\"R106\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-014\", \"source_ids\": [\"R106\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-129\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Q1 2027 CONNECT major AI release was deployed by end of research window.", "reason": "2026 announcement states this major release is planned for Q1 2027; future roadmap is outside window.", "source_url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/"}], "Audit_review": {"id": "AI-129", "company": "AVEVA", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/"], "findings": [{"severity": "medium", "field": "stage", "issue": "AVEVA source includes future capabilities and planned releases; it is a product roadmap, not proof those functions were shipped/deployed in May 2026.", "correction": "Use future tense for roadmap items and distinguish available integrations from planned agentic features.", "source_url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-130", "Company": "Emerson / AspenTech", "Company_primary": "Emerson", "Segment": "Software incumbent", "HQ": "United States", "Theme": "Investment AI", "Initiative": "Emerson completes AspenTech buyout for industrial AI", "Extracted_finding": "Emerson bought the remaining AspenTech shares at $265 per share in cash, making AspenTech a wholly owned subsidiary. The deal combines automation hardware with optimization, APM and subsurface software aimed at autonomous operations.", "Context": "Emerson bought the remaining AspenTech shares at $265 per share in cash, making AspenTech a wholly owned subsidiary. The deal combines automation hardware with optimization, APM and subsurface software aimed at autonomous operations.", "Implementation": "Disclosed partners / technologies: Emerson, AspenTech. Full supplied disclosure is retained in the timeline below. Detailed implementation has not been independently checked.", "Assessment": "Automation vendors are consolidating industrial software to own the path to autonomy, which leaves independent AI players to become partners or acquisition targets.", "Reported_metric": "Minority stake ~$7.2B; total AspenTech investment ~$17B (reported)", "Stage_at_disclosure": "As described in supplied reference; review pending", "Interpretation_limit": "Imported claim and interpretation from the supplied reference. Source text, date, scope and outcome attribution require independent verification.", "Source_IDs": "R107", "Source_date": "2025-03", "Source_type": "Emerson", "Source_locator": "See source and disclosure timeline", "Source_URL": "https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "Evidence_status": "Imported reference; source verification pending", "Review_tier": "Reference pending review", "Time_scope": "Reference date within window; date not independently verified", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-107", "Initiative_ID": "AI-130", "Title": "Emerson completes AspenTech buyout for industrial AI", "Date": "2025-03", "Finding": "Emerson bought the remaining AspenTech shares at $265 per share in cash, making AspenTech a wholly owned subsidiary. The deal combines automation hardware with optimization, APM and subsurface software aimed at autonomous operations.", "Metric": "Minority stake ~$7.2B; total AspenTech investment ~$17B (reported)", "Source_IDs": "R107", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Automation vendors are consolidating industrial software to own the path to autonomy, which leaves independent AI players to become partners or acquisition targets.", "Partners": "Emerson, AspenTech", "Original_date": "2025-03", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-107"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-03", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_status": "Inherited profile; expansion review incomplete", "Research_packets": [], "Research_category": "Investment or acquisition", "Audit_flags": [{"claim": "Emerson completion of remaining AspenTech acquisition.", "reason": "Verified in Emerson primary release on 2025-03-12; acquisition transaction is ownership context and does not itself establish an AI deployment. Legacy classification as an AI initiative should be treated as transaction context.", "source_url": "https://ir.emerson.com/news-events/press-releases/detail/11/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech"}, {"claim": "Completion of Emerson acquisition of remaining AspenTech shares as an AI implementation.", "reason": "The transaction is verified, but the official acquisition notice establishes ownership and deal completion, not an AI workflow. Keep as corporate transaction context.", "source_url": "https://ir.emerson.com/news-events/press-releases/detail/11/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech"}, {"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-130", "company": "Emerson / AspenTech", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "pending_review", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "No expanded source review recorded", "claim_status": "Imported reference; source verification pending", "independent_validation": "Not established by this collection"}}, {"ID": "AI-131", "Company": "Aker BP", "Company_primary": "Aker BP", "Segment": "Independent E&P", "HQ": "Norway", "Theme": "Enterprise AI", "Initiative": "Copilot rollout, employee agents and Knowledge Agents", "Extracted_finding": "Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.", "Context": "Employees need more efficient knowledge retrieval and routine writing/meeting work while safe operations require governance and support for broad AI adoption.", "Implementation": "Aker BP expanded Copilot and Copilot Studio access, training, help desk and champion network. Knowledge Agents answer repeated document questions and route complex requests to experts; employees and central specialist team build agents.", "Assessment": "Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.", "Reported_metric": "Initial Copilot licenses: 300 (Customer statement in vendor case); Regular active users: 96% over preceding four weeks (Vendor-reported customer metric)", "Stage_at_disclosure": "Microsoft customer story dated 2025-06-23; rollout began with 300 licenses in 2023", "Interpretation_limit": "Microsoft-published customer narrative; 96% metric lacks denominator and independent audit. License expansion plan is not proof all licenses were delivered.", "Source_IDs": "R126", "Source_date": "2025-06-23", "Source_type": "partner case study", "Source_locator": "Employee Copilot rollout, governance, Knowledge Agents", "Source_URL": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Dated source: 2025-06-23; see chronology for operating dates.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "REF-126", "Initiative_ID": "AI-131", "Title": "Aker BP: Copilot for all staff plus champion network", "Date": "2025-06", "Finding": "Aker BP started with 300 Copilot licences and committed to giving every employee one in 2025. It created an AI adoption champion network with accountability at executive-committee level. Training includes help-desk support, AI-fundamentals learning paths and hands-on prompt-engineering sessions, and 'Digital and AI competence' became one of five core competencies for all employees. Staff and small teams are building their own agents, such as knowledge agents.", "Metric": "300 initial licences to all employees; 96% active users over four weeks (Viva dashboard); AI competence is 1 of 5 core competencies", "Source_IDs": "R126", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Putting AI competence into the formal competency model and making executives accountable for champions is what moves usage to 96%.", "Partners": "Microsoft", "Original_date": "2025-06", "Related_IDs": []}], "Related_IDs": [], "Reference_IDs": ["REF-126"], "Provenance": ["User-supplied reference"], "Latest_disclosure": "2025-06-23", "Has_pending_disclosures": true, "Assessment_origin": "Interpretation supplied with reference; not independently endorsed", "Research_packets": ["majors_backfill-011"], "Research_search": " {\"key\": \"akerbp-copilot\", \"existing_ids\": [\"AI-131\"], \"related_existing_ids\": [], \"company\": \"Aker BP / Microsoft\", \"title\": \"Copilot rollout, employee agents and Knowledge Agents\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Microsoft customer story dated 2025-06-23; rollout began with 300 licenses in 2023\", \"practical_relevance\": \"Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.\", \"business_problem\": \"Employees need more efficient knowledge retrieval and routine writing/meeting work while safe operations require governance and support for broad AI adoption.\", \"workflow\": \"Aker BP expanded Copilot and Copilot Studio access, training, help desk and champion network. Knowledge Agents answer repeated document questions and route complex requests to experts; employees and central specialist team build agents.\", \"technology_data\": \"Microsoft 365 Copilot and Copilot Studio; knowledge agents use generative AI, low-code capabilities, topics and human-in-the-loop features; detailed corpus/integration not stated.\", \"integration_human_role\": \"Agents route complex queries to experts; central team supports makers; security ratings and governance precede wider adoption.\", \"deployment_scope\": \"Company-wide rollout planned for 2025; 300 starting licenses. User activity and agent-live counts are from Microsoft customer case.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Initial Copilot licenses\", \"value\": \"300\", \"period\": \"Early deployment; 2023 pilot context\", \"attribution\": \"Customer statement in vendor case\", \"source_ids\": [\"R126\"]}, {\"metric\": \"Regular active users\", \"value\": \"96% over preceding four weeks\", \"period\": \"As described 2025-06-23\", \"attribution\": \"Vendor-reported customer metric\", \"source_ids\": [\"R126\"]}], \"economics\": \"No quantified labor-savings/economic outcome.\", \"timeline\": [{\"date\": \"2025-06-23\", \"event\": \"Copilot rollout, employee agents and Knowledge Agents documented or disclosed.\", \"source_ids\": [\"R126\"]}], \"limitations\": [\"Microsoft-published customer narrative; 96% metric lacks denominator and independent audit.\", \"License expansion plan is not proof all licenses were delivered.\"], \"assessment\": \"Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.\", \"sources\": [{\"id\": \"R126\", \"title\": \"Aker BP builds AI-first future with Microsoft 365 Copilot and AI agents\", \"url\": \"https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot\", \"publisher\": \"Microsoft\", \"date\": \"2025-06-23\", \"type\": \"partner case study\", \"locator\": \"Employee Copilot rollout, governance, Knowledge Agents\", \"review_status\": \"full text reviewed\", \"supports\": [\"Aker BP started with 300 Copilot licenses and planned all-staff licensing in 2025; trained employees, set up help desk and champion/accountability network.\", \"Employee/small-team-built agents, including Knowledge Agents, answer repetitive documentation questions and route harder queries; several agents live.\", \"Microsoft case quotes 96% active users in prior four weeks; vendor-published customer metric, not independent audit.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-014\"}], \"field_sources\": {\"business_problem\": [\"R126\"], \"workflow\": [\"R126\"], \"technology_data\": [\"R126\"], \"integration_human_role\": [\"R126\"], \"deployment_scope\": [\"R126\"], \"economics\": [\"R126\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-011\", \"source_ids\": [\"R126\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-131\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-131", "company": "Aker BP", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "300 Copilot licences", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-132", "Company": "OMV", "Company_primary": "OMV", "Segment": "Integrated", "HQ": "Austria", "Theme": "Operational AI", "Initiative": "Norwegian AI Companion geological and engineering search", "Extracted_finding": "Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.", "Context": "Engineers face large archives of geology and petroleum-engineering material on the Norwegian Continental Shelf.", "Implementation": "Companion synthesizes answers to complex geology/geoscience/petroleum engineering questions from a large curated source corpus.", "Assessment": "Analyst inference: Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.", "Reported_metric": "Corpus: >2.5 million pages (Company-reported)", "Stage_at_disclosure": "Reported in OMV 2025 annual report", "Interpretation_limit": "OMV describes a digital assistant, but details on what counts as actionable response and hallucination controls are absent.", "Source_IDs": "R037", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Digitalization; AI Across the Value Chain", "Source_URL": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "ADD-AI-132", "Initiative_ID": "AI-132", "Title": "Norwegian AI Companion", "Date": "2026-04-02", "Finding": "NAIC answers Norwegian Continental Shelf technical questions using more than 2.5 million pages.", "Metric": "2.5m+ pages", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}], "Related_IDs": ["AI-084"], "Reference_IDs": [], "Provenance": ["Primary-source expansion"], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-009"], "Research_search": " {\"key\": \"omv-norwegian-companion\", \"existing_ids\": [\"AI-132\"], \"company\": \"OMV\", \"title\": \"Norwegian AI Companion geological and engineering search\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Reported in OMV 2025 annual report\", \"practical_relevance\": \"Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.\", \"business_problem\": \"Engineers face large archives of geology and petroleum-engineering material on the Norwegian Continental Shelf.\", \"workflow\": \"Companion synthesizes answers to complex geology/geoscience/petroleum engineering questions from a large curated source corpus.\", \"technology_data\": \"Searches 2.5m pages of public and internal geology, geophysics and reports; precise model and retrieval methods not specified.\", \"integration_human_role\": \"Engineering teams consume concise answers and focus on technical work; validation workflow not described.\", \"deployment_scope\": \"Relevant to Norwegian Continental Shelf teams; user count and rollout stage not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Corpus\", \"value\": \">2.5 million pages\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R037\"]}], \"economics\": \"No output quality or time-saved result stated.\", \"timeline\": [{\"date\": \"2026-04-02\", \"event\": \"2025 annual report describes NAIC.\", \"source_ids\": [\"R037\"]}], \"limitations\": [\"OMV describes a digital assistant, but details on what counts as actionable response and hallucination controls are absent.\"], \"assessment\": \"Analyst inference: Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.\", \"sources\": [{\"id\": \"R037\", \"title\": \"Digitalization — OMV Combined Annual Report 2025\", \"url\": \"https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html\", \"publisher\": \"OMV\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Digitalization; AI Across the Value Chain\", \"review_status\": \"full text reviewed\", \"supports\": [\"OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.\", \"OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.\", \"Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.\", \"iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.\", \"Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-012\"}], \"field_sources\": {\"business_problem\": [\"R037\"], \"workflow\": [\"R037\"], \"technology_data\": [\"R037\"], \"integration_human_role\": [\"R037\"], \"deployment_scope\": [\"R037\"], \"economics\": [\"R037\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-009\", \"source_ids\": [\"R037\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-132\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-132", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-133", "Company": "OMV", "Company_primary": "OMV", "Segment": "Integrated", "HQ": "Austria", "Theme": "Operational AI", "Initiative": "Agent-based low-carbon field-development planning", "Extracted_finding": "OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI.", "Context": "Low-carbon field development needs better decisions under subsurface uncertainty and safety constraints.", "Implementation": "Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "Assessment": "Test whether the selected plan works after execution. Faster scenario analysis and early project-value claims are not collected cash.", "Reported_metric": "OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "Stage_at_disclosure": "CCS planning pilot; early internal results", "Interpretation_limit": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132).", "Source_IDs": "R037; WATCH18PM-OMV", "Source_date": "2026-04-14", "Source_type": "Operator technical account", "Source_locator": "Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker", "Source_URL": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "Evidence_status": "Original dated operator technical account reviewed; early internal planning results", "Review_tier": "Reviewed source", "Time_scope": "Norwegian CCS planning pilot; article published 14 April 2026. Test dates, asset name and project count unknown.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "ADD-AI-133", "Initiative_ID": "AI-133", "Title": "Agent-based field development", "Date": "2026-04-02", "Finding": "OMV is testing agents with Stanford and TerraAI for Norwegian low-carbon field development.", "Metric": "No quantified return disclosed", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}, {"ID": "WATCH18PM-AI-133", "Initiative_ID": "AI-133", "Title": "OMV tests field-development decisions with analytical agents", "Date": "2026-04-14", "Finding": "OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI. Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "Metric": "OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "Source_IDs": "WATCH18PM-OMV", "Source_URL": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "Evidence": "Original OMV article reviewed; early internal claims not independently validated", "Origin": "18 October evening source review", "Assessment": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132)."}], "Related_IDs": ["AI-084"], "Reference_IDs": [], "Provenance": ["Primary-source expansion"], "Latest_disclosure": "2026-04-14", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-012"], "Research_search": "{\"key\": \"omv-field-agent\", \"existing_ids\": [\"AI-133\"], \"company\": \"OMV\", \"title\": \"OMV tests field-development decisions with analytical agents\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"CCS planning pilot; early internal results\", \"practical_relevance\": \"Do better plans still outperform after engineering review and actual field execution?\", \"business_problem\": \"Low-carbon field development needs better decisions under subsurface uncertainty and safety constraints.\", \"workflow\": \"Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.\", \"technology_data\": \"Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.\", \"integration_human_role\": \"Engineers challenge assumptions and validate recommendations before action.\", \"deployment_scope\": \"Norwegian CCS planning pilot; article published 14 April 2026. Test dates, asset name and project count unknown.\", \"partners\": [\"Stanford University\", \"TerraAI\"], \"disclosed_results\": [], \"economics\": \"OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.\", \"timeline\": [{\"date\": \"2026-04-02\", \"event\": \"OMV describes ongoing test partnership in annual report.\", \"source_ids\": [\"R037\"]}, {\"date\": \"2026-04-14\", \"review_on\": \"2026-10-18\", \"event\": \"OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI. OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.\", \"source_ids\": [\"WATCH18PM-OMV\"]}], \"limitations\": [\"No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132).\"], \"assessment\": \"Test whether the selected plan works after execution. Faster scenario analysis and early project-value claims are not collected cash.\", \"sources\": [{\"id\": \"R037\", \"title\": \"Digitalization — OMV Combined Annual Report 2025\", \"url\": \"https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html\", \"publisher\": \"OMV\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Digitalization; AI Across the Value Chain\", \"review_status\": \"full text reviewed\", \"supports\": [\"OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.\", \"OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.\", \"Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.\", \"iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.\", \"Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-012\"}, {\"id\": \"WATCH18PM-OMV\", \"title\": \"How AI is powering better decisions in the race to net zero\", \"url\": \"https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero\", \"date\": \"2026-04-14\", \"publisher\": \"OMV\", \"locator\": \"Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker\", \"excerpt\": \"not generative AI, but analytical, agent-based AI\", \"access\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\", \"review_status\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"R037\", \"WATCH18PM-OMV\"], \"workflow\": [\"R037\", \"WATCH18PM-OMV\"], \"technology_data\": [\"R037\", \"WATCH18PM-OMV\"], \"deployment_scope\": [\"R037\", \"WATCH18PM-OMV\"], \"economics\": [\"R037\", \"WATCH18PM-OMV\"], \"integration_human_role\": [\"R037\", \"WATCH18PM-OMV\"], \"practical_relevance\": [\"R037\", \"WATCH18PM-OMV\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-012\", \"source_ids\": [\"R037\", \"WATCH18PM-OMV\"], \"report_date\": \"2026-10-18\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-133\"], \"earlier_review\": {\"report_date\": \"2026-09-27\", \"assessment\": \"Analyst inference: Early-stage named technical collaboration with preliminary outcomes; claims remain company-reported and unquantified.\", \"limitations\": [\"Early results are qualitative; do not treat as deployed or proven value.\"], \"note\": \"Superseded by targeted 18 October 2026 source check.\"}, \"last_source_check\": \"2026-10-18\"}", "Research_status": "Targeted source detail checked 18 October 2026 evening", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-133", "company": "OMV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [{"severity": "medium", "field": "stage", "issue": "OMV annual report says low-carbon field-development agentic AI work is being tested; no oil-and-gas deployment or quantified outcome.", "correction": "Keep as test-stage adjacent low-carbon planning.", "source_url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"}], "residual_gaps": ["Metric definitions, validation sample, comparator and full costs remain undisclosed."], "partition": 2, "reviewer_outcome": "partially supported", "source_recheck_on": "2026-10-18", "resolution": "The dated April operator article establishes the analytical CCS pilot and early internal planning figures. It does not establish executed field results."}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132).", "Outcome_counted": false, "Outcome_count_note": "Early internal planning metrics; no realized production, collected cash or independent net ROI established.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "CCS planning pilot; early internal results", "independent_validation": "Not established by this collection"}, "Earlier_source_review": {"as_of": "Before 18 October evening source check", "finding": "Early-stage named technical collaboration with preliminary outcomes; claims remain company-reported and unquantified.", "limit": "Early results are qualitative; do not treat as deployed or proven value.", "source_date": "Undated source; see chronology", "note": "Earlier review retained for provenance; current fields below supersede it."}, "Current_source_check": "2026-10-18", "Research_origin": ["research/watch-2026-10-18-pm/README.md"]}, {"ID": "AI-134", "Company": "OMV", "Company_primary": "OMV", "Segment": "Integrated", "HQ": "Austria", "Theme": "Operational AI", "Initiative": "iDEA well-engineering knowledge system", "Extracted_finding": "A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.", "Context": "Drilling design can repeat avoidable problems when prior global well experience is hard to find.", "Implementation": "Knowledge system identifies global analogues and lessons learned during active drilling-program design to flag potential issues and costs.", "Assessment": "Analyst inference: A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.", "Reported_metric": "No quantified results disclosed", "Stage_at_disclosure": "Introduced by 2025 annual-report period", "Interpretation_limit": "System is described as knowledge-based; avoid asserting generative AI or autonomous design.", "Source_IDs": "R037", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Digitalization; AI Across the Value Chain", "Source_URL": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "ADD-AI-134", "Initiative_ID": "AI-134", "Title": "iDEA well engineering", "Date": "2026-04-02", "Finding": "iDEA links drilling program design to global analogues and lessons learned.", "Metric": "No quantified return disclosed", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}], "Related_IDs": ["AI-084"], "Reference_IDs": [], "Provenance": ["Primary-source expansion"], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-010"], "Research_search": " {\"key\": \"omv-iDEA\", \"existing_ids\": [\"AI-134\"], \"company\": \"OMV\", \"title\": \"iDEA well-engineering knowledge system\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Introduced by 2025 annual-report period\", \"practical_relevance\": \"A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.\", \"business_problem\": \"Drilling design can repeat avoidable problems when prior global well experience is hard to find.\", \"workflow\": \"Knowledge system identifies global analogues and lessons learned during active drilling-program design to flag potential issues and costs.\", \"technology_data\": \"Connects current well design to OMV and OMV Petrom global experience; underlying AI/model details not disclosed.\", \"integration_human_role\": \"Drilling engineers use timely insights to inform program; final design accountability remains with engineers.\", \"deployment_scope\": \"OMV and OMV Petrom well-design workflow; exact adoption and well count not reported.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Avoided incidents/cost savings not quantified.\", \"timeline\": [{\"date\": \"2026-04-02\", \"event\": \"2025 annual report describes introduction of iDEA.\", \"source_ids\": [\"R037\"]}], \"limitations\": [\"System is described as knowledge-based; avoid asserting generative AI or autonomous design.\"], \"assessment\": \"Analyst inference: A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.\", \"sources\": [{\"id\": \"R037\", \"title\": \"Digitalization — OMV Combined Annual Report 2025\", \"url\": \"https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html\", \"publisher\": \"OMV\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Digitalization; AI Across the Value Chain\", \"review_status\": \"full text reviewed\", \"supports\": [\"OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.\", \"OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.\", \"Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.\", \"iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.\", \"Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-012\"}], \"field_sources\": {\"business_problem\": [\"R037\"], \"workflow\": [\"R037\"], \"technology_data\": [\"R037\"], \"integration_human_role\": [\"R037\"], \"deployment_scope\": [\"R037\"], \"economics\": [\"R037\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-010\", \"source_ids\": [\"R037\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-134\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "The reviewed annual report gives use for drilling programs, analogues and lessons but no technical description, deployment count or quantified result.", "source_url": ""}], "Audit_review": {"id": "AI-134", "company": "OMV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [{"severity": "low", "field": "technology_data", "issue": "OMV’s report places iDEA among examples illustrating AI across the value chain, but calls the tool a knowledge-based system and gives no model details.", "correction": "Retain as issuer-framed AI use; state that the AI technique is undisclosed rather than declassifying for lack of architecture detail.", "source_url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"}], "residual_gaps": ["The reviewed annual report gives use for drilling programs, analogues and lessons but no technical description, deployment count or quantified result."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-135", "Company": "OMV", "Company_primary": "OMV", "Segment": "Integrated", "HQ": "Austria", "Theme": "Operational AI", "Initiative": "Predictive pump-failure analysis in Romania", "Extracted_finding": "Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.", "Context": "Critical well failures and process interventions can interrupt production; proactive planning can reduce downtime.", "Implementation": "Model analyzes historical pump-operation and event data to predict potential pump failures up to 90 days ahead.", "Assessment": "Analyst inference: Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.", "Reported_metric": "Forecast horizon: up to 90 days (Company-reported)", "Stage_at_disclosure": "Machine-learning pilot reported in 2025 annual report", "Interpretation_limit": "Pilot only; predictive horizon is not accuracy.", "Source_IDs": "R037", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Digitalization; AI Across the Value Chain", "Source_URL": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "Evidence_status": "verified primary", "Review_tier": "Reviewed source", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Coverage_bucket": "Consolidated reference addition", "PE_relationship": null, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for ambition shortlist", "Ambition_assessment_basis": "Breadth, autonomy, integration and scale assessed separately from evidence maturity.", "Scope_type": "Program or workflow reported in supplied reference", "Disclosures": [{"ID": "ADD-AI-135", "Initiative_ID": "AI-135", "Title": "Pump-failure prediction", "Date": "2026-04-02", "Finding": "A Romanian ML pilot analyzes pump histories to anticipate well failures.", "Metric": "Up to 90 days advance warning", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}], "Related_IDs": ["AI-084"], "Reference_IDs": [], "Provenance": ["Primary-source expansion"], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-011"], "Research_search": " {\"key\": \"omv-pump-ml\", \"existing_ids\": [\"AI-135\"], \"company\": \"OMV\", \"title\": \"Predictive pump-failure analysis in Romania\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Machine-learning pilot reported in 2025 annual report\", \"practical_relevance\": \"Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.\", \"business_problem\": \"Critical well failures and process interventions can interrupt production; proactive planning can reduce downtime.\", \"workflow\": \"Model analyzes historical pump-operation and event data to predict potential pump failures up to 90 days ahead.\", \"technology_data\": \"Machine-learning failure prediction; specific model, test performance and deployment integration not disclosed.\", \"integration_human_role\": \"Supports proactive maintenance planning; human maintenance intervention implied.\", \"deployment_scope\": \"Romania pilot; critical well failures; no number of wells/pumps given.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Forecast horizon\", \"value\": \"up to 90 days\", \"period\": \"Romania pilot\", \"attribution\": \"Company-reported\", \"source_ids\": [\"R037\"]}], \"economics\": \"Downtime avoided or prediction precision not reported.\", \"timeline\": [{\"date\": \"2026-04-02\", \"event\": \"OMV reports Romanian ML pilot and 90-day forecast horizon.\", \"source_ids\": [\"R037\"]}], \"limitations\": [\"Pilot only; predictive horizon is not accuracy.\"], \"assessment\": \"Analyst inference: Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.\", \"sources\": [{\"id\": \"R037\", \"title\": \"Digitalization — OMV Combined Annual Report 2025\", \"url\": \"https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html\", \"publisher\": \"OMV\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Digitalization; AI Across the Value Chain\", \"review_status\": \"full text reviewed\", \"supports\": [\"OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.\", \"OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.\", \"Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.\", \"iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.\", \"Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-012\"}], \"field_sources\": {\"business_problem\": [\"R037\"], \"workflow\": [\"R037\"], \"technology_data\": [\"R037\"], \"integration_human_role\": [\"R037\"], \"deployment_scope\": [\"R037\"], \"economics\": [\"R037\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-011\", \"source_ids\": [\"R037\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-135\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-135", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-136", "Company": "bp", "Company_primary": "bp", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Real-time driver hazard alerts in Permian operations", "Extracted_finding": "A frontline safety workflow with a reported operating-period signal: bp says the AI alerts reduced two driver-risk events in bpx Permian operations, while not disclosing fleet denominators or independent validation.", "Stage_at_disclosure": "Implemented across bpx Permian operations in 2024", "Reported_metric": "Driver distraction events: 89% reduction (Company-reported); Tailgating events: 75% reduction (Company-reported)", "Interpretation_limit": "Direct report PDF fetch returned 403; substantive passage was reviewed through the official indexed report excerpt. bp provides percentage changes without absolute event counts, fleet denominator, or independent validation. Do not generalize beyond bpx Permian operations or infer collision/injury reduction.", "Source_IDs": "RS002", "Source_date": "Undated source; see chronology", "Source_type": "sustainability report", "Source_locator": "bp Sustainability Report 2024, p. 9 (archived copy)", "Source_URL": "https://web.archive.org/web/20260201003758/https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Driving hazards create collision and injury exposure for onshore field staff.", "Implementation": "An AI system assesses driving hazards and risks in real time and provides in-vehicle alerts intended to prompt preventative action.", "Assessment": "Analyst inference: in-vehicle alerts can place risk signals inside a driver’s response loop; the company-reported event reductions suggest a measurable operational feedback loop, but denominator and validation limits prevent estimating fleet-wide safety impact.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors-002-1", "Initiative_ID": "AI-136", "Title": "Real-time driver hazard alerts in Permian operations", "Date": "2024", "Finding": "bpx implemented the AI technology across Permian operations.", "Metric": "", "Source_IDs": "RS002", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-002"], "Research_search": " {\"key\": \"bp-bpx-driving-safety\", \"existing_ids\": [], \"company\": \"bp\", \"title\": \"Real-time driver hazard alerts in Permian operations\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Implemented across bpx Permian operations in 2024\", \"practical_relevance\": \"A frontline safety workflow with a reported operating-period signal: bp says the AI alerts reduced two driver-risk events in bpx Permian operations, while not disclosing fleet denominators or independent validation.\", \"business_problem\": \"Driving hazards create collision and injury exposure for onshore field staff.\", \"workflow\": \"An AI system assesses driving hazards and risks in real time and provides in-vehicle alerts intended to prompt preventative action.\", \"technology_data\": \"bp does not identify model, sensor inputs, vendor, coverage, or alert-performance metrics.\", \"integration_human_role\": \"Drivers receive alerts and must respond; precise escalation and supervision procedures are not disclosed.\", \"deployment_scope\": \"bpx-operated Permian Basin operations; extent by vehicle or fleet not stated.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Driver distraction events\", \"value\": \"89% reduction\", \"period\": \"December 2023 to December 2024\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS002\"]}, {\"metric\": \"Tailgating events\", \"value\": \"75% reduction\", \"period\": \"December 2023 to December 2024\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS002\"]}], \"economics\": \"No financial economics disclosed. bp reports reductions in two driving-risk event types; this is not a quantified cost, injury, or collision reduction.\", \"timeline\": [{\"date\": \"2024\", \"event\": \"bpx implemented the AI technology across Permian operations.\", \"source_ids\": [\"RS002\"]}], \"limitations\": [\"Direct report PDF fetch returned 403; substantive passage was reviewed through the official indexed report excerpt.\", \"bp provides percentage changes without absolute event counts, fleet denominator, or independent validation.\", \"Do not generalize beyond bpx Permian operations or infer collision/injury reduction.\"], \"assessment\": \"Analyst inference: in-vehicle alerts can place risk signals inside a driver’s response loop; the company-reported event reductions suggest a measurable operational feedback loop, but denominator and validation limits prevent estimating fleet-wide safety impact.\", \"sources\": [{\"id\": \"RS002\", \"title\": \"bp Sustainability Report 2024\", \"url\": \"https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf\", \"publisher\": \"bp\", \"date\": null, \"type\": \"annual report\", \"locator\": \"Advancing safety through AI, bpx energy\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"bp states bpx implemented AI technology across Permian operations in 2024 to assess driving hazards and risks in real time and provide in-vehicle alerts.\", \"The indexed official report passage attributes 89% fewer driver-distraction events and 75% fewer tailgating events between Dec 2023 and Dec 2024 to the technology; direct PDF retrieval returned 403.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"RS002\"}], \"field_sources\": {\"business_problem\": [\"RS002\"], \"workflow\": [\"RS002\"], \"technology_data\": [\"RS002\"], \"integration_human_role\": [\"RS002\"], \"deployment_scope\": [\"RS002\"], \"economics\": [\"RS002\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-002\", \"source_ids\": [\"RS002\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-136\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-136", "company": "bp", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf"], "findings": [{"severity": "medium", "field": "disclosed_results", "issue": "Indexed official report text adds company-reported results: driver-distraction events down 89% and tailgating events down 75% between Dec 2023 and Dec 2024. Direct PDF fetch returned 403; evidence is excerpt-level.", "correction": "Add both metrics with the stated Dec 2023–Dec 2024 comparison period and company attribution; do not characterize as independently validated or as incident-rate reductions.", "source_url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf"}], "residual_gaps": ["Report does not disclose fleet denominator, sensor/model details, absolute event counts, or independent validation; PDF direct fetch remained blocked, so the report passage was reviewed via official search indexing only."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "bp's original PDF URL now 301-redirects to the reporting-centre archive page; Source_URL points to the Wayback Machine copy captured 1 Feb 2026, re-read 5 Oct 2026 (p. 9: distraction events -89%, tailgating -75%, Dec 2023–Dec 2024).", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-137", "Company": "Eni", "Company_primary": "Eni", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "Open access to supercomputing for external AI innovation", "Extracted_finding": "Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.", "Stage_at_disclosure": "HPC6/HPC7 access program announced July 2026", "Reported_metric": "Combined compute: Exascale-class performance (Company-reported)", "Interpretation_limit": "Kept separate from data-center supply/infrastructure activity; disclosure concerns research compute made available to third parties.", "Source_IDs": "RS008", "Source_date": "2026-07-28", "Source_type": "company announcement", "Source_locator": "Research and Development", "Source_URL": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-07-28; see chronology for operating dates.", "Context": "Startups, companies and research organizations may lack compute and specialized expertise needed to develop AI solutions.", "Implementation": "Eni makes HPC capacity, expertise and tools available to external firms, startups and research centers developing AI and digital innovation.", "Assessment": "Analyst inference: Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.", "Primary_department": "Research & innovation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "AI infrastructure & enablement", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors-013-1", "Initiative_ID": "AI-137", "Title": "Open access to supercomputing for external AI innovation", "Date": "2026-07-28", "Finding": "Eni announces access program for HPC infrastructure.", "Metric": "", "Source_IDs": "RS008", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-07-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-013"], "Research_search": " {\"key\": \"eni-open-hpc-ai\", \"existing_ids\": [], \"company\": \"Eni\", \"title\": \"Open access to supercomputing for external AI innovation\", \"department\": \"Research & innovation\", \"secondary_departments\": [], \"category\": \"AI infrastructure & enablement\", \"stage\": \"HPC6/HPC7 access program announced July 2026\", \"practical_relevance\": \"Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.\", \"business_problem\": \"Startups, companies and research organizations may lack compute and specialized expertise needed to develop AI solutions.\", \"workflow\": \"Eni makes HPC capacity, expertise and tools available to external firms, startups and research centers developing AI and digital innovation.\", \"technology_data\": \"Green Data Center hosts HPC6/HPC7; combined performance described by Eni as above exascale-class. The infrastructure supports broad HPC/AI experimentation, not a specific oilfield AI workflow.\", \"integration_human_role\": \"External researchers/companies access Eni tools and expertise; project selection/terms are not in reviewed announcement.\", \"deployment_scope\": \"Innovation ecosystem access; specific program scale, users and compute allocation not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Combined compute\", \"value\": \"Exascale-class performance\", \"period\": \"July 2026\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS008\"]}], \"economics\": \"No commercial returns or AI project outcomes quantified.\", \"timeline\": [{\"date\": \"2026-07-28\", \"event\": \"Eni announces access program for HPC infrastructure.\", \"source_ids\": [\"RS008\"]}], \"limitations\": [\"Kept separate from data-center supply/infrastructure activity; disclosure concerns research compute made available to third parties.\"], \"assessment\": \"Analyst inference: Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.\", \"sources\": [{\"id\": \"RS008\", \"title\": \"Eni opens its supercomputing capabilities to the innovation ecosystem\", \"url\": \"https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html\", \"publisher\": \"Eni\", \"date\": \"2026-07-28\", \"type\": \"company announcement\", \"locator\": \"Research and Development\", \"review_status\": \"full text reviewed\", \"supports\": [\"Eni opened HPC infrastructure, expertise and tools to companies, startups and research centres to develop AI and digital innovation; HPC6 and HPC7 together exceed exascale-class performance.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-013\"}], \"field_sources\": {\"business_problem\": [\"RS008\"], \"workflow\": [\"RS008\"], \"technology_data\": [\"RS008\"], \"integration_human_role\": [\"RS008\"], \"deployment_scope\": [\"RS008\"], \"economics\": [\"RS008\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"AI-027 is an aggregate portfolio record; this is a named 2026 infrastructure initiative and not evidence of a particular oil/gas workflow.\", \"related_existing_ids\": [\"AI-027\"], \"cohort\": \"majors\", \"packet_id\": \"majors-013\", \"source_ids\": [\"RS008\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-137\"]}", "Research_status": "Expanded source review", "Research_category": "AI infrastructure & enablement", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-137", "company": "Eni", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "infrastructure", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-138", "Company": "Aker BP", "Company_primary": "Aker BP", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI agents for Yggdrasil remote operations", "Extracted_finding": "A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.", "Stage_at_disclosure": "Planned from day one of Yggdrasil operations in 2027; disclosed 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company claims operational agents already reduce troubleshooting time, but no example or quantified result was found in reviewed material.", "Source_IDs": "RS010; RS011", "Source_date": "2026-02-11", "Source_type": "company announcement", "Source_locator": "Yggdrasil remote operations", "Source_URL": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-02-11; see chronology for operating dates.", "Context": "Remote/low-manned operations need rapid condition monitoring, maintenance planning and troubleshooting with safe operations.", "Implementation": "AI agents planned for condition-based maintenance and planning at launch, with scaling to more operational areas as solutions mature; data to be shared with onshore control room and external experts.", "Assessment": "Analyst inference: A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors-015-1", "Initiative_ID": "AI-138", "Title": "AI agents for Yggdrasil remote operations", "Date": "2026-04-29", "Finding": "Aker BP described day-one AI agent plans for Yggdrasil.", "Metric": "", "Source_IDs": "RS010", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "majors-015-2", "Initiative_ID": "AI-138", "Title": "AI agents for Yggdrasil remote operations", "Date": "2026-02-11", "Finding": "CEO said AI agents reduce operational troubleshooting time and improve uptime; company-wide capability claim.", "Metric": "", "Source_IDs": "RS011", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-02-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-015"], "Research_search": " {\"key\": \"aker-yggdrasil-agents\", \"existing_ids\": [], \"company\": \"Aker BP\", \"title\": \"AI agents for Yggdrasil remote operations\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Planned from day one of Yggdrasil operations in 2027; disclosed 2026\", \"practical_relevance\": \"A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.\", \"business_problem\": \"Remote/low-manned operations need rapid condition monitoring, maintenance planning and troubleshooting with safe operations.\", \"workflow\": \"AI agents planned for condition-based maintenance and planning at launch, with scaling to more operational areas as solutions mature; data to be shared with onshore control room and external experts.\", \"technology_data\": \"Cognite Data Fusion structures thousands of sensors and plant data; detailed agent design and safeguards not disclosed.\", \"integration_human_role\": \"Agents work alongside control-room technicians; people decide on operations and gradual scope increase is planned.\", \"deployment_scope\": \"Yggdrasil remote onshore operations centre planned from 2027.\", \"partners\": [\"Cognite\"], \"disclosed_results\": [], \"economics\": \"No measured uptime or cost benefit disclosed.\", \"timeline\": [{\"date\": \"2026-04-29\", \"event\": \"Aker BP described day-one AI agent plans for Yggdrasil.\", \"source_ids\": [\"RS010\"]}, {\"date\": \"2026-02-11\", \"event\": \"CEO said AI agents reduce operational troubleshooting time and improve uptime; company-wide capability claim.\", \"source_ids\": [\"RS011\"]}], \"limitations\": [\"Company claims operational agents already reduce troubleshooting time, but no example or quantified result was found in reviewed material.\"], \"assessment\": \"Analyst inference: A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.\", \"sources\": [{\"id\": \"RS010\", \"title\": \"Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation\", \"url\": \"https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/\", \"publisher\": \"Aker BP\", \"date\": null, \"type\": \"company announcement\", \"locator\": \"Yggdrasil remote operations\", \"review_status\": \"full text reviewed\", \"supports\": [\"Yggdrasil operations from onshore centre planned from 2027; AI agents planned from day one for condition-based maintenance and planning, with scaling to other workflows as mature.\", \"Aker BP says industrial data platform Cognite Data Fusion structures thousands of sensor streams and data is to be shared with people and AI agents.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-016\"}, {\"id\": \"RS011\", \"title\": \"Aker BP Q4 2025 transcript\", \"url\": \"https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf\", \"publisher\": \"Aker BP\", \"date\": \"2026-02-11\", \"type\": \"earnings call\", \"locator\": \"PDF pp. 5–6\", \"review_status\": \"full text reviewed\", \"supports\": [\"CEO describes AI-enabled earlier exploration decisions and AI agents reducing operational troubleshooting time and improving uptime.\", \"The CEO describes high-quality structured real-time data connection across the company; 10 years of digital investment is contextual foundation.\", \"Microsoft speaker (partner attribution) says Copilot adoption nearly 100%; not independently audited.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"majors-015\"}], \"field_sources\": {\"business_problem\": [\"RS010\"], \"workflow\": [\"RS011\", \"RS010\"], \"technology_data\": [\"RS010\"], \"integration_human_role\": [\"RS010\"], \"deployment_scope\": [\"RS010\"], \"economics\": [\"RS011\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [\"AI-131\"], \"cohort\": \"majors\", \"packet_id\": \"majors-015\", \"source_ids\": [\"RS010\", \"RS011\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-138\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-138", "company": "Aker BP", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-139", "Company": "Santos", "Company_primary": "Santos", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "Production AI, virtual pressure measurement and drilling controls", "Extracted_finding": "Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.", "Stage_at_disclosure": "Company reports production-AI benefits in 2025; pressure-model savings are potential; deployment population undisclosed.", "Reported_metric": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "Interpretation_limit": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology.", "Source_IDs": "RS012", "Source_date": "2026-05-26", "Source_type": "investor presentation", "Source_locator": "Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross", "Source_URL": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "Evidence_status": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-05; see chronology for operating dates.", "Context": "Santos targets production uplift and cost reduction across drilling, completions, projects and facilities.", "Implementation": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "Assessment": "Virtual pressure measurement could avoid spending on physical instrumentation. That is a different value mechanism from production optimization. The potential benefit still depends on where engineers accept the model as a replacement and what it costs to keep it reliable.", "Primary_department": "Production & field operations", "Secondary_departments": ["Reservoir engineering & field development", "Drilling & completions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Strategy only", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors-016-1", "Initiative_ID": "AI-139", "Title": "Digital and AI across integrated value chain", "Date": "2026-05-20", "Finding": "Investor briefing described technology program across value chain.", "Metric": "", "Source_IDs": "RS012", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "OCT14B-SANTOS-72", "Initiative_ID": "AI-139", "Title": "Production AI, virtual pressure measurement and drilling controls", "Date": "2026-05-26", "Finding": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "Metric": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "Source_IDs": "RS012", "Source_URL": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "Evidence": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "Origin": "14 October 2026 original slide review", "Assessment": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology."}], "Latest_disclosure": "2026-05-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-016"], "Research_search": "{\"key\": \"santos-ai-program\", \"existing_ids\": [], \"company\": \"Santos\", \"title\": \"Production AI, virtual pressure measurement and drilling controls\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company reports production-AI benefits in 2025; pressure-model savings are potential; deployment population undisclosed.\", \"practical_relevance\": \"Virtual pressure measurement could avoid spending on physical instrumentation. That is a different value mechanism from production optimization. The potential benefit still depends on where engineers accept the model as a replacement and what it costs to keep it reliable.\", \"business_problem\": \"Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.\", \"workflow\": \"The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.\", \"technology_data\": \"The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.\", \"integration_human_role\": \"Drilling controls are monitored from an operating center; pressure-model acceptance and oversight are not described.\", \"deployment_scope\": \"26 May 2026 investor briefing; the deck reports 2025 production-AI benefits and 2029 potential. The stimulation column compares 2026 productivity with a 2023 baseline. No well/rig population or model-specific measurement period is supplied.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Column-specific achieved benefits and potential\", \"value\": \"Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.\", \"period\": \"26 May 2026 investor briefing; the deck reports 2025 production-AI benefits and 2029 potential. The stimulation column compares 2026 productivity with a 2023 baseline. No well/rig population or model-specific measurement period is supplied.\", \"attribution\": \"Santos; not independently validated; excluded from numerical counts pending scope reconciliation\", \"source_ids\": [\"RS012\"]}], \"economics\": \"Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.\", \"timeline\": [{\"date\": \"2026-05-26\", \"event\": \"Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.\", \"source_ids\": [\"RS012\"]}], \"limitations\": [\"The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology.\"], \"assessment\": \"Virtual pressure measurement could avoid spending on physical instrumentation. That is a different value mechanism from production optimization. The potential benefit still depends on where engineers accept the model as a replacement and what it costs to keep it reliable.\", \"sources\": [{\"id\": \"RS012\", \"title\": \"Santos 2026 Investor Briefing Day\", \"url\": \"https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf\", \"publisher\": \"Santos\", \"date\": \"2026-05-26\", \"type\": \"investor presentation\", \"locator\": \"Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross\", \"review_status\": \"Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.\", \"supports\": [\"Santos says technology program applies digital and AI capability across integrated value chain, including cost reduction and production uplift across drilling, completions, projects and facilities; specific quantified AI-only impacts not isolated.\"], \"accessed\": \"2026-10-14\", \"original_id\": \"majors-017\"}], \"field_sources\": {\"business_problem\": [\"RS012\"], \"workflow\": [\"RS012\"], \"technology_data\": [\"RS012\"], \"integration_human_role\": [\"RS012\"], \"deployment_scope\": [\"RS012\"], \"economics\": [\"RS012\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"AI-083 is secondary report on Xecta and should not be certified from this briefing excerpt; Xecta source verification remains outstanding.\", \"related_existing_ids\": [\"AI-083\"], \"cohort\": \"majors\", \"packet_id\": \"majors-016\", \"source_ids\": [\"RS012\"], \"report_date\": \"2026-10-14\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-139\"]}", "Research_status": "Expanded 14 October 2026 · original slide reviewed", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-139", "company": "Santos", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "Source is explicitly an AI/digital program across Santos’s integrated value chain, with no one segment identified as primary. Upstream classification would overstate workflow specificity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "Workflow population undisclosed", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology.", "Outcome_counted": false, "Outcome_count_note": "Production-AI benefit is program-level and may overlap AI-083/Xecta; excluded pending scope reconciliation. Potential pressure-model capex reduction is not an achieved result.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original PDF text and rendered slide reviewed", "claim_status": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "independent_validation": "Not established"}, "Current_source_check": "2026-10-14"}, {"ID": "AI-140", "Company": "Shell", "Company_primary": "Shell", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Real-time machine vision and equipment health AI at Pearl GTL", "Extracted_finding": "Near-real-time identification of unsafe conditions and equipment issues supports earlier action and reduces manual inspection needs.", "Stage_at_disclosure": "Deployed by September 2025", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Shell does not report model accuracy, alert volume, incidents avoided or maintenance savings.", "Source_IDs": "RS014; OCT5-S009", "Source_date": "Undated source; see chronology", "Source_type": "company announcement", "Source_locator": "Being smart about safety", "Source_URL": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Large plant work and critical machinery create safety risks and inspection burden; timely signals can allow intervention before incidents or equipment problems escalate.", "Implementation": "Computer vision identifies unsafe behavior and hazardous conditions and alerts supervisors; a separate AI capability provides near-real-time equipment-health insights for field intervention.", "Assessment": "Analyst inference: specific operating use of AI in safety and asset reliability is verified, although outcome evidence is limited.", "Primary_department": "HSE & emissions", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors-019-1", "Initiative_ID": "AI-140", "Title": "Real-time machine vision and equipment health AI at Pearl GTL", "Date": "2025-09", "Finding": "Pearl GTL recognized as WEF Global Lighthouse; Shell describes deployed safety and equipment-health AI.", "Metric": "", "Source_IDs": "RS014", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "OCT5-AI-140-1", "Initiative_ID": "AI-140", "Title": "Pearl GTL machine-vision hazard alerts and AI equipment health", "Date": "2026-05-13", "Finding": "Shell's FY2025 annual report says Pearl GTL uses machine-vision systems to detect hazards in real time and alert supervisors, while AI also monitors equipment health. It adds that the World Economic Forum's Global Lighthouse Network recognised this in September 2025.", "Metric": "", "Source_IDs": "OCT5-S009", "Source_URL": "https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section/promo.multi.stream/1779352356739/36306d968747b6079d3f800b2b1552a033856b5d/shell-integrated-annual-and-sustainability-report.pdf", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review"}], "Latest_disclosure": "2026-05-13", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors-019"], "Research_search": " {\"key\": \"shell-pearl-machine-vision\", \"existing_ids\": [], \"company\": \"Shell\", \"title\": \"Real-time machine vision and equipment health AI at Pearl GTL\", \"department\": \"HSE & emissions\", \"secondary_departments\": [\"Maintenance & reliability\"], \"category\": \"Internal adoption\", \"stage\": \"Deployed by September 2025\", \"practical_relevance\": \"Near-real-time identification of unsafe conditions and equipment issues supports earlier action and reduces manual inspection needs.\", \"business_problem\": \"Large plant work and critical machinery create safety risks and inspection burden; timely signals can allow intervention before incidents or equipment problems escalate.\", \"workflow\": \"Computer vision identifies unsafe behavior and hazardous conditions and alerts supervisors; a separate AI capability provides near-real-time equipment-health insights for field intervention.\", \"technology_data\": \"Machine vision and AI; video/imagery inputs or sensor types and model details are not disclosed.\", \"integration_human_role\": \"Supervisors receive alerts; operations teams act on equipment insights.\", \"deployment_scope\": \"Pearl GTL plant in Qatar; Shell says the site was recognized by WEF Global Lighthouse Network in Sep 2025.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific safety or maintenance outcomes quantified.\", \"timeline\": [{\"date\": \"2025-09\", \"event\": \"Pearl GTL recognized as WEF Global Lighthouse; Shell describes deployed safety and equipment-health AI.\", \"source_ids\": [\"RS014\"]}], \"limitations\": [\"Shell does not report model accuracy, alert volume, incidents avoided or maintenance savings.\"], \"assessment\": \"Analyst inference: specific operating use of AI in safety and asset reliability is verified, although outcome evidence is limited.\", \"sources\": [{\"id\": \"RS014\", \"title\": \"Safety at Pearl: Every second counts, every role matters\", \"url\": \"https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html\", \"publisher\": \"Shell\", \"date\": null, \"accessed\": \"2026-09-27\", \"type\": \"company announcement\", \"locator\": \"Being smart about safety\", \"review_status\": \"full text reviewed\", \"supports\": [\"Shell says Pearl GTL uses machine vision to flag unsafe behaviors and hazardous conditions in real time and alert supervisors.\", \"Pearl uses AI for near-real-time equipment health insights to reduce manual inspections and support earlier intervention.\", \"The plant was recognized by the World Economic Forum Global Lighthouse Network in September 2025.\"], \"original_id\": \"majors-019\"}], \"field_sources\": {\"business_problem\": [\"RS014\"], \"workflow\": [\"RS014\"], \"technology_data\": [\"RS014\"], \"integration_human_role\": [\"RS014\"], \"deployment_scope\": [\"RS014\"], \"economics\": [\"RS014\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"majors\", \"packet_id\": \"majors-019\", \"source_ids\": [\"RS014\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-140\"]}", "Research_status": "Expanded source review · Supplemented 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-140", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-141", "Company": "Aramco", "Company_primary": "Aramco", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Reinforcement-learning control for acid gas removal at Fadhili", "Extracted_finding": "This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.", "Stage_at_disclosure": "Permanent adoption at Fadhili after pilot; additional facilities assessed for suitability", "Reported_metric": "amine and steam use reduction: 10%-15% (Company-reported); power use reduction: around 5% (Company-reported)", "Interpretation_limit": "Metric attribution is company-reported and evaluation details are not disclosed. The company source page is undated; accessed 2026-09-27. The specific month permanent adoption occurred is not disclosed.", "Source_IDs": "RS015", "Source_date": "Undated; accessed 2026-09-27", "Source_type": "company announcement", "Source_locator": "Fadhili Gas Plant section", "Source_URL": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Acid-gas removal performance affects amine and steam consumption, power, stability and operator workload.", "Implementation": "The AI system monitors the existing AGR process and makes ongoing adjustments to improve sour-to-sweet gas conversion. The pilot was conducted Oct 2024-Apr 2025, followed by permanent adoption.", "Assessment": "Analyst inference: This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-001-1", "Initiative_ID": "AI-141", "Title": "Reinforcement-learning control for acid gas removal at Fadhili", "Date": "2024-10 to 2025-04", "Finding": "Fadhili pilot period", "Metric": "", "Source_IDs": "RS015", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "nocs-001-2", "Initiative_ID": "AI-141", "Title": "Reinforcement-learning control for acid gas removal at Fadhili", "Date": "After pilot; exact date undisclosed", "Finding": "Company says technology permanently adopted at Fadhili; other sites assessed", "Metric": "", "Source_IDs": "RS015", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-001"], "Research_search": " {\"key\": \"aramco-fadhili-agr\", \"existing_ids\": [], \"company\": \"Aramco\", \"title\": \"Reinforcement-learning control for acid gas removal at Fadhili\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Permanent adoption at Fadhili after pilot; additional facilities assessed for suitability\", \"practical_relevance\": \"This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.\", \"business_problem\": \"Acid-gas removal performance affects amine and steam consumption, power, stability and operator workload.\", \"workflow\": \"The AI system monitors the existing AGR process and makes ongoing adjustments to improve sour-to-sweet gas conversion. The pilot was conducted Oct 2024-Apr 2025, followed by permanent adoption.\", \"technology_data\": \"Yokogawa-developed reinforcement-learning AI; data sources, controls architecture and operating envelope are not disclosed.\", \"integration_human_role\": \"Operators remain responsible for plant operations; Aramco reports fewer manual interventions. Human approval/safety interlocks are not described.\", \"deployment_scope\": \"Fadhili Gas Plant; further Aramco sites under assessment.\", \"partners\": [\"Yokogawa Middle East & Africa\", \"Yokogawa Digital\"], \"disclosed_results\": [{\"metric\": \"amine and steam use reduction\", \"value\": \"10%-15%\", \"period\": \"Fadhili pilot, Oct 2024-Apr 2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS015\"]}, {\"metric\": \"power use reduction\", \"value\": \"around 5%\", \"period\": \"Fadhili pilot, Oct 2024-Apr 2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS015\"]}], \"economics\": \"Company reports 10-15% reduction in amine and steam use, ~5% lower power use, improved stability and fewer manual interventions. Quantified dollar savings/emissions are not disclosed.\", \"timeline\": [{\"date\": \"2024-10 to 2025-04\", \"event\": \"Fadhili pilot period\", \"source_ids\": [\"RS015\"]}, {\"date\": \"After pilot; exact date undisclosed\", \"event\": \"Company says technology permanently adopted at Fadhili; other sites assessed\", \"source_ids\": [\"RS015\"]}], \"limitations\": [\"Metric attribution is company-reported and evaluation details are not disclosed.\", \"The company source page is undated; accessed 2026-09-27. The specific month permanent adoption occurred is not disclosed.\"], \"assessment\": \"Analyst inference: This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.\", \"sources\": [{\"id\": \"RS015\", \"title\": \"Aramco AI and Big Data\", \"url\": \"https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data\", \"publisher\": \"Aramco\", \"date\": \"Undated; accessed 2026-09-27\", \"type\": \"company announcement\", \"locator\": \"Fadhili Gas Plant section\", \"review_status\": \"full text reviewed\", \"supports\": [\"RL pilot for AGR optimization ran Oct 2024-Apr 2025; Aramco says permanent adoption followed.\", \"Company reports 10-15% lower amine and steam use, around 5% lower power use, greater stability and fewer manual operator interventions.\"], \"accessed_date\": \"2026-09-27\", \"original_id\": \"aramco_fadhili\"}], \"field_sources\": {\"business_problem\": [\"RS015\"], \"workflow\": [\"RS015\"], \"technology_data\": [\"RS015\"], \"integration_human_role\": [\"RS015\"], \"deployment_scope\": [\"RS015\"], \"economics\": [\"RS015\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [\"AI-038\"], \"cohort\": \"nocs\", \"packet_id\": \"nocs-001\", \"source_ids\": [\"RS015\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-141\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-141", "company": "Aramco", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data"], "findings": [{"severity": "medium", "field": "dates", "issue": "Aramco AI and Big Data page is undated; 2026-09-27 is access date only.", "correction": "Use “Undated; accessed 2026-09-27”; keep pilot timing to separately stated source dates.", "source_url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-142", "Company": "PETRONAS", "Company_primary": "PETRONAS", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "ERMAI subsurface opportunity identification", "Extracted_finding": "A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.", "Stage_at_disclosure": "In-house AI solution; value uplift reported for 2025", "Reported_metric": "subsurface analysis time: ~6 days to 6 seconds (Company-reported)", "Interpretation_limit": "Source is excerpt review of the annual report; details of the metric baseline and the value-uplift calculation are not disclosed.", "Source_IDs": "RS017", "Source_date": "2026-02-27", "Source_type": "annual report", "Source_locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)", "Source_URL": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-02-27; see chronology for operating dates.", "Context": "Find bypassed hydrocarbons and accelerate subsurface analysis.", "Implementation": "ERMAI analyzes subsurface data for behind-casing opportunities and real-time petrophysics.", "Assessment": "Analyst inference: A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-004-1", "Initiative_ID": "AI-142", "Title": "ERMAI subsurface opportunity identification", "Date": "2025-12-31", "Finding": "2025 performance reported in Integrated Report 2025", "Metric": "", "Source_IDs": "RS017", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-02-27", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-004"], "Research_search": " {\"key\": \"petronas-ermai\", \"existing_ids\": [], \"company\": \"PETRONAS\", \"title\": \"ERMAI subsurface opportunity identification\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"In-house AI solution; value uplift reported for 2025\", \"practical_relevance\": \"A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.\", \"business_problem\": \"Find bypassed hydrocarbons and accelerate subsurface analysis.\", \"workflow\": \"ERMAI analyzes subsurface data for behind-casing opportunities and real-time petrophysics.\", \"technology_data\": \"In-house AI; models/data integration unspecified.\", \"integration_human_role\": \"Decision rights/validation not described.\", \"deployment_scope\": \"PETRONAS upstream; asset count unspecified.\", \"partners\": [\"Internal PETRONAS development\"], \"disclosed_results\": [{\"metric\": \"subsurface analysis time\", \"value\": \"~6 days to 6 seconds\", \"period\": \"2025 report\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS017\"]}], \"economics\": \"PETRONAS reports analysis time falling from ~6 days to 6 seconds and 2025 recovery value uplift; amount not given.\", \"timeline\": [{\"date\": \"2025-12-31\", \"event\": \"2025 performance reported in Integrated Report 2025\", \"source_ids\": [\"RS017\"]}], \"limitations\": [\"Source is excerpt review of the annual report; details of the metric baseline and the value-uplift calculation are not disclosed.\"], \"assessment\": \"Analyst inference: A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.\", \"sources\": [{\"id\": \"RS017\", \"title\": \"PETRONAS Integrated Report 2025\", \"url\": \"https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf\", \"publisher\": \"PETRONAS\", \"date\": \"2026-02-27\", \"type\": \"annual report\", \"locator\": \"Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"ERMAI subsurface analysis and company-reported analysis-time/recovery claims.\"], \"original_id\": \"petronas_report\"}], \"field_sources\": {\"business_problem\": [\"RS017\"], \"workflow\": [\"RS017\"], \"technology_data\": [\"RS017\"], \"integration_human_role\": [\"RS017\"], \"deployment_scope\": [\"RS017\"], \"economics\": [\"RS017\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Exact-match audit: ERMAI is a separate initiative; AI-050 (TIRAI) and AI-080 (myPROdata) are related, not exact matches.\", \"related_existing_ids\": [\"AI-050\", \"AI-080\"], \"cohort\": \"nocs\", \"packet_id\": \"nocs-004\", \"source_ids\": [\"RS017\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-142\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-142", "company": "PETRONAS", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf"], "findings": [{"severity": "high", "field": "source access", "issue": "PETRONAS integrated report PDF could not be fully opened (viewer error); excerpt/search evidence is insufficient to certify every ERMAI technical detail.", "correction": "Mark excerpt-only/unverified where assertions exceed exposed text; do not call full report reviewed.", "source_url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and the AI passages were located in the source.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-143", "Company": "PETRONAS", "Company_primary": "PETRONAS", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "TriCiptaAI exploration products and surface-equipment development", "Extracted_finding": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.", "Stage_at_disclosure": "Earlier exploration tools deployed; new surface-equipment work remains development", "Reported_metric": "No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.", "Interpretation_limit": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.", "Source_IDs": "RS018; RS017; M29-011; OCT14C-PETRONAS-TRICIPTA", "Source_date": "2026-07-08", "Source_type": "Company release", "Source_locator": "8 July 2026 release: third JDA scope and final two paragraphs on earlier tools", "Source_URL": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Company release dated 8 July 2026. Earlier exploration tools came from 2025 agreements; the third agreement was newly signed.", "Context": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.", "Implementation": "The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "Assessment": "What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?", "Primary_department": "Exploration & geoscience", "Secondary_departments": ["Production & field operations", "Maintenance & reliability"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-005-1", "Initiative_ID": "AI-143", "Title": "TriCiptaAI exploration AI products", "Date": "2025-06-18", "Finding": "TriCipta AI joint development agreement announced", "Metric": "", "Source_IDs": "RS018", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "nocs-005-2", "Initiative_ID": "AI-143", "Title": "TriCiptaAI exploration AI products", "Date": "2025-11", "Finding": "AI.SEEK and GEB first MVPs delivered, per 2025 Integrated Report", "Metric": "", "Source_IDs": "RS017", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "M29-D-AI-143", "Initiative_ID": "AI-143", "Title": "TriCipta AI third JDA: surface equipment optimization", "Date": "2026-07-08", "Finding": "Third joint development agreement extends the TriCipta partnership into AI for upstream surface equipment optimization, targeting production, maintenance decisions and asset value.", "Metric": "No operating or economic metric for surface equipment optimization disclosed. Prior-JDA AI.SEEK and GEB deployment is reported without quantified result.", "Source_IDs": "M29-011", "Evidence": "Full issuer media release reviewed.", "Origin": "29 September 2026 gap review"}, {"ID": "OCT14C-AI-143", "Initiative_ID": "AI-143", "Title": "Deployed exploration tools and a separate equipment-development agreement", "Date": "2026-07-08", "Finding": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools. The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "Metric": "No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.", "Source_IDs": "OCT14C-PETRONAS-TRICIPTA", "Source_URL": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed."}], "Latest_disclosure": "2026-07-08", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-143", "nocs-005", "oct14c-ai-143"], "Research_search": "{\"packet_id\": \"oct14c-ai-143\", \"key\": \"oct14c-ai-143\", \"record_ids\": [\"AI-143\"], \"company\": \"PETRONAS\", \"title\": \"Deployed exploration tools and a separate equipment-development agreement\", \"category\": \"Internal adoption\", \"department\": \"Exploration & geoscience\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Earlier exploration tools deployed; new surface-equipment work remains development\", \"business_problem\": \"PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.\", \"workflow\": \"The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.\", \"technology_data\": \"The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.\", \"deployment_scope\": \"Company release dated 8 July 2026. Earlier exploration tools came from 2025 agreements; the third agreement was newly signed.\", \"integration_human_role\": \"This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.\", \"economics\": \"No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.\", \"practical_relevance\": \"What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?\", \"assessment\": \"What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?\", \"limitations\": [\"This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.\"], \"source_ids\": [\"OCT14C-PETRONAS-TRICIPTA\"], \"sources\": [{\"id\": \"OCT14C-PETRONAS-TRICIPTA\", \"title\": \"PETRONAS advances TriCipta AI through the third joint development agreement\", \"url\": \"https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm\", \"date\": \"2026-07-08\", \"type\": \"Company release\", \"publisher\": \"PETRONAS\", \"locator\": \"8 July 2026 release: third JDA scope and final two paragraphs on earlier tools\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-07-08\", \"event\": \"PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.\", \"source_ids\": [\"OCT14C-PETRONAS-TRICIPTA\"]}], \"field_sources\": {\"business_problem\": [\"OCT14C-PETRONAS-TRICIPTA\"], \"workflow\": [\"OCT14C-PETRONAS-TRICIPTA\"], \"technology_data\": [\"OCT14C-PETRONAS-TRICIPTA\"], \"deployment_scope\": [\"OCT14C-PETRONAS-TRICIPTA\"], \"economics\": [\"OCT14C-PETRONAS-TRICIPTA\"]}}", "Research_status": "Clarified 14 October 2026 · third pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-143", "company": "PETRONAS", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/bm/node/1688"], "findings": [{"severity": "medium", "field": "stage", "issue": "TriCipta release supports venture/product launch but does not establish named customer production deployment or measured subsurface outcome.", "correction": "Separate announced AI venture and intended product suite from operator use.", "source_url": "https://www.petronas.com/bm/node/1688"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.", "Outcome_counted": false, "Outcome_count_note": "No quantified AI outcome supplied.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Earlier exploration tools deployed; new surface-equipment work remains development", "independent_validation": "Not established"}, "Current_source_check": "2026-10-14", "Research_origin": ["research/gap-swarm-oct14c/README.md"]}, {"ID": "AI-144", "Company": "Sinopec", "Company_primary": "Sinopec", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "Jianghan Oilfield AI for shale gas, drilling, safety and production", "Extracted_finding": "This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.", "Stage_at_disclosure": "Multiple use cases described as in use in 2025", "Reported_metric": "fracture simulation accuracy: >80% (Company-reported); safety inspection efficiency: >50% increase (Company-reported); liquid-loading alert accuracy: >95% (Company-reported)", "Interpretation_limit": "Chinese subsidiary/field source; group-wide deployment is not established. Performance measures are company-reported.", "Source_IDs": "RS021", "Source_date": "2025-05-28", "Source_type": "company announcement", "Source_locator": "Jianghan Oilfield AI deployment section", "Source_URL": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-05-28; see chronology for operating dates.", "Context": "Improve shale-gas development and field work by refreshing models, warning about drilling hazards, tuning fracturing, monitoring safety and alerting on liquid loading.", "Implementation": "AI use cases include a shale-gas digital twin with dynamically updated geologic models; drilling risk list generation based on planned well path and complex-event proximity; fracture parameter optimization; video detection of unsafe acts/PPE; drone/video/human inspection; liquid-loading warning; and Jianghan Zhiyan LLM for routine work.", "Assessment": "Analyst inference: This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-008-1", "Initiative_ID": "AI-144", "Title": "Jianghan Oilfield AI for shale gas, drilling, safety and production", "Date": "2025-05-28", "Finding": "Jianghan deployment portfolio and use cases published", "Metric": "", "Source_IDs": "RS021", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-05-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-008"], "Research_search": " {\"key\": \"sinopec-jianghan\", \"existing_ids\": [], \"company\": \"Sinopec\", \"title\": \"Jianghan Oilfield AI for shale gas, drilling, safety and production\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Multiple use cases described as in use in 2025\", \"practical_relevance\": \"This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.\", \"business_problem\": \"Improve shale-gas development and field work by refreshing models, warning about drilling hazards, tuning fracturing, monitoring safety and alerting on liquid loading.\", \"workflow\": \"AI use cases include a shale-gas digital twin with dynamically updated geologic models; drilling risk list generation based on planned well path and complex-event proximity; fracture parameter optimization; video detection of unsafe acts/PPE; drone/video/human inspection; liquid-loading warning; and Jianghan Zhiyan LLM for routine work.\", \"technology_data\": \"AI models, video analytics, digital twin and LLM using geological, drilling, fracture, production and safety data. Vendor/model and integration architecture not stated.\", \"integration_human_role\": \"Warnings and recommendations support field staff/engineers. Automated surveillance and alerting are described; full autonomous control is not.\", \"deployment_scope\": \"Jianghan Oilfield and Fuling shale-gas operating area, including >800 well-instance liquid-loading warning applications.\", \"partners\": [\"Sinopec Jianghan Oilfield\"], \"disclosed_results\": [{\"metric\": \"fracture simulation accuracy\", \"value\": \">80%\", \"period\": \"2025 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS021\"]}, {\"metric\": \"safety inspection efficiency\", \"value\": \">50% increase\", \"period\": \"2025 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS021\"]}, {\"metric\": \"liquid-loading alert accuracy\", \"value\": \">95%\", \"period\": \"2025 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS021\"]}], \"economics\": \"Company-reported: fracturing simulation accuracy >80%; safety supervision efficiency +50%, missed-violation rate <10%; liquid-loading warnings accuracy >95%, response efficiency >90%. Causal attribution/baselines not fully described.\", \"timeline\": [{\"date\": \"2025-05-28\", \"event\": \"Jianghan deployment portfolio and use cases published\", \"source_ids\": [\"RS021\"]}], \"limitations\": [\"Chinese subsidiary/field source; group-wide deployment is not established. Performance measures are company-reported.\"], \"assessment\": \"Analyst inference: This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.\", \"sources\": [{\"id\": \"RS021\", \"title\": \"〖中国石化报〗多场景推进人工智能技术应用\", \"url\": \"https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml\", \"publisher\": \"Sinopec / Jianghan Oilfield\", \"date\": \"2025-05-28\", \"type\": \"company announcement\", \"locator\": \"Jianghan Oilfield AI deployment section\", \"review_status\": \"full text reviewed\", \"supports\": [\"Jianghan Oilfield describes AI for shale-gas digital-twin reservoir updates, drilling risk alerts, fracture optimization, video safety monitoring, field inspection and work assistance.\", \"Reported examples include fracture simulation accuracy above 80%, video safety inspection efficiency up over 50% and missed-violation detection below 10%.\", \"A liquid-loading warning use case had more than 800 well instances, parsed 150,000 records/day, and company-reported warning accuracy above 95%; treatment efficiency above 90%.\"], \"original_id\": \"sinopec\"}], \"field_sources\": {\"business_problem\": [\"RS021\"], \"workflow\": [\"RS021\"], \"technology_data\": [\"RS021\"], \"integration_human_role\": [\"RS021\"], \"deployment_scope\": [\"RS021\"], \"economics\": [\"RS021\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-008\", \"source_ids\": [\"RS021\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-144\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-144", "company": "Sinopec", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-145", "Company": "CNOOC", "Company_primary": "CNOOC", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "AI+ smart field, intelligent drilling and typhoon response", "Extracted_finding": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.", "Stage_at_disclosure": "Operating integrated digital field; AI contribution not isolated", "Reported_metric": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI. Earlier smart-drilling demonstration: 26% faster drilling, also not AI-only.", "Interpretation_limit": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved.", "Source_IDs": "RS022; WATCH15AM-CNOOC", "Source_date": "2026-04-09", "Source_type": "Company publication", "Source_locator": "Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026", "Source_URL": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "Evidence_status": "Relevant 2025 ESG report passage verified; field results remain company-reported and are not AI-only.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "2025 reporting period; report released 9 April 2026. Start date, observation periods and deployed user count are not specified. More than 26,000 points measures instrumentation, not people or wells.", "Context": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.", "Implementation": "More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval. Earlier coverage also describes intelligent drilling/completions and typhoon response.", "Assessment": "The competitive advantage may come from connecting data and decisions across a field. Ask what changed in field work, how recommendations are checked, and which benefits remain after the cost of keeping the system reliable.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-009-1", "Initiative_ID": "AI-145", "Title": "AI+ smart field, intelligent drilling and typhoon response", "Date": "2025-08-27", "Finding": "CNOOC mid-year results describe AI+ scenes, injection-production, smart drilling and weather response", "Metric": "", "Source_IDs": "RS022", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "WATCH15AM-CNOOC", "Initiative_ID": "AI-145", "Title": "One oilfield connects reservoir, production and equipment data", "Date": "2026-04-09", "Finding": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans. More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.", "Metric": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.", "Source_IDs": "WATCH15AM-CNOOC", "Source_URL": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "Evidence": "Original company ESG report reviewed", "Origin": "15 October morning watch; existing program expanded", "Assessment": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved."}], "Latest_disclosure": "2026-04-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch15am-ai-145", "nocs-009"], "Research_search": "{\"packet_id\": \"watch15am-ai-145\", \"key\": \"watch15am-ai-145\", \"record_ids\": [\"AI-145\"], \"company\": \"CNOOC\", \"title\": \"One oilfield connects reservoir, production and equipment data\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-15\", \"verification\": \"verified primary\", \"stage\": \"Operating integrated digital field; AI contribution not isolated\", \"business_problem\": \"CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.\", \"workflow\": \"More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.\", \"technology_data\": \"More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.\", \"deployment_scope\": \"2025 reporting period; report released 9 April 2026. Start date, observation periods and deployed user count are not specified. More than 26,000 points measures instrumentation, not people or wells.\", \"integration_human_role\": \"Automatic plan generation is reported; execution authority, review and safety limits are not disclosed.\", \"economics\": \"CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.\", \"practical_relevance\": \"The competitive advantage may come from connecting data and decisions across a field. Ask what changed in field work, how recommendations are checked, and which benefits remain after the cost of keeping the system reliable.\", \"assessment\": \"The competitive advantage may come from connecting data and decisions across a field. Ask what changed in field work, how recommendations are checked, and which benefits remain after the cost of keeping the system reliable.\", \"limitations\": [\"Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved.\"], \"source_ids\": [\"WATCH15AM-CNOOC\"], \"sources\": [{\"id\": \"WATCH15AM-CNOOC\", \"title\": \"2025 Environmental, Social and Governance Report\", \"url\": \"https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf\", \"date\": \"2026-04-09\", \"publisher\": \"CNOOC Limited\", \"locator\": \"Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026\", \"access\": \"Parent retrieved original PDF and checked the relevant page on 15 October 2026; results remain source-reported.\"}], \"disclosed_results\": [{\"metric\": \"Integrated field outcomes; not isolated AI benefits\", \"value\": \"CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.\", \"period\": \"2025 report; measurement period not supplied\", \"attribution\": \"CNOOC; source-reported\", \"source_ids\": [\"WATCH15AM-CNOOC\"]}], \"timeline\": [{\"date\": \"2026-04-09\", \"event\": \"CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.\", \"source_ids\": [\"WATCH15AM-CNOOC\"]}], \"field_sources\": {\"business_problem\": [\"WATCH15AM-CNOOC\"], \"workflow\": [\"WATCH15AM-CNOOC\"], \"technology_data\": [\"WATCH15AM-CNOOC\"], \"deployment_scope\": [\"WATCH15AM-CNOOC\"], \"integration_human_role\": [\"WATCH15AM-CNOOC\"], \"economics\": [\"WATCH15AM-CNOOC\"]}}", "Research_status": "Expanded 15 October 2026 · named field inside existing program", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-145", "company": "CNOOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved.", "Outcome_counted": true, "Outcome_count_note": "Existing source-reported numerical-result profile. Integrated smart-field and drilling results do not isolate AI contribution; this update adds no result count.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-15", "Research_origin": ["research/watch-2026-10-15-am/README.md"]}, {"ID": "AI-146", "Company": "Petrobras", "Company_primary": "Petrobras", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI industrial-flare control and Smart Tocha monitoring", "Extracted_finding": "Petrobras reports AI control of industrial flares in real time; earlier corporate material separately describes Smart Tocha automatic steam-flow adjustment.", "Stage_at_disclosure": "Annualized emissions result reported; Smart Tocha’s full rollout is not established by this slide.", "Reported_metric": "10,000 tCO2/year reduction attributed to real-time AI industrial-flare control.", "Interpretation_limit": "Do not assign the full result specifically to Smart Tocha: the slide does not name it. No baseline, measurement period, plant count or emissions calculation method. P-52 is newly identified scope, not proof that the strategy-deck 10,000 tCO2/year result belongs to P-52 or Smart Tocha.", "Source_IDs": "RS023; SW2-006; OCT5-S040", "Source_date": "2025-11", "Source_type": "Issuer investor presentation", "Source_locator": "Slide 158, Accelerate value generation through digital technologies", "Source_URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Evidence_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "Context": "Adds quantified company evidence to the existing flare-control topic without claiming a confirmed product-level attribution.", "Implementation": "The plan slide adds an aggregate flare-control outcome but does not name Smart Tocha, sites or the rollout population. The earlier named product and this result are retained as related flare-workflow evidence.", "Assessment": "Adds quantified company evidence to the existing flare-control topic without claiming a confirmed product-level attribution.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-010-1", "Initiative_ID": "AI-146", "Title": "Smart Tocha AI flare combustion monitoring and adjustment", "Date": "2026", "Finding": "Company-stated planned availability across new units by 2026; actual completion not verified", "Metric": "", "Source_IDs": "RS023", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "SW2-D-AI-146", "Initiative_ID": "AI-146", "Title": "AI industrial-flare control and Smart Tocha monitoring", "Date": "2025-11", "Finding": "Petrobras reports AI control of industrial flares in real time; earlier corporate material separately describes Smart Tocha automatic steam-flow adjustment.", "Metric": "10,000 tCO2/year reduction attributed to real-time AI industrial-flare control.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}, {"ID": "OCT5-AI-146-1", "Initiative_ID": "AI-146", "Title": "P-52 flare-reduction application: platform scope added", "Date": "2026-05", "Finding": "Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.", "Metric": "", "Source_IDs": "OCT5-S040", "Source_URL": "https://transparencia.petrobras.com.br/documents/41711919/42223019/Relatorio%2Bde%2BGestao%2B-%2B2025.pdf/9ef35500-1f38-0cb6-539d-072b0569514f?download=true&t=1779977455000&version=1.0", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Latest_disclosure": "2026-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-three-rounds-ai-146", "swarm2-ai-146", "nocs-010"], "Research_search": "{\"packet_id\": \"oct6-three-rounds-ai-146\", \"review_date\": \"2026-10-06\", \"title\": \"P-52 flare-reduction application: platform scope added\", \"workflow\": \"Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.\", \"technology_data\": \"Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.\", \"deployment_scope\": \"AI applied to P-52 in 2025; no launch month or P-52-specific result. Separate Petrobras field source locates P-52 at Roncador in Campos Basin.\", \"result_label\": \"Supplemental source: scope and limits\", \"attribution\": \"Operator-reported\", \"economics\": \"No project cost and measured cash-benefit pair disclosed.\", \"limitations\": [\"P-52 is newly identified scope, not proof that the strategy-deck 10,000 tCO2/year result belongs to P-52 or Smart Tocha.\"], \"verification\": \"verified primary\", \"source_review\": \"Full original source reviewed in Luna verification and editor checks on 6 October 2026.\", \"source_locator\": \"2025 innovation highlights, physical/printed p. 133\", \"timeline_date\": \"2026-05\", \"timeline_event\": \"P-52 flare-reduction application: platform scope added\", \"stage\": \"Annualized emissions result reported; Smart Tocha’s full rollout is not established by this slide.\", \"key\": \"oct6-three-rounds-ai-146\", \"record_ids\": [\"AI-146\"], \"company\": \"Petrobras\", \"category\": \"Internal adoption\", \"department\": \"Refining, LNG & process operations\", \"source_ids\": [\"OCT5-S040\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S040\"], \"workflow\": [\"OCT5-S040\"], \"technology_data\": [\"OCT5-S040\"], \"deployment_scope\": [\"OCT5-S040\"], \"practical_relevance\": [\"OCT5-S040\"], \"economics\": [\"OCT5-S040\"], \"integration_human_role\": [\"OCT5-S040\"]}, \"disclosed_results\": [{\"metric\": \"Supplemental source: scope and limits\", \"value\": \"No P-52-specific outcome reported. The existing 10,000 tCO2/year deck claim remains at company/program scope.\", \"period\": \"2026-05\", \"attribution\": \"Operator-reported\", \"source_ids\": [\"OCT5-S040\"]}], \"timeline\": [{\"date\": \"2026-05\", \"event\": \"P-52 flare-reduction application: platform scope added\", \"source_ids\": [\"OCT5-S040\"]}], \"sources\": [{\"id\": \"OCT5-S040\", \"title\": \"Petrobras Relatório de Gestão 2025\", \"url\": \"https://transparencia.petrobras.com.br/documents/41711919/42223019/Relatorio%2Bde%2BGestao%2B-%2B2025.pdf/9ef35500-1f38-0cb6-539d-072b0569514f?download=true&t=1779977455000&version=1.0\", \"date\": \"2026-05\", \"type\": \"Issuer management report\", \"publisher\": \"Petrobras\", \"locator\": \"2025 innovation highlights, physical/printed p. 133\", \"review_status\": \"Full original source reviewed in Luna verification and editor checks on 6 October 2026.\"}]}", "Research_status": "Updated · second September review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-146", "company": "Petrobras", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://petrobras.com.br/inovacao-e-tecnologia"], "findings": [{"severity": "high", "field": "source support", "issue": "Petrobras generic innovation page did not expose the named Smart Tocha deployment details on reopen.", "correction": "Specific workflow/date/metrics remain unverified until a direct Petrobras technical or operations source is located.", "source_url": "https://petrobras.com.br/inovacao-e-tecnologia"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Asset_location": "Assets not specified on slide", "Followup_review": {"date": "2026-09-29", "decision": "Enriched", "note": "Deduplicated, original-source checked and separately reviewed."}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-147", "Company": "Ecopetrol", "Company_primary": "Ecopetrol", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "AI and advanced analytics ecosystem; intelligent upstream planning", "Extracted_finding": "The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.", "Stage_at_disclosure": "Ecosystem implemented in 2025; maturity assessment identified constraints", "Reported_metric": "AI investment US$1.88m in 2025; benefits US$3.5m (revenue +US$2.06m, costs −US$1.44m); cumulative ROI since 2021 estimated at 7% (company-reported); 4,609 labor hours released by intelligent planner", "Interpretation_limit": "Company-reported AI portfolio investment and revenue/cost benefits, not audited project cash ROI. No all-in project cost, cash-conversion bridge or comparable asset-level denominator; the cumulative ROI uses a different period.", "Source_IDs": "RS024; SW6-004", "Source_date": "2026-03-04", "Source_type": "annual report", "Source_locator": "2025 sustainability annex, PDF page 209 (printed page 209); issued 4 March 2026", "Source_URL": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "2025 portfolio benefits; sustainability annex issued 4 March 2026. Cumulative ROI uses 2021 onward.", "Context": "Create shared AI/analytics capabilities and enable business initiatives; an upstream planner addresses repeated planning work.", "Implementation": "Ecopetrol reports an AI and advanced analytics ecosystem and an intelligent upstream planner that released staff hours. Other named application details were not disclosed in this report.", "Assessment": "Analyst inference: The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-011-1", "Initiative_ID": "AI-147", "Title": "AI and advanced analytics ecosystem; intelligent upstream planning", "Date": "2025-12-31", "Finding": "Ecosystem and planner outcomes included in 2025 report", "Metric": "", "Source_IDs": "RS024", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "SW6-D-AI-147", "Initiative_ID": "AI-147", "Title": "AI portfolio: 2025 investment and benefits disclosure", "Date": "2026", "Finding": "Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.", "Metric": "2025: US$1.88m reported AI investment, US$2.06m revenue increase and US$1.44m cost reduction (US$3.50m total benefits). Reported ROI: 7%, cumulative since 2021.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}], "Latest_disclosure": "2026-03-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm6-ai-147", "nocs-011"], "Research_search": " {\"key\": \"ecopetrol-ai-ecosystem\", \"existing_ids\": [], \"company\": \"Ecopetrol\", \"title\": \"AI and advanced analytics ecosystem; intelligent upstream planning\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Ecosystem implemented in 2025; maturity assessment identified constraints\", \"practical_relevance\": \"The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.\", \"business_problem\": \"Create shared AI/analytics capabilities and enable business initiatives; an upstream planner addresses repeated planning work.\", \"workflow\": \"Ecopetrol reports an AI and advanced analytics ecosystem and an intelligent upstream planner that released staff hours. Other named application details were not disclosed in this report.\", \"technology_data\": \"Shared capabilities, resources and infrastructure; report cites limited data access and autonomy as maturity gaps.\", \"integration_human_role\": \"Business teams and AI community; human decision processes implied, details not disclosed.\", \"deployment_scope\": \"Group ecosystem; intelligent planner in upstream, asset count not specified.\", \"partners\": [\"Internal Ecopetrol teams\"], \"disclosed_results\": [{\"metric\": \"labor hours released by intelligent planner\", \"value\": \"4,609 hours\", \"period\": \"2025\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS024\"]}], \"economics\": \"Planner reportedly released 4,609 labor hours. Company reporting does not establish this as a monetary return or validate the causal contribution of AI.\", \"timeline\": [{\"date\": \"2025-12-31\", \"event\": \"Ecosystem and planner outcomes included in 2025 report\", \"source_ids\": [\"RS024\"]}], \"limitations\": [\"AI maturity and access issues remain; annual-report excerpt reviewed, so claims limited to the surfaced passages.\"], \"assessment\": \"Analyst inference: The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.\", \"sources\": [{\"id\": \"RS024\", \"title\": \"Informe de gestión y resultados 2025\", \"url\": \"https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES\", \"publisher\": \"Ecopetrol\", \"date\": \"2026-03-31\", \"type\": \"annual report\", \"locator\": \"Technology and innovation section; AI ecosystem and upstream planner\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Ecopetrol reports establishing an AI and advanced analytics ecosystem and identifies data access and autonomy limits in its maturity assessment.\", \"Upstream intelligent planner reportedly released 4,609 labor hours; report does not attribute all AI ecosystem benefits to that planner or quantify causal AI-specific savings.\"], \"original_id\": \"ecopetrol_2025\"}], \"field_sources\": {\"business_problem\": [\"RS024\"], \"workflow\": [\"RS024\"], \"technology_data\": [\"RS024\"], \"integration_human_role\": [\"RS024\"], \"deployment_scope\": [\"RS024\"], \"economics\": [\"RS024\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-011\", \"source_ids\": [\"RS024\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-147\"]} {\"key\": \"swarm6-ai-147\", \"packet_id\": \"swarm6-ai-147\", \"record_ids\": [\"AI-147\"], \"existing_ids\": [\"AI-147\"], \"related_existing_ids\": [], \"company\": \"Ecopetrol\", \"title\": \"AI portfolio: 2025 investment and benefits disclosure\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"2025 portfolio disclosure; initiative count does not prove all tools are live.\", \"practical_relevance\": \"Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.\", \"business_problem\": \"\", \"workflow\": \"Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.\", \"technology_data\": \"More than 35 initiatives generated during 2025; eight analytics hubs. Costs and outcomes are aggregated, not allocated to individual tools.\", \"integration_human_role\": \"Not specified.\", \"deployment_scope\": \"2025 portfolio disclosure; initiative count does not prove all tools are live.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"2025: US$1.88m reported AI investment, US$2.06m revenue increase and US$1.44m cost reduction (US$3.50m total benefits). Reported ROI: 7%, cumulative since 2021.\", \"period\": \"2025 annual metrics; ROI cumulative from 2021 through 2025.\", \"attribution\": \"Company-reported; no initiative-specific independent assurance established.\", \"source_ids\": [\"SW6-004\"]}], \"economics\": \"Some portfolio economics are disclosed, but the cost perimeter is incomplete. Do not compute a one-year ROI from US$3.50m / US$1.88m; spending, benefit timing and the stated ROI have different scopes.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"2026 report on 2025 activities; precise publication day not established.\", \"source_ids\": [\"SW6-004\"]}], \"limitations\": [\"The 7% uses five-year consolidated, EBITDA-aligned figures with lagged benefits. The investment measure excludes other AI spending without direct financial impact. No project-level payback or independently verified causal return.\"], \"assessment\": \"Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.\", \"sources\": [{\"title\": \"Ecopetrol 2025 Integrated Management Report — AI KPIs\", \"url\": \"https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf\", \"date\": \"2026 AGM report; covers 2025\", \"locator\": \"Printed pp. 318 and 447; physical PDF pp. 319 and 448, table 7.115 and explanatory note.\", \"type\": \"Operator annual report\", \"review_status\": \"Original PDF downloaded and relevant full pages read; source text and page locators independently checked.\", \"id\": \"SW6-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW6-004\"], \"field_sources\": {\"workflow\": [\"SW6-004\"], \"technology_data\": [\"SW6-004\"], \"deployment_scope\": [\"SW6-004\"], \"practical_relevance\": [\"SW6-004\"]}, \"verification\": \"verified primary\", \"cohort\": \"Sixth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported; no initiative-specific independent assurance established.\"}", "Research_status": "Supplemental source added · October gap review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-147", "company": "Ecopetrol", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES"], "findings": [{"severity": "medium", "field": "scope", "issue": "Ecopetrol 2025 report supports group AI/analytics activity but not all profile details about a specific upstream planning workflow.", "correction": "Do not attribute general R&D, digital or group results to AI or to upstream planning without direct wording.", "source_url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "Record combines Ecopetrol-wide AI/advanced-analytics capability and an intelligent upstream planner; the latter is Upstream, while the ecosystem is enterprise-wide. Cross-segment is defensible for the combined record, but split if the upstream planner is the primary counted initiative.", "Segment_classification_confidence": "medium", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Ecopetrol reports US$1.88m of AI-project investment and US$3.50m of financial benefits in 2025 (US$2.06m higher revenue and US$1.44m lower costs), across its AI portfolio. This is company-reported portfolio accounting, not a project-level reconciliation of total cost and collected cash. Cumulative ROI since 2021, tied to EBITDA margin, is estimated at 7%; it uses a different period and basis. The planner released 4,609 labor hours.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": true, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-148", "Company": "Ecopetrol", "Company_primary": "Ecopetrol", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "CADE energy dispatch decision platform", "Extracted_finding": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI.", "Stage_at_disclosure": "CADE described in use; Rubiales dispatch-model MVP under construction in 2025", "Reported_metric": "No measured AI-specific energy, emissions, reliability, cost or production benefit disclosed. The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.", "Interpretation_limit": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given. May 2026 follow-up: No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.", "Source_IDs": "RS025; SW6-004; OCT14C-ECOPETROL-CADE; WATCH20PM-CADE", "Source_date": "2026 (2025 annual report)", "Source_type": "Operator annual report", "Source_locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "Source_URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "2025 operating and development activity, described in the 2026 annual-report edition. Exact publication day and MVP commissioning date are not established.", "Context": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI.", "Implementation": "A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.\n\nMay 2026 coverage: Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.", "Assessment": "Are Rubiales recommendations now used, who approves a setpoint, and how is the result compared with the previous dispatch method?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-012-1", "Initiative_ID": "AI-148", "Title": "CADE energy dispatch decision platform", "Date": "2026-05-04", "Finding": "CADE opened; company says AI/analytics will be developed for energy optimization", "Metric": "", "Source_IDs": "RS025", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "SW6-D-AI-148", "Initiative_ID": "AI-148", "Title": "CADE energy dispatch: Rubiales model development update", "Date": "2026", "Finding": "Consolidate electrical operating signals and develop dispatch recommendations for Rubiales operations.", "Metric": "No quantified AI-attributed result disclosed.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}, {"ID": "OCT14C-AI-148", "Initiative_ID": "AI-148", "Title": "An operating energy desk, with a dispatch model still being built", "Date": "2026 (2025 annual report)", "Finding": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI. A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.", "Metric": "No measured energy, emissions, reliability, cost or production benefit is disclosed for these AI components.", "Source_IDs": "OCT14C-ECOPETROL-CADE", "Source_URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given."}, {"ID": "WATCH20PM-AI-148", "Initiative_ID": "AI-148", "Title": "Rubiales adds a dated operating footprint to CADE", "Date": "2026-05-04", "Finding": "Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.", "Metric": "The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.", "Source_IDs": "WATCH20PM-CADE", "Source_URL": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "Evidence": "Original operator announcement read; power coverage is not AI performance", "Origin": "20 October evening Luna research; editor original-source review", "Assessment": "No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population."}], "Latest_disclosure": "2026-05-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch20pm-ai-148", "swarm6-ai-148", "nocs-012", "oct14c-ai-148"], "Research_search": "{\"packet_id\": \"watch20pm-ai-148\", \"key\": \"watch20pm-ai-148\", \"record_ids\": [\"AI-148\"], \"company\": \"Ecopetrol\", \"title\": \"Rubiales adds a dated operating footprint to CADE\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-20\", \"verification\": \"verified primary\", \"stage\": \"CADE operating coverage; specific AI-model commissioning unverified\", \"business_problem\": \"Choose the timing and mix of self-generation and purchased electricity for operating assets.\", \"workflow\": \"CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.\", \"technology_data\": \"2025 report describes analytics and GenAI in CADE, with a separate Rubiales dispatch-model MVP under construction. May 2026 release adds platform coverage, not model-specific commissioning.\", \"integration_human_role\": \"60 workers trained, not 60 verified active users. Source does not define approval, override or control permissions.\", \"deployment_scope\": \"Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. Issuer announcement May 4, 2026; Rubiales coverage reported from March 2026. Older evidence newly checked, not a new October launch.\", \"economics\": \"The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.\", \"practical_relevance\": \"Which decisions are actually automated, and what changes in delivered electricity cost after all platform and operating costs?\", \"assessment\": \"Trace whether an operating decision desk translates into lower delivered electricity cost. Capacity and a name implying autonomy cannot answer that question.\", \"limitations\": [\"No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.\", \"The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given. May 2026 follow-up: No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.\"], \"source_ids\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"partners\": [\"ISA\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2025\", \"event\": \"Rubiales model proof of concept and MVP construction described in earlier annual report.\", \"source_ids\": [\"OCT14C-ECOPETROL-CADE\"]}, {\"date\": \"2026-05-04\", \"event\": \"Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid.\", \"source_ids\": [\"WATCH20PM-CADE\"]}], \"sources\": [{\"id\": \"OCT14C-ECOPETROL-CADE\", \"title\": \"Ecopetrol Integrated Management Report 2025\", \"url\": \"https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf\", \"date\": \"2026 (2025 annual report)\", \"publisher\": \"Ecopetrol\", \"type\": \"Operator annual report\", \"locator\": \"PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}, {\"id\": \"WATCH20PM-CADE\", \"title\": \"Grupo Ecopetrol inaugura centro para optimizar su gestión energética\", \"url\": \"https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica\", \"date\": \"2026-05-04\", \"publisher\": \"Ecopetrol\", \"type\": \"Operator announcement\", \"locator\": \"May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs\", \"review_status\": \"Full original operator page retrieved (HTTP200) and read by editor. Coverage and plans remain company-reported; no independent cost or performance validation.\"}], \"field_sources\": {\"business_problem\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"workflow\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"technology_data\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"integration_human_role\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"deployment_scope\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"], \"economics\": [\"OCT14C-ECOPETROL-CADE\", \"WATCH20PM-CADE\"]}}", "Research_status": "Clarified 14 October 2026 · third pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-148", "company": "Ecopetrol", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/"], "findings": [{"severity": "high", "field": "source support", "issue": "CADE announcement URL failed with cache miss, and the profile title alone does not establish that the decision platform uses AI.", "correction": "Mark inaccessible/not re-verified, not disproved; keep AI status unverified.", "source_url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "CADE is an energy-resource dispatch decision platform spanning self-generation, renewable supply and grid purchases. It is not clearly an oil/gas production segment workflow; cross-segment enterprise is the closest current bucket.", "Segment_classification_confidence": "medium", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given. May 2026 follow-up: No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.", "Outcome_counted": false, "Outcome_count_note": "No quantified AI outcome supplied.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "CADE described in use; Rubiales dispatch-model MVP under construction in 2025", "independent_validation": "Not established"}, "Current_source_check": "2026-10-20", "Research_origin": ["research/gap-swarm-oct14c/README.md"]}, {"ID": "AI-149", "Company": "YPF", "Company_primary": "YPF", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "Generative AI for contract cataloging, customer service and refinery inspection", "Extracted_finding": "Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.", "Stage_at_disclosure": "Use cases carried out in 2024 and reported in sustainability report", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "2024 initiatives fall inside window but original launch dates or scale are not specified. Annual-report text accessed as excerpt.", "Source_IDs": "RS026", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Innovation and Technology / Use of artificial intelligence", "Source_URL": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Increase productivity in contract cataloging, make strategic customer information easier to access, and improve inspector decisions/training.", "Implementation": "GenAI/NLP automates contract classification; a chatbot organizes strategic information for commercial decisions; a refinery inspection assistant draws on reports, plant shutdown records and historical asset data to support material choice, practices and technical learning.", "Assessment": "Analyst inference: Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-013-1", "Initiative_ID": "AI-149", "Title": "Generative AI for contract cataloging, customer service and refinery inspection", "Date": "2024", "Finding": "YPF reports the three AI projects as use cases conducted in 2024", "Metric": "", "Source_IDs": "RS026", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-013"], "Research_search": " {\"key\": \"ypf-three-workflows\", \"existing_ids\": [], \"company\": \"YPF\", \"title\": \"Generative AI for contract cataloging, customer service and refinery inspection\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Use cases carried out in 2024 and reported in sustainability report\", \"practical_relevance\": \"Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.\", \"business_problem\": \"Increase productivity in contract cataloging, make strategic customer information easier to access, and improve inspector decisions/training.\", \"workflow\": \"GenAI/NLP automates contract classification; a chatbot organizes strategic information for commercial decisions; a refinery inspection assistant draws on reports, plant shutdown records and historical asset data to support material choice, practices and technical learning.\", \"technology_data\": \"Generative AI and NLP; the inspection assistant connects technical document/history sources. Exact models, integrations and user counts not disclosed.\", \"integration_human_role\": \"Provides assistance to cataloguers, commercial teams and refinery inspectors; YPF positions AI as human-empowering. No autonomous decisions reported.\", \"deployment_scope\": \"Corporate cataloging and customer support; refinery inspection deployment scope not stated.\", \"partners\": [\"Internal YPF development\"], \"disclosed_results\": [], \"economics\": \"No quantified results; company says contract cataloging productivity/information quality improve and inspection support can reduce risk, improve material durability and accelerate learning.\", \"timeline\": [{\"date\": \"2024\", \"event\": \"YPF reports the three AI projects as use cases conducted in 2024\", \"source_ids\": [\"RS026\"]}], \"limitations\": [\"2024 initiatives fall inside window but original launch dates or scale are not specified. Annual-report text accessed as excerpt.\"], \"assessment\": \"Analyst inference: Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.\", \"sources\": [{\"id\": \"RS026\", \"title\": \"Reporte de Sustentabilidad 2024\", \"url\": \"https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf\", \"publisher\": \"YPF\", \"date\": \"2025; publication day not confirmed\", \"type\": \"annual report\", \"locator\": \"Innovation and Technology / Use of artificial intelligence\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"YPF lists 2024 use cases: generative AI/NLP contract cataloging; customer-service chatbot; and refinery inspection virtual assistant.\", \"The inspection assistant uses inspection reports, shutdown records and historical asset data to support material selection, good practices and technical learning; YPF says it can reduce operational risk.\", \"YPF reports governance standards for LLM security and generative AI and general principles for privacy, cybersecurity, bias and labeling.\"], \"original_id\": \"ypf_report\"}], \"field_sources\": {\"business_problem\": [\"RS026\"], \"workflow\": [\"RS026\"], \"technology_data\": [\"RS026\"], \"integration_human_role\": [\"RS026\"], \"deployment_scope\": [\"RS026\"], \"economics\": [\"RS026\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-013\", \"source_ids\": [\"RS026\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-149\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved.", "source_url": ""}], "Audit_review": {"id": "AI-149", "company": "YPF", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and the AI section was located in the source.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-150", "Company": "Pertamina", "Company_primary": "Pertamina", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "PINTAR GenAI HSE question-answering for Rokan workers", "Extracted_finding": "A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.", "Stage_at_disclosure": "Described in Pertamina Hulu Energi 2024 annual report", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Primary annual-report excerpt reviewed; product scope and AI design details are sparse.", "Source_IDs": "RS027", "Source_date": "2025-09-01", "Source_type": "company announcement", "Source_locator": "PINTAR description", "Source_URL": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-09-01; see chronology for operating dates.", "Context": "Help workers find HSE rules/procedures and improve practical safety awareness at work.", "Implementation": "Conversational GenAI answers questions about HSE regulations and procedures; intended to support workplace safety learning.", "Assessment": "Analyst inference: A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.", "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-014-1", "Initiative_ID": "AI-150", "Title": "PINTAR GenAI HSE question-answering for Rokan workers", "Date": "2024", "Finding": "PINTAR reported in Pertamina Hulu Energi annual report, published 2025", "Metric": "", "Source_IDs": "RS027", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-09-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-014"], "Research_search": " {\"key\": \"pertamina-pintar\", \"existing_ids\": [], \"company\": \"Pertamina\", \"title\": \"PINTAR GenAI HSE question-answering for Rokan workers\", \"department\": \"Workforce & learning\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Described in Pertamina Hulu Energi 2024 annual report\", \"practical_relevance\": \"A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.\", \"business_problem\": \"Help workers find HSE rules/procedures and improve practical safety awareness at work.\", \"workflow\": \"Conversational GenAI answers questions about HSE regulations and procedures; intended to support workplace safety learning.\", \"technology_data\": \"ChatGPT-powered GenAI; underlying retrieval corpus, access controls, validation and interface not disclosed.\", \"integration_human_role\": \"Worker self-service Q&A; no human-review pathway or escalation behavior described.\", \"deployment_scope\": \"Pertamina Hulu Rokan Zone; user count, production deployment scope and usage volume not disclosed.\", \"partners\": [\"Pertamina Hulu Rokan\"], \"disclosed_results\": [], \"economics\": \"No outcome metric disclosed.\", \"timeline\": [{\"date\": \"2024\", \"event\": \"PINTAR reported in Pertamina Hulu Energi annual report, published 2025\", \"source_ids\": [\"RS027\"]}], \"limitations\": [\"Primary annual-report excerpt reviewed; product scope and AI design details are sparse.\"], \"assessment\": \"Analyst inference: A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.\", \"sources\": [{\"id\": \"RS027\", \"title\": \"PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone\", \"url\": \"https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf\", \"publisher\": \"Pertamina Hulu Rokan\", \"date\": \"2025-09-01\", \"type\": \"company announcement\", \"locator\": \"PINTAR description\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"PINTAR is a generative AI conversational platform for answering workplace health, safety and environment regulations and procedures and supporting worker safety learning.\", \"Source is a 2024 upstream-subholding annual-report document excerpt; user scale, deployment metrics and outcome measures are not disclosed.\"], \"original_id\": \"pertamina_pintar\"}], \"field_sources\": {\"business_problem\": [\"RS027\"], \"workflow\": [\"RS027\"], \"technology_data\": [\"RS027\"], \"integration_human_role\": [\"RS027\"], \"deployment_scope\": [\"RS027\"], \"economics\": [\"RS027\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-014\", \"source_ids\": [\"RS027\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-150\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [], "Audit_review": {"id": "AI-150", "company": "Pertamina", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf"], "findings": [{"severity": "high", "field": "source access", "issue": "PINTAR primary PDF returned request rejected during review.", "correction": "Mark as unverified on this pass; access failure is not a contradiction.", "source_url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-151", "Company": "Pertamina", "Company_primary": "Pertamina", "Segment": "Geothermal energy (adjacent energy; not oil and gas)", "Theme": "Operational AI", "Initiative": "PGE G-Bionic machine learning and digital twins", "Extracted_finding": "Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.", "Stage_at_disclosure": "Digital program described Oct 2025; value figures are estimates", "Reported_metric": "estimated production increase: up to 6% (Company-reported estimate; machine learning/digital twin program); estimated annual net income equivalent value creation: 5%-7% (Company-reported estimate; not AI-only)", "Interpretation_limit": "Figures are estimates for a combined digital transformation program and should not be attributed solely to AI. Pertamina Geothermal Energy is a geothermal subsidiary; this is an adjacent-energy example rather than oil and gas operations.", "Source_IDs": "RS028", "Source_date": "2025-10-27", "Source_type": "company announcement", "Source_locator": "G-Bionic value case", "Source_URL": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-10-27; see chronology for operating dates.", "Context": "Improve geothermal drilling economics, plant efficiency, well productivity and lost-production performance.", "Implementation": "PGE links machine learning and digital twins to well optimization, plant efficiency and reduced lost production opportunities. Specific workflows or asset-level examples not detailed in this release.", "Assessment": "Analyst inference: Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-015-1", "Initiative_ID": "AI-151", "Title": "PGE G-Bionic machine learning and digital twins", "Date": "2025-10-27", "Finding": "G-Bionic program benefits presented at PGE Digital Day 2025", "Metric": "", "Source_IDs": "RS028", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-10-27", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-015"], "Research_search": " {\"key\": \"pertamina-geothermal-gbionic\", \"existing_ids\": [], \"company\": \"Pertamina\", \"title\": \"PGE G-Bionic machine learning and digital twins\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Digital program described Oct 2025; value figures are estimates\", \"practical_relevance\": \"Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.\", \"business_problem\": \"Improve geothermal drilling economics, plant efficiency, well productivity and lost-production performance.\", \"workflow\": \"PGE links machine learning and digital twins to well optimization, plant efficiency and reduced lost production opportunities. Specific workflows or asset-level examples not detailed in this release.\", \"technology_data\": \"Machine learning and digital twins; source does not detail data, models, controls or specific production integrations.\", \"integration_human_role\": \"Engineering and operating teams presumably apply recommendations; exact human role not stated.\", \"deployment_scope\": \"PGE geothermal operations; site coverage not disclosed.\", \"partners\": [\"Pertamina Geothermal Energy internal teams\"], \"disclosed_results\": [{\"metric\": \"estimated production increase\", \"value\": \"up to 6%\", \"period\": \"program estimate in Oct 2025\", \"attribution\": \"Company-reported estimate; machine learning/digital twin program\", \"source_ids\": [\"RS028\"]}, {\"metric\": \"estimated annual net income equivalent value creation\", \"value\": \"5%-7%\", \"period\": \"program estimate in Oct 2025\", \"attribution\": \"Company-reported estimate; not AI-only\", \"source_ids\": [\"RS028\"]}], \"economics\": \"PGE estimates up to 6% drilling cost reduction, production increase up to 6%, and value equivalent to 5-7% annual net income increase. These are program estimates, not realized/independently audited AI-only returns.\", \"timeline\": [{\"date\": \"2025-10-27\", \"event\": \"G-Bionic program benefits presented at PGE Digital Day 2025\", \"source_ids\": [\"RS028\"]}], \"limitations\": [\"Figures are estimates for a combined digital transformation program and should not be attributed solely to AI.\", \"Pertamina Geothermal Energy is a geothermal subsidiary; this is an adjacent-energy example rather than oil and gas operations.\"], \"assessment\": \"Analyst inference: Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.\", \"sources\": [{\"id\": \"RS028\", \"title\": \"PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness\", \"url\": \"https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness\", \"publisher\": \"Pertamina Geothermal Energy\", \"date\": \"2025-10-27\", \"type\": \"company announcement\", \"locator\": \"G-Bionic value case\", \"review_status\": \"full text reviewed\", \"supports\": [\"PGE describes a digital transformation program using machine learning and digital twins for well optimization, plant efficiency and lost-production opportunity reduction.\", \"PGE estimates potential production increases up to 6%, drilling cost reductions up to 6%, and value equivalent to 5-7% net income annually; these are program estimates, not audited realized AI benefits.\"], \"original_id\": \"pertamina_geothermal\"}], \"field_sources\": {\"business_problem\": [\"RS028\"], \"workflow\": [\"RS028\"], \"technology_data\": [\"RS028\"], \"integration_human_role\": [\"RS028\"], \"deployment_scope\": [\"RS028\"], \"economics\": [\"RS028\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-015\", \"source_ids\": [\"RS028\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-151\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-151", "company": "Pertamina", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness"], "findings": [{"severity": "medium", "field": "classification", "issue": "PGE G-Bionic source is geothermal operations, an adjacent energy business rather than oil-and-gas production.", "correction": "Label geothermal scope and avoid combining program-level efficiency gains as AI-specific.", "source_url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-152", "Company": "ONGC", "Company_primary": "ONGC", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI/ML-enabled offshore gas-lift optimization procurement specification", "Extracted_finding": "The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.", "Stage_at_disclosure": "Tender specification issued Nov 2025; procurement is not evidence of operational deployment", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Tender publication date metadata surfaced as 3 Nov 2025; full procurement result and commissioning were not located in this pass.", "Source_IDs": "RS029", "Source_date": "2025-11-03", "Source_type": "company announcement", "Source_locator": "AI-driven optimization and predictive maintenance requirements", "Source_URL": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-11-03; see chronology for operating dates.", "Context": "Automate gas-lift monitoring and optimize injection while detecting equipment degradation to reduce downtime.", "Implementation": "Tender calls for real-time well-data analysis, gas-lift injection optimization and predictive maintenance alerts from valve/pressure behavior, integrated with existing wellhead control panels and SCADA.", "Assessment": "Analyst inference: The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "nocs-016-1", "Initiative_ID": "AI-152", "Title": "AI/ML-enabled offshore gas-lift optimization procurement specification", "Date": "2025-11-03", "Finding": "Tender requirements include AI-driven optimization and predictive maintenance", "Metric": "", "Source_IDs": "RS029", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["nocs-016"], "Research_search": " {\"key\": \"ongc-gaslift-tender\", \"existing_ids\": [], \"company\": \"ONGC\", \"title\": \"AI/ML-enabled offshore gas-lift optimization procurement specification\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Tender specification issued Nov 2025; procurement is not evidence of operational deployment\", \"practical_relevance\": \"The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.\", \"business_problem\": \"Automate gas-lift monitoring and optimize injection while detecting equipment degradation to reduce downtime.\", \"workflow\": \"Tender calls for real-time well-data analysis, gas-lift injection optimization and predictive maintenance alerts from valve/pressure behavior, integrated with existing wellhead control panels and SCADA.\", \"technology_data\": \"AI/ML specified; no selected vendor, algorithm, data history or acceptance results disclosed.\", \"integration_human_role\": \"Operators/engineers receive local HMI and remote SCADA visibility; autonomous lift behavior is included as demonstration requirement, but actual commissioned control is unverified.\", \"deployment_scope\": \"Offshore wells under the tender scope; installed site count not verified.\", \"partners\": [\"Prospective bidder(s) not identified\"], \"disclosed_results\": [], \"economics\": \"No outcome or cost data; tender is a planned/procurement specification.\", \"timeline\": [{\"date\": \"2025-11-03\", \"event\": \"Tender requirements include AI-driven optimization and predictive maintenance\", \"source_ids\": [\"RS029\"]}], \"limitations\": [\"Tender publication date metadata surfaced as 3 Nov 2025; full procurement result and commissioning were not located in this pass.\"], \"assessment\": \"Analyst inference: The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.\", \"sources\": [{\"id\": \"RS029\", \"title\": \"ONGC tender for automated gas-lift system\", \"url\": \"https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207\", \"publisher\": \"ONGC\", \"date\": \"2025-11-03\", \"type\": \"company announcement\", \"locator\": \"AI-driven optimization and predictive maintenance requirements\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Tender technical scope requires AI/ML algorithms to analyze well data and optimize gas-lift injection in real time, with predictive maintenance based on valve response and pressure trends.\", \"SCADA and existing wellhead control panel integration are specified. Procurement requirements show intended scope, not completed deployment.\"], \"original_id\": \"ongc_tender\"}], \"field_sources\": {\"business_problem\": [\"RS029\"], \"workflow\": [\"RS029\"], \"technology_data\": [\"RS029\"], \"integration_human_role\": [\"RS029\"], \"deployment_scope\": [\"RS029\"], \"economics\": [\"RS029\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"related_existing_ids\": [], \"cohort\": \"nocs\", \"packet_id\": \"nocs-016\", \"source_ids\": [\"RS029\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-152\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-152", "company": "ONGC", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207"], "findings": [{"severity": "low", "field": "stage", "issue": "The official tender specifies AI-driven optimization/predictive maintenance and gas-lift adjustment using SCADA/HMI with safe limits, but is an EOI/pilot solicitation rather than proof of awarded, commissioned deployment.", "correction": "Keep stage as proposed pilot / market exploration; do not represent as live operating AI.", "source_url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207"}], "residual_gaps": ["Tender source defines requirements and solicitation intent; no award, commissioning, or measured results located in this packet review."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-153", "Company": "Occidental", "Company_primary": "Occidental", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "AI Center of Excellence and 2025 project portfolio", "Extracted_finding": "The proxy reports 20 AI projects developed in 2025. Separately, it describes reservoir-characterization AI tools integrated with geoscience platforms; their inclusion in the count and link to the Center of Excellence are not specified.", "Stage_at_disclosure": "At year-end 2025: 14 projects deployed, four piloting and two under development.", "Reported_metric": "20 projects; each assigned $10–90 million in estimated annual benefit. No realized portfolio total disclosed.", "Interpretation_limit": "“Deployed” is not defined by asset count or rollout breadth. Incentive objectives refer to realized value, but the supporting results paragraph uses estimated benefits.", "Source_IDs": "RS035; SW2-004", "Source_date": "2026-03-19", "Source_type": "Issuer proxy statement", "Source_locator": "Compensation Discussion and Analysis, footnote (4), printed page 43 (PDF index 42)", "Source_URL": "https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf", "Evidence_status": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "Proxy furnished 19 March 2026; underlying project status at 2025 year-end.", "Context": "Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.", "Implementation": "The Compensation Committee discusses AI alongside technology objectives. It reports improved reservoir predictions in blind testing without publishing the test design or accuracy gain.", "Assessment": "Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "north_america-005-1", "Initiative_ID": "AI-153", "Title": "AI Center of Excellence and cross-functional operating use cases", "Date": "2025-02-19", "Finding": "CEO named AI Center of Excellence and Gulf of America use areas.", "Metric": "", "Source_IDs": "RS035", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "SW2-D-AI-153", "Initiative_ID": "AI-153", "Title": "AI Center of Excellence and 2025 project portfolio", "Date": "2026-03-19", "Finding": "The proxy reports 20 AI projects developed in 2025. Separately, it describes reservoir-characterization AI tools integrated with geoscience platforms; their inclusion in the count and link to the Center of Excellence are not specified.", "Metric": "20 projects; each assigned $10–90 million in estimated annual benefit. No realized portfolio total disclosed.", "Source_IDs": "SW2-004", "Evidence": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2026-03-19", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-153", "north_america-005"], "Research_search": "{\"key\": \"swarm2-ai-153\", \"packet_id\": \"swarm2-ai-153\", \"record_ids\": [\"AI-153\"], \"existing_ids\": [\"AI-153\"], \"related_existing_ids\": [], \"company\": \"Occidental\", \"title\": \"AI Center of Excellence and 2025 project portfolio\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [\"Reservoir engineering & field development\"], \"category\": \"Internal adoption\", \"stage\": \"At year-end 2025: 14 projects deployed, four piloting and two under development.\", \"practical_relevance\": \"Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.\", \"business_problem\": \"\", \"workflow\": \"The proxy reports 20 AI projects developed in 2025. Separately, it describes reservoir-characterization AI tools integrated with geoscience platforms; their inclusion in the count and link to the Center of Excellence are not specified.\", \"technology_data\": \"The Compensation Committee discusses AI alongside technology objectives. It reports improved reservoir predictions in blind testing without publishing the test design or accuracy gain.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"At year-end 2025: 14 projects deployed, four piloting and two under development.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"20 projects; each assigned $10–90 million in estimated annual benefit. No realized portfolio total disclosed.\", \"period\": \"2025 year-end status\", \"attribution\": \"Issuer proxy disclosure; estimated economic benefits, not independently established returns.\", \"source_ids\": [\"SW2-004\"]}], \"economics\": \"Estimates are not net cash returns. Deployment costs, overlap between benefits and project-level realization remain undisclosed; do not sum the range into a portfolio valuation.\", \"timeline\": [{\"date\": \"2026-03-19\", \"event\": \"Proxy furnished 19 March 2026; underlying project status at 2025 year-end.\", \"source_ids\": [\"SW2-004\"]}], \"limitations\": [\"“Deployed” is not defined by asset count or rollout breadth. Incentive objectives refer to realized value, but the supporting results paragraph uses estimated benefits.\"], \"assessment\": \"Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.\", \"sources\": [{\"title\": \"2026 Oxy Proxy Statement\", \"url\": \"https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf\", \"date\": \"2026-03-19\", \"type\": \"Issuer proxy statement\", \"locator\": \"Compensation Discussion and Analysis, footnote (4), printed page 43 (PDF index 42)\", \"review_status\": \"Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.\", \"publisher\": \"Occidental\", \"id\": \"SW2-004\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-004\"], \"field_sources\": {\"workflow\": [\"SW2-004\"], \"technology_data\": [\"SW2-004\"], \"deployment_scope\": [\"SW2-004\"], \"practical_relevance\": [\"SW2-004\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Issuer proxy disclosure; estimated economic benefits, not independently established returns.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-153", "company": "Occidental", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The disclosed operating use cases are Gulf of America operations (supply chain, asset integrity and reservoir characterization), which are upstream-centered; the AI Center of Excellence is company-wide. Keep Upstream as primary workflow but tag the record cross-functional or split the CoE from use cases.", "Segment_classification_confidence": "medium", "Asset_location": "Company-wide; asset-level scope undisclosed", "Followup_review": {"date": "2026-09-29", "decision": "Enriched", "note": "Deduplicated, original-source checked and separately reviewed."}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "USD 10–90m per project is an estimated annual benefit assigned in the proxy, not a realized result; no realized portfolio total disclosed.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-154", "Company": "Suncor Energy", "Company_primary": "Suncor Energy", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "Dynamic AI dispatch assigns mine haul trucks", "Extracted_finding": "A named AI system automates routine truck assignments while keeping dispatchers active to improve route and ore-blend decisions.", "Stage_at_disclosure": "Operating system described on an undated company page; initial deployment date not disclosed.", "Reported_metric": "Routine dispatch functions: AI assigns dumps/refueling and reroutes around slowdowns; human dispatchers retained (Company-reported)", "Interpretation_limit": "Source page is undated; “recently implemented” has no dated reference point. No named vendor, system architecture, AI method, user count or quantitative outcome disclosed. This workflow is distinct from the autonomous haulage fleet in AI-099.", "Source_IDs": "RS041", "Source_date": "Undated source; see chronology", "Source_type": "company announcement", "Source_locator": "AI dispatch system; human dispatcher workflow", "Source_URL": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Mine dispatchers must continuously route haul trucks among shovels, dumps and refueling while responding to slowdowns.", "Implementation": "Suncor says dynamic dispatch assigns trucks to dump stations, sends them to refueling and reroutes around slowdowns; dispatchers then review mine maps and optimize routes and blends.", "Assessment": "This is a practical AI decision-support/assignment workflow, not evidence that haulage fleet autonomy and dispatch are one system. Human dispatchers remain part of operations.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "north_america-022-1", "Initiative_ID": "AI-154", "Title": "Dynamic AI dispatch assigns mine haul trucks", "Date": "undated", "Finding": "Company page describes dynamic dispatch as recently implemented; publication and initial deployment dates are not stated.", "Metric": "", "Source_IDs": "RS041", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["north_america-022"], "Research_search": " {\"key\": \"suncor-dynamic-dispatch-ai\", \"existing_ids\": [], \"company\": \"Suncor Energy\", \"title\": \"Dynamic AI dispatch assigns mine haul trucks\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [\"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"Operating system described on an undated company page; initial deployment date not disclosed.\", \"practical_relevance\": \"A named AI system automates routine truck assignments while keeping dispatchers active to improve route and ore-blend decisions.\", \"business_problem\": \"Mine dispatchers must continuously route haul trucks among shovels, dumps and refueling while responding to slowdowns.\", \"workflow\": \"Suncor says dynamic dispatch assigns trucks to dump stations, sends them to refueling and reroutes around slowdowns; dispatchers then review mine maps and optimize routes and blends.\", \"technology_data\": \"Suncor identifies an AI dispatch system; model, optimization method and input data are not disclosed.\", \"integration_human_role\": \"Human dispatchers remain on shift, monitor routes and use the system’s assignments to improve mine routing and blends.\", \"deployment_scope\": \"Suncor describes implementation at Syncrude’s Mildred Lake site; truck count, start date and broader site coverage are not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Routine dispatch functions\", \"value\": \"AI assigns dumps/refueling and reroutes around slowdowns; human dispatchers retained\", \"period\": \"undated company page\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS041\"]}], \"economics\": \"No dispatch-specific production, cost or revenue outcome is quantified.\", \"timeline\": [{\"date\": \"undated\", \"event\": \"Company page describes dynamic dispatch as recently implemented; publication and initial deployment dates are not stated.\", \"source_ids\": [\"RS041\"]}], \"limitations\": [\"Source page is undated; “recently implemented” has no dated reference point.\", \"No named vendor, system architecture, AI method, user count or quantitative outcome disclosed.\", \"This workflow is distinct from the autonomous haulage fleet in AI-099.\"], \"assessment\": \"This is a practical AI decision-support/assignment workflow, not evidence that haulage fleet autonomy and dispatch are one system. Human dispatchers remain part of operations.\", \"sources\": [{\"id\": \"RS041\", \"title\": \"Dispatching innovation: Suncor's new dispatch system improves mine operations\", \"url\": \"https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations\", \"publisher\": \"Suncor\", \"date\": null, \"type\": \"company announcement\", \"locator\": \"AI dispatch system; human dispatcher workflow\", \"review_status\": \"full text reviewed\", \"supports\": [\"Dynamic dispatch automates dump/refueling assignments and slowdown rerouting; human dispatchers remain active.\"], \"accessed\": \"2026-09-27\", \"publication_date_status\": \"undated page; accessed 2026-09-27\", \"original_id\": \"na-008\"}], \"field_sources\": {\"business_problem\": [\"RS041\"], \"workflow\": [\"RS041\"], \"technology_data\": [\"RS041\"], \"integration_human_role\": [\"RS041\"], \"deployment_scope\": [\"RS041\"], \"economics\": [\"RS041\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"New distinct workflow identified in Suncor source; related to, but not the same initiative as, AI-099.\", \"related_existing_ids\": [\"AI-099\"], \"cohort\": \"north_america\", \"packet_id\": \"north_america-022\", \"source_ids\": [\"RS041\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-154\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-154", "company": "Suncor Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-155", "Company": "Petoro", "Company_primary": "Petoro", "Segment": "Upstream", "Theme": "AI products", "Initiative": "Petoro AI well-data extraction and field-development screening", "Extracted_finding": "A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.", "Stage_at_disclosure": "Pilot/prototype workflow and customer validation reported June 2026", "Reported_metric": "well schematic quality-control throughput: 2 to 6-7 schematics/day (Vendor-reported); productivity improvement: 3x (Vendor-reported)", "Interpretation_limit": "Vendor-authored case study; Petoro 2025 report separately corroborates reservoir-complexity tool in Delfi, but not the specific GenAI schematic agent or productivity numbers.", "Source_IDs": "RS044; RS045", "Source_date": "2026-06-05", "Source_type": "partner case study", "Source_locator": "Objectives, workflow, impact and deployment sections", "Source_URL": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-06-05; see chronology for operating dates.", "Context": "Review a large, fragmented set of well records and sidetrack candidates to improve field-development choices in mature Norwegian fields.", "Implementation": "Well Schematic Extractor/Well History agents extract casing, completion and formation facts from reports/logs; deterministic checks plus visual QC help staff verify output. FDPlan and agent workflows aid opportunity selection and drilling portfolio optimization.", "Assessment": "Analyst inference: A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-003-1", "Initiative_ID": "AI-155", "Title": "Petoro AI well-data extraction and field-development screening", "Date": "2025-10-17", "Finding": "SLB Q3 earnings says Petoro awarded generative AI field-planning solution", "Metric": "", "Source_IDs": "RS045", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "services-003-2", "Initiative_ID": "AI-155", "Title": "Petoro AI well-data extraction and field-development screening", "Date": "2026-06-05", "Finding": "Detailed Petoro well-data extraction and preliminary productivity published", "Metric": "", "Source_IDs": "RS044", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2026-06-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-003"], "Research_search": " {\"key\": \"slb-petoro\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Petoro\", \"title\": \"Petoro AI well-data extraction and field-development screening\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Pilot/prototype workflow and customer validation reported June 2026\", \"practical_relevance\": \"A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.\", \"business_problem\": \"Review a large, fragmented set of well records and sidetrack candidates to improve field-development choices in mature Norwegian fields.\", \"workflow\": \"Well Schematic Extractor/Well History agents extract casing, completion and formation facts from reports/logs; deterministic checks plus visual QC help staff verify output. FDPlan and agent workflows aid opportunity selection and drilling portfolio optimization.\", \"technology_data\": \"Lumi AI workspace, LLMs, domain context, FDPlan, Delfi and OSDU-aligned data foundation.\", \"integration_human_role\": \"Petoro analysts validate extraction via visual QC and use insights for plans; user decisions remain with engineers/planners.\", \"deployment_scope\": \"Petoro NCS assets; field/schematic coverage is not quantified.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"well schematic quality-control throughput\", \"value\": \"2 to 6-7 schematics/day\", \"period\": \"preliminary user testing reported Jun 2026\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS044\"]}, {\"metric\": \"productivity improvement\", \"value\": \"3x\", \"period\": \"2026 case study\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS044\"]}], \"economics\": \"SLB reports 3x productivity; preliminary QC throughput went from 2 schematics/day to 6-7/day. Vendor case-study figures; no economics disclosed.\", \"timeline\": [{\"date\": \"2025-10-17\", \"event\": \"SLB Q3 earnings says Petoro awarded generative AI field-planning solution\", \"source_ids\": [\"RS045\"]}, {\"date\": \"2026-06-05\", \"event\": \"Detailed Petoro well-data extraction and preliminary productivity published\", \"source_ids\": [\"RS044\"]}], \"limitations\": [\"Vendor-authored case study; Petoro 2025 report separately corroborates reservoir-complexity tool in Delfi, but not the specific GenAI schematic agent or productivity numbers.\"], \"assessment\": \"Analyst inference: A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.\", \"sources\": [{\"id\": \"RS044\", \"title\": \"Petoro and SLB: Pioneering AI-driven well planning on the Norwegian continental shelf\", \"url\": \"https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning\", \"publisher\": \"SLB\", \"date\": \"2026-06-05\", \"type\": \"partner case study\", \"locator\": \"Objectives, workflow, impact and deployment sections\", \"review_status\": \"full text reviewed\", \"supports\": [\"Petoro and SLB built Lumi/Innovation Factori workflows to extract data from well schematics/reports and rank field-development/sidetrack opportunities.\", \"Well Schematic Extractor and Well History use LLMs to extract casing, completion and formation information, with deterministic checks and user visual QC.\", \"SLB reports threefold productivity improvement; preliminary QC rose from two schematics/day to six or seven. This is vendor case-study reporting.\"], \"original_id\": \"petoro\"}, {\"id\": \"RS045\", \"title\": \"SLB Announces Third-Quarter 2025 Results\", \"url\": \"https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609\", \"publisher\": \"SLB\", \"date\": \"2025-10-17\", \"type\": \"earnings release\", \"locator\": \"Digital highlights / Petoro award\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Q3 earnings: Petoro awarded a generative-AI solution to optimize drilling resources using FDPlan and Lumi.\"], \"original_id\": \"slb_q3\"}], \"field_sources\": {\"business_problem\": [\"RS044\", \"RS045\"], \"workflow\": [\"RS044\", \"RS045\"], \"technology_data\": [\"RS044\", \"RS045\"], \"integration_human_role\": [\"RS044\"], \"deployment_scope\": [\"RS044\", \"RS045\"], \"economics\": [\"RS044\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-003\", \"source_ids\": [\"RS044\", \"RS045\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-155\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-155", "company": "Petoro", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-156", "Company": "ONGC", "Company_primary": "ONGC", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "DrillOps intelligent well delivery across ONGC rig fleet", "Extracted_finding": "This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.", "Stage_at_disclosure": "Three-year enterprise contract; rollout scope up to 120 rigs announced July 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Vendor-announced contract; no ONGC announcement or measured outcomes found. AI contribution is not explicit in the deployment description.", "Source_IDs": "S27", "Source_date": "2026-07-24", "Source_type": "earnings release", "Source_locator": "Quarterly highlights, contract awards and technology highlights", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-07-24; see chronology for operating dates.", "Context": "Coordinate large drilling fleet.", "Implementation": "DrillOps monitoring, predictive analytics, central review.", "Assessment": "Analyst inference: This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-008-1", "Initiative_ID": "AI-156", "Title": "DrillOps intelligent well delivery across ONGC rig fleet", "Date": "2026-07-24", "Finding": "Three-year contract announced; up to 120 rigs in scope", "Metric": "", "Source_IDs": "S27", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2026-07-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-008"], "Research_search": " {\"key\": \"slb-ongc-drillops\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"ONGC\", \"title\": \"DrillOps intelligent well delivery across ONGC rig fleet\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Three-year enterprise contract; rollout scope up to 120 rigs announced July 2026\", \"practical_relevance\": \"This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.\", \"business_problem\": \"Coordinate large drilling fleet.\", \"workflow\": \"DrillOps monitoring, predictive analytics, central review.\", \"technology_data\": \"Models/sensors absent.\", \"integration_human_role\": \"Central review; no autonomous control claim.\", \"deployment_scope\": \"Three years; up to 120 rigs.\", \"partners\": [\"SLB\"], \"disclosed_results\": [], \"economics\": \"Value/results absent.\", \"timeline\": [{\"date\": \"2026-07-24\", \"event\": \"Three-year contract announced; up to 120 rigs in scope\", \"source_ids\": [\"S27\"]}], \"limitations\": [\"Vendor-announced contract; no ONGC announcement or measured outcomes found. AI contribution is not explicit in the deployment description.\"], \"assessment\": \"Analyst inference: This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.\", \"sources\": [{\"id\": \"S27\", \"title\": \"SLB Announces Second-Quarter 2026 Results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results\", \"publisher\": \"SLB\", \"date\": \"2026-07-24\", \"type\": \"earnings release\", \"locator\": \"Quarterly highlights, contract awards and technology highlights\", \"review_status\": \"full text reviewed\", \"supports\": [\"Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs.\"], \"original_id\": \"slb_q2\"}], \"field_sources\": {\"business_problem\": [\"S27\"], \"workflow\": [\"S27\"], \"technology_data\": [\"S27\"], \"integration_human_role\": [\"S27\"], \"deployment_scope\": [\"S27\"], \"economics\": [\"S27\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-008\", \"source_ids\": [\"S27\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-156\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-156", "company": "ONGC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-157", "Company": "Halliburton", "Company_primary": "Halliburton", "Segment": "Upstream services & technology", "Theme": "Automation context", "Initiative": "Automated on-bottom directional drilling for Equinor", "Extracted_finding": "Named live-well deployment and three-system integration make this a concrete autonomous-drilling case. Commercial scaling evidence is not in the source.", "Stage_at_disclosure": "One North Sea Equinor well delivered Feb 2025", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company-reported or vendor case study; no independent validation identified in reviewed sources. Source describes automated/autonomous drilling; AI/ML method not disclosed.", "Source_IDs": "RS050", "Source_date": "2025-02-26", "Source_type": "company announcement", "Source_locator": "Deployment description", "Source_URL": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-02-26; see chronology for operating dates.", "Context": "Coordinate downhole, hydraulics and surface rig controls to optimize well placement and drilling parameters in real time.", "Implementation": "Integrated LOGIX remote automation, Sekal Drilltronics and rig controls perform autonomous directional drilling with automated hydraulics and dynamic surface equipment control.", "Assessment": "Analyst inference: The one-well integrated deployment is a material drilling-automation case; it should not be counted as AI without explicit AI evidence.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Automation or digital workflow; explicit AI basis not established", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-013-1", "Initiative_ID": "AI-157", "Title": "Automated on-bottom directional drilling for Equinor", "Date": "2025-02-26", "Finding": "Halliburton and Sekal reported delivery of an Equinor well with integrated automated on-bottom drilling.", "Metric": "", "Source_IDs": "RS050", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2025-02-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-013"], "Research_search": " {\"key\": \"halliburton-equinor-onbottom\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Halliburton\", \"title\": \"Automated on-bottom directional drilling for Equinor\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Automation / AI not established\", \"stage\": \"One North Sea Equinor well delivered Feb 2025\", \"practical_relevance\": \"Named live-well deployment and three-system integration make this a concrete autonomous-drilling case. Commercial scaling evidence is not in the source.\", \"business_problem\": \"Coordinate downhole, hydraulics and surface rig controls to optimize well placement and drilling parameters in real time.\", \"workflow\": \"Integrated LOGIX remote automation, Sekal Drilltronics and rig controls perform autonomous directional drilling with automated hydraulics and dynamic surface equipment control.\", \"technology_data\": \"Closed-loop models for subsurface, wellbore fluid and pressure systems integrated with rig automation.\", \"integration_human_role\": \"A one-push initiation is described; safe limits cited. Human override/monitoring architecture not detailed.\", \"deployment_scope\": \"One Equinor well on Norwegian Continental Shelf; future scale not disclosed.\", \"partners\": [\"Equinor\", \"Sekal\"], \"disclosed_results\": [], \"economics\": \"No ROP/time/cost measure disclosed.\", \"timeline\": [{\"date\": \"2025-02-26\", \"event\": \"Halliburton and Sekal reported delivery of an Equinor well with integrated automated on-bottom drilling.\", \"source_ids\": [\"RS050\"]}], \"limitations\": [\"Company-reported or vendor case study; no independent validation identified in reviewed sources.\", \"Source describes automated/autonomous drilling; AI/ML method not disclosed.\"], \"assessment\": \"Analyst inference: The one-well integrated deployment is a material drilling-automation case; it should not be counted as AI without explicit AI evidence.\", \"sources\": [{\"id\": \"RS050\", \"title\": \"Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea\", \"url\": \"https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom\", \"publisher\": \"Halliburton\", \"date\": \"2025-02-26\", \"type\": \"company announcement\", \"locator\": \"Deployment description\", \"review_status\": \"full text reviewed\", \"supports\": [\"A well for Equinor on the Norwegian Continental Shelf used LOGIX, Drilltronics and rig control for autonomous directional drilling, automated wellbore hydraulics and surface equipment control.\", \"Integrated models optimize drilling parameters and well placement in real time; a single action initiates the workflow.\"], \"original_id\": \"RS050\"}], \"field_sources\": {\"business_problem\": [\"RS050\"], \"workflow\": [\"RS050\"], \"technology_data\": [\"RS050\"], \"integration_human_role\": [\"RS050\"], \"deployment_scope\": [\"RS050\"], \"economics\": [\"RS050\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-013\", \"source_ids\": [\"RS050\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-157\"]}", "Research_status": "Expanded source review", "Research_category": "Automation / AI not established", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-157", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton/Sekal source establishes automated on-bottom drilling, control integration and autonomy, without AI/ML identification.", "correction": "Classify as drilling automation; remove it from AI adoption counts absent explicit AI evidence.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom"}], "residual_gaps": ["No AI/ML method identified in the primary deployment report or cited product material."], "partition": 2, "reviewer_outcome": "corrected"}, "Audit_outcome": "corrected", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "not_ai", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-158", "Company": "NOV", "Company_primary": "NOV", "Segment": "Upstream services & technology", "Theme": "Operational AI", "Initiative": "Drilling Beliefs & Analytics anomaly warning in Egypt", "Extracted_finding": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.", "Stage_at_disclosure": "Vendor reports expansion from two rigs to a double-digit fleet", "Reported_metric": "NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.", "Interpretation_limit": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.", "Source_IDs": "RS053; WATCH14PM-NOV", "Source_date": "2026-03-11", "Source_type": "Vendor deployment article", "Source_locator": "11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example", "Source_URL": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Article published 11 March 2026. Egypt introduction dates to 2023; the source does not date each intervention or disclose exact rig/well counts.", "Context": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.", "Implementation": "Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "Assessment": "How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-016-1", "Initiative_ID": "AI-158", "Title": "Drilling Beliefs & Analytics anomaly warning in Egypt", "Date": "2025-01", "Finding": "Bit Balling Belief alert prompted crew response in the Egypt example.", "Metric": "", "Source_IDs": "RS053", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "services-016-2", "Initiative_ID": "AI-158", "Title": "Drilling Beliefs & Analytics anomaly warning in Egypt", "Date": "2025-06-27", "Finding": "NOV published details of the case and Egypt remote operations center.", "Metric": "", "Source_IDs": "RS053", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "WATCH14PM-AI-158", "Initiative_ID": "AI-158", "Title": "Egyptian drilling teams expand an edge-based warning system", "Date": "2026-03-11", "Finding": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings. Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "Metric": "NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.", "Source_IDs": "WATCH14PM-NOV", "Source_URL": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established."}], "Latest_disclosure": "2026-03-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-016", "watch14pm-ai-158"], "Research_search": "{\"packet_id\": \"watch14pm-ai-158\", \"key\": \"watch14pm-ai-158\", \"record_ids\": [\"AI-158\"], \"company\": \"NOV\", \"title\": \"Egyptian drilling teams expand an edge-based warning system\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Vendor reports expansion from two rigs to a double-digit fleet\", \"business_problem\": \"NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.\", \"workflow\": \"Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.\", \"technology_data\": \"Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.\", \"deployment_scope\": \"Article published 11 March 2026. Egypt introduction dates to 2023; the source does not date each intervention or disclose exact rig/well counts.\", \"integration_human_role\": \"Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.\", \"economics\": \"NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.\", \"practical_relevance\": \"How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?\", \"assessment\": \"How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?\", \"limitations\": [\"Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.\"], \"source_ids\": [\"WATCH14PM-NOV\"], \"sources\": [{\"id\": \"WATCH14PM-NOV\", \"title\": \"NOV’s Drilling Beliefs & Analytics advances digital operations in Egypt\", \"url\": \"https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt\", \"date\": \"2026-03-11\", \"type\": \"Vendor deployment article\", \"publisher\": \"NOV\", \"locator\": \"11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-03-11\", \"event\": \"NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.\", \"source_ids\": [\"WATCH14PM-NOV\"]}], \"field_sources\": {\"business_problem\": [\"WATCH14PM-NOV\"], \"workflow\": [\"WATCH14PM-NOV\"], \"technology_data\": [\"WATCH14PM-NOV\"], \"deployment_scope\": [\"WATCH14PM-NOV\"], \"economics\": [\"WATCH14PM-NOV\"]}}", "Research_status": "Reviewed 14 October 2026 · evening watch", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-158", "company": "NOV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "NOV Egypt source primarily supports remote operations and operational performance; the reopened announcement is vendor authored and AI method/effect must not be assumed.", "correction": "Separate remote-operation achievement from any unproven AI analytics attribution.", "source_url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-14", "Research_origin": ["research/watch-2026-10-14-pm/README.md"]}, {"ID": "AI-159", "Company": "Precision Drilling", "Company_primary": "Precision Drilling", "Segment": "Upstream services & technology", "Theme": "Operational AI", "Initiative": "Alpha drilling analytics and predictive maintenance", "Extracted_finding": "This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.", "Stage_at_disclosure": "78 rigs equipped with Alpha digital technology at end-2024; report published Apr 2025", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company-reported or vendor case study; no independent validation identified in reviewed sources.", "Source_IDs": "RS059", "Source_date": "2025-04", "Source_type": "sustainability report", "Source_locator": "Alpha Suite technologies", "Source_URL": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "Evidence_status": "Audit could not verify the cited claims", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-04; see chronology for operating dates.", "Context": "Use rig/drilling data to improve performance, adapt drilling in real time and plan equipment maintenance before failures.", "Implementation": "Alpha suite aggregates drilling, equipment, geology and environmental data; AI algorithms identify patterns/trends, adjust operations and support predictive maintenance scheduling.", "Assessment": "Analyst inference: This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-022-1", "Initiative_ID": "AI-159", "Title": "Alpha drilling analytics and predictive maintenance", "Date": "2024-12-31", "Finding": "Precision reports Alpha digital technology deployed on 78 rigs.", "Metric": "", "Source_IDs": "RS059", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "services-022-2", "Initiative_ID": "AI-159", "Title": "Alpha drilling analytics and predictive maintenance", "Date": "2025-04", "Finding": "2024 ESG report published.", "Metric": "", "Source_IDs": "RS059", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-022"], "Research_search": " {\"key\": \"precision-alpha-ai\", \"existing_ids\": [], \"related_existing_ids\": [\"2024 status baseline inside window; ESG report is company-reported and no AI-attributed metric or named asset case was identified.\"], \"company\": \"Precision Drilling\", \"title\": \"Alpha drilling analytics and predictive maintenance\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"78 rigs equipped with Alpha digital technology at end-2024; report published Apr 2025\", \"practical_relevance\": \"This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.\", \"business_problem\": \"Use rig/drilling data to improve performance, adapt drilling in real time and plan equipment maintenance before failures.\", \"workflow\": \"Alpha suite aggregates drilling, equipment, geology and environmental data; AI algorithms identify patterns/trends, adjust operations and support predictive maintenance scheduling.\", \"technology_data\": \"Digital automation and analytics suite with AI; vendor detail/model/connection architecture not in report.\", \"integration_human_role\": \"Drilling staff use recommendations and automation; independent impact or autonomy boundary not detailed.\", \"deployment_scope\": \"78 rigs in North America at end-2024.\", \"partners\": [\"Internal Precision Drilling teams\"], \"disclosed_results\": [], \"economics\": \"No quantified AI-only uptime, downtime, cost or performance results.\", \"timeline\": [{\"date\": \"2024-12-31\", \"event\": \"Precision reports Alpha digital technology deployed on 78 rigs.\", \"source_ids\": [\"RS059\"]}, {\"date\": \"2025-04\", \"event\": \"2024 ESG report published.\", \"source_ids\": [\"RS059\"]}], \"limitations\": [\"Company-reported or vendor case study; no independent validation identified in reviewed sources.\"], \"assessment\": \"Analyst inference: This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.\", \"sources\": [{\"id\": \"RS059\", \"title\": \"Precision Drilling 2024 ESG Report\", \"url\": \"https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf\", \"publisher\": \"Precision Drilling\", \"date\": \"2025-04\", \"type\": \"sustainability report\", \"locator\": \"Alpha Suite technologies\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Precision reports 78 rigs with Alpha digital technology at end-2024.\", \"Alpha suite combines digital automation and analytics with AI algorithms for drilling performance pattern recognition, real-time adjustments and predictive maintenance; report does not break out AI-specific results.\"], \"original_id\": \"precision\"}], \"field_sources\": {\"business_problem\": [\"RS059\"], \"workflow\": [\"RS059\"], \"technology_data\": [\"RS059\"], \"integration_human_role\": [\"RS059\"], \"deployment_scope\": [\"RS059\"], \"economics\": [\"RS059\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-022\", \"source_ids\": [\"RS059\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-159\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "unable to verify", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-159", "company": "Precision Drilling", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, "Audit_outcome": "unverified", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "scaled", "Scale_note": "78 rigs", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and the Alpha/AI passages were located in the source. Audit outcome 'unverified' retained pending a full re-read; grade stays D.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "unable to verify", "claim_status": "Audit could not verify the cited claims", "independent_validation": "Not established by this collection"}}, {"ID": "AI-160", "Company": "Saipem", "Company_primary": "Saipem", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "AI Project Schedule Onshore for EPC risk forecasting", "Extracted_finding": "Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.", "Stage_at_disclosure": "AI engineering initiative launched 2024; 13 use cases explored by 2025", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company-reported or vendor case study; no independent validation identified in reviewed sources.", "Source_IDs": "RS060", "Source_date": "2026-04", "Source_type": "sustainability report", "Source_locator": "Harnessing AI for engineering section", "Source_URL": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-04; see chronology for operating dates.", "Context": "Anticipate construction delays and bottlenecks and improve EPC project planning decisions using prior project data.", "Implementation": "Project Schedule Onshore analyzes historical project data with predictive models to estimate construction risk and delays for engineering/procurement/construction planning.", "Assessment": "Analyst inference: Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-023-1", "Initiative_ID": "AI-160", "Title": "AI Project Schedule Onshore for EPC risk forecasting", "Date": "2024", "Finding": "Saipem launched its two-year AI for Engineering initiative.", "Metric": "", "Source_IDs": "RS060", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "services-023-2", "Initiative_ID": "AI-160", "Title": "AI Project Schedule Onshore for EPC risk forecasting", "Date": "2025", "Finding": "Saipem named 13 use cases and AI Project Schedule Onshore.", "Metric": "", "Source_IDs": "RS060", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-023"], "Research_search": " {\"key\": \"saipem-ai-schedule\", \"existing_ids\": [], \"related_existing_ids\": [\"Sustainability report excerpt reviewed; no customer/project-specific result identified.\"], \"company\": \"Saipem\", \"title\": \"AI Project Schedule Onshore for EPC risk forecasting\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"AI engineering initiative launched 2024; 13 use cases explored by 2025\", \"practical_relevance\": \"Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.\", \"business_problem\": \"Anticipate construction delays and bottlenecks and improve EPC project planning decisions using prior project data.\", \"workflow\": \"Project Schedule Onshore analyzes historical project data with predictive models to estimate construction risk and delays for engineering/procurement/construction planning.\", \"technology_data\": \"Predictive models on historical project data; features, training set and validation metrics not disclosed.\", \"integration_human_role\": \"Planning teams use risk signals to make faster decisions; not autonomous schedule control.\", \"deployment_scope\": \"Saipem engineering/projects; number of projects or teams using the tool not disclosed.\", \"partners\": [\"Internal Saipem teams\"], \"disclosed_results\": [], \"economics\": \"No quantified schedule accuracy or financial return disclosed.\", \"timeline\": [{\"date\": \"2024\", \"event\": \"Saipem launched its two-year AI for Engineering initiative.\", \"source_ids\": [\"RS060\"]}, {\"date\": \"2025\", \"event\": \"Saipem named 13 use cases and AI Project Schedule Onshore.\", \"source_ids\": [\"RS060\"]}], \"limitations\": [\"Company-reported or vendor case study; no independent validation identified in reviewed sources.\"], \"assessment\": \"Analyst inference: Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.\", \"sources\": [{\"id\": \"RS060\", \"title\": \"Saipem Sustainability Overview 2025\", \"url\": \"https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf\", \"publisher\": \"Saipem\", \"date\": \"2026-04\", \"type\": \"sustainability report\", \"locator\": \"Harnessing AI for engineering section\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Saipem’s AI for Engineering is a two-year cross-functional effort launched 2024; 13 use cases explored by 2025.\", \"AI Project Schedule Onshore analyzes historical project data with predictive models for EPC construction risks/delays and bottlenecks; 12 additional apps were planned for 2026.\"], \"original_id\": \"saipem\"}], \"field_sources\": {\"business_problem\": [\"RS060\"], \"workflow\": [\"RS060\"], \"technology_data\": [\"RS060\"], \"integration_human_role\": [\"RS060\"], \"deployment_scope\": [\"RS060\"], \"economics\": [\"RS060\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-023\", \"source_ids\": [\"RS060\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-160\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-160", "company": "Saipem", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-161", "Company": "Transocean", "Company_primary": "Transocean", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "AI tools for office and administrative productivity", "Extracted_finding": "This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.", "Stage_at_disclosure": "AI tools introduced by 2025; no use in offshore drilling stated", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company-reported or vendor case study; no independent validation identified in reviewed sources.", "Source_IDs": "RS061", "Source_date": "2026-02-19", "Source_type": "annual report", "Source_locator": "Technology and AI discussion", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-02-19; see chronology for operating dates.", "Context": "Reduce effort in high-volume office and administrative tasks.", "Implementation": "Company says it has introduced AI tools for unspecified high-volume administrative tasks and is exploring more productivity use cases.", "Assessment": "Analyst inference: This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-024-1", "Initiative_ID": "AI-161", "Title": "AI tools for office and administrative productivity", "Date": "2026-02-19", "Finding": "2025 Form 10-K reports AI use in office tasks and no AI use in offshore drilling operations.", "Metric": "", "Source_IDs": "RS061", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-02-19", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-024"], "Research_search": " {\"key\": \"transocean-office-ai\", \"existing_ids\": [], \"related_existing_ids\": [\"Company explicitly limits current AI use to office/admin tasks; operational use is being explored, not deployed.\"], \"company\": \"Transocean\", \"title\": \"AI tools for office and administrative productivity\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"AI tools introduced by 2025; no use in offshore drilling stated\", \"practical_relevance\": \"This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.\", \"business_problem\": \"Reduce effort in high-volume office and administrative tasks.\", \"workflow\": \"Company says it has introduced AI tools for unspecified high-volume administrative tasks and is exploring more productivity use cases.\", \"technology_data\": \"AI tool types and workflow data are not disclosed; governance committee oversees tool selection, security, training and usage.\", \"integration_human_role\": \"Corporate users; human decision processes implied but details not disclosed.\", \"deployment_scope\": \"Office/administrative workflows only; Transocean explicitly says AI is not currently used in offshore drilling operations.\", \"partners\": [\"Internal Transocean teams\"], \"disclosed_results\": [], \"economics\": \"No productivity, time or cost metric disclosed.\", \"timeline\": [{\"date\": \"2026-02-19\", \"event\": \"2025 Form 10-K reports AI use in office tasks and no AI use in offshore drilling operations.\", \"source_ids\": [\"RS061\"]}], \"limitations\": [\"Company-reported or vendor case study; no independent validation identified in reviewed sources.\"], \"assessment\": \"Analyst inference: This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.\", \"sources\": [{\"id\": \"RS061\", \"title\": \"Transocean Ltd. 2025 Form 10-K\", \"url\": \"https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm\", \"publisher\": \"Transocean\", \"date\": \"2026-02-19\", \"type\": \"annual report\", \"locator\": \"Technology and AI discussion\", \"review_status\": \"full text reviewed\", \"supports\": [\"Transocean explicitly says AI is not currently used in offshore drilling operations; it has introduced AI tools for high-volume office/administrative tasks and is exploring productivity use cases.\", \"Company describes AI governance committee; fleet automation/robotics systems are separate from AI.\"], \"original_id\": \"transocean\"}], \"field_sources\": {\"business_problem\": [\"RS061\"], \"workflow\": [\"RS061\"], \"technology_data\": [\"RS061\"], \"integration_human_role\": [\"RS061\"], \"deployment_scope\": [\"RS061\"], \"economics\": [\"RS061\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"services\", \"packet_id\": \"services-024\", \"source_ids\": [\"RS061\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-161\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-161", "company": "Transocean", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The only disclosed workflow is generic office and administrative productivity. Transocean’s identity as an offshore drilling contractor does not make this workflow an upstream-services operation.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. SEC EDGAR blocks automated retrieval; not re-read on 5 Oct 2026.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-162", "Company": "Kuwait Oil Company", "Company_primary": "Kuwait Oil Company", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "Ahmadi Innovation Valley seven-year technology evaluation and deployment contract", "Extracted_finding": "The program creates an evaluation-to-deployment channel spanning nearly 100 initiatives, but the source does not identify a named AI workflow that can yet be assessed as a deployment.", "Stage_at_disclosure": "Seven-year contract awarded June 30, 2026; planned facility construction 2026 and opening 2028", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Broad multi-domain program, not evidence of a specific AI model or operating workflow. Facility dates are plans, not completed milestones. AI-specific economics and deployments not disclosed.", "Source_IDs": "RS075", "Source_date": "2026-06-30", "Source_type": "company announcement", "Source_locator": "Contract scope, key points", "Source_URL": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-06-30; see chronology for operating dates.", "Context": "Accelerate evaluation and deployment of solutions for KOC operational and strategic priorities.", "Implementation": "SLB will support applied research, technology evaluation, testing, and deployment across a multi-domain project portfolio.", "Assessment": "Analyst inference: This is an enabling research-to-deployment framework with substantial potential breadth, but it should not be counted as a proven AI deployment until named projects and outcomes emerge.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Production & field operations", "Reservoir engineering & field development"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Strategy only", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "services-036-1", "Initiative_ID": "AI-162", "Title": "Ahmadi Innovation Valley seven-year technology evaluation and deployment contract", "Date": "2026-06-30", "Finding": "KOC awarded SLB a seven-year Ahmadi Innovation Valley contract.", "Metric": "", "Source_IDs": "RS075", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "services-036-2", "Initiative_ID": "AI-162", "Title": "Ahmadi Innovation Valley seven-year technology evaluation and deployment contract", "Date": "2026-2028 planned", "Finding": "Facility construction planned to start in 2026 and opening planned in 2028.", "Metric": "", "Source_IDs": "RS075", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-06-30", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["services-036"], "Research_search": " {\"key\": \"koc-ahmadi-innovation-valley\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-014\", \"AI-066\"], \"company\": \"Kuwait Oil Company\", \"title\": \"Ahmadi Innovation Valley seven-year technology evaluation and deployment contract\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [\"Production & field operations\", \"Reservoir engineering & field development\"], \"category\": \"Strategy only\", \"stage\": \"Seven-year contract awarded June 30, 2026; planned facility construction 2026 and opening 2028\", \"practical_relevance\": \"The program creates an evaluation-to-deployment channel spanning nearly 100 initiatives, but the source does not identify a named AI workflow that can yet be assessed as a deployment.\", \"business_problem\": \"Accelerate evaluation and deployment of solutions for KOC operational and strategic priorities.\", \"workflow\": \"SLB will support applied research, technology evaluation, testing, and deployment across a multi-domain project portfolio.\", \"technology_data\": \"AI is one of several domains; IIoT, production optimization, reservoir technologies, water management and energy transition are also included. No model or data integration is specified.\", \"integration_human_role\": \"KOC, operational teams, researchers and technology providers participate; the source emphasizes local capability and knowledge transfer.\", \"deployment_scope\": \"Seven-year KOC initiative with nearly 100 projects; dedicated facility planned for Kuwait. Specific field AI installation and scale undisclosed.\", \"partners\": [\"SLB\", \"KOC\", \"Industry, academic and technology providers (not individually named)\"], \"disclosed_results\": [], \"economics\": \"Contract value and AI-specific returns not disclosed.\", \"timeline\": [{\"date\": \"2026-06-30\", \"event\": \"KOC awarded SLB a seven-year Ahmadi Innovation Valley contract.\", \"source_ids\": [\"RS075\"]}, {\"date\": \"2026-2028 planned\", \"event\": \"Facility construction planned to start in 2026 and opening planned in 2028.\", \"source_ids\": [\"RS075\"]}], \"limitations\": [\"Broad multi-domain program, not evidence of a specific AI model or operating workflow.\", \"Facility dates are plans, not completed milestones.\", \"AI-specific economics and deployments not disclosed.\"], \"assessment\": \"Analyst inference: This is an enabling research-to-deployment framework with substantial potential breadth, but it should not be counted as a proven AI deployment until named projects and outcomes emerge.\", \"sources\": [{\"id\": \"RS075\", \"title\": \"SLB Awarded Seven-Year Contract Under Kuwait Oil Company’s Ahmadi Innovation Valley Initiative\", \"url\": \"https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1\", \"publisher\": \"SLB\", \"date\": \"2026-06-30\", \"type\": \"company announcement\", \"locator\": \"Contract scope, key points\", \"review_status\": \"full text reviewed\", \"supports\": [\"KOC awarded SLB a seven-year AIV contract to evaluate/test/deploy technologies; scope spans nearly 100 projects across AI, IIoT, production optimization, reservoirs, water and energy transition.\", \"A dedicated facility is planned, with construction expected to start 2026 and opening in 2028; the announcement names no specific AI workflow or outcome.\"], \"original_id\": \"koc_ahmadi\"}], \"field_sources\": {\"business_problem\": [\"RS075\"], \"workflow\": [\"RS075\"], \"technology_data\": [\"RS075\"], \"integration_human_role\": [\"RS075\"], \"deployment_scope\": [\"RS075\"], \"economics\": [\"RS075\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"New multi-domain KOC contract, related to but not an exact match for KOC seismic records AI-014/AI-066.\", \"cohort\": \"services\", \"packet_id\": \"services-036\", \"source_ids\": [\"RS075\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-162\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-162", "company": "Kuwait Oil Company", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1"], "findings": [{"severity": "medium", "field": "stage", "issue": "Procurement specifications/contracts establish solicitation or award, not necessarily operational commissioning.", "correction": "Preserve planned/contract stage unless evidence verifies a deployed asset workflow.", "source_url": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "KOC is the operator/customer and the AIV’s stated strategic technical challenges are upstream. The contract spans AI, production optimization and reservoir technology. “Upstream services & technology” misclassifies the customer’s initiative as a vendor/service-provider workflow.", "Segment_classification_confidence": "high", "Record_type": "strategy", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-163", "Company": "Novi Labs", "Company_primary": "Novi Labs", "Segment": "Upstream services & technology", "Theme": "AI products", "Initiative": "Novi Intelligence upstream research and forecasting", "Extracted_finding": "A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.", "Stage_at_disclosure": "Launched / unveiled 2025-09-10", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No independent benchmark, named customer deployment or workflow economics.", "Source_IDs": "RS077", "Source_date": "2025-09-10", "Source_type": "company announcement", "Source_locator": "Novi Intelligence product", "Source_URL": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-09-10; see chronology for operating dates.", "Context": "Analysts and operators need comparable forecasts and transparent evidence about well quality, completions, depletion and local inventory when screening assets and allocating development capital.", "Implementation": "Novi Intelligence surfaces forward-looking well forecasts and proprietary diagnostics (rock quality, completion intensity, prior depletion, well density); clients can inspect underlying sourced data to check and contextualize outputs.", "Assessment": "A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-005-1", "Initiative_ID": "AI-163", "Title": "Novi Intelligence upstream research and forecasting", "Date": "2025-09-10", "Finding": "Novi Intelligence unveiled.", "Metric": "", "Source_IDs": "RS077", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-09-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-005"], "Research_search": " {\"key\": \"novi-intelligence\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-122\"], \"company\": \"Novi Labs\", \"title\": \"Novi Intelligence upstream research and forecasting\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Launched / unveiled 2025-09-10\", \"practical_relevance\": \"A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.\", \"business_problem\": \"Analysts and operators need comparable forecasts and transparent evidence about well quality, completions, depletion and local inventory when screening assets and allocating development capital.\", \"workflow\": \"Novi Intelligence surfaces forward-looking well forecasts and proprietary diagnostics (rock quality, completion intensity, prior depletion, well density); clients can inspect underlying sourced data to check and contextualize outputs.\", \"technology_data\": \"Machine-learning forecasts over operator-sourced datasets and proprietary derived metrics; specific model classes, validation design and error rates are not disclosed.\", \"integration_human_role\": \"Client can validate underlying data; human review is described at a high level, no formal approval workflow disclosed.\", \"deployment_scope\": \"Offered to Novi customers; no customer count, basin coverage or deployment dates disclosed in launch.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No independently tested accuracy or realized economics. Vendor’s scale claims about capital decisions should not be read as causal savings.\", \"timeline\": [{\"date\": \"2025-09-10\", \"event\": \"Novi Intelligence unveiled.\", \"source_ids\": [\"RS077\"]}], \"limitations\": [\"No independent benchmark, named customer deployment or workflow economics.\"], \"assessment\": \"A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.\", \"sources\": [{\"id\": \"RS077\", \"title\": \"Novi Labs unveils Novi Intelligence\", \"url\": \"https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/\", \"publisher\": \"Novi Labs\", \"date\": \"2025-09-10\", \"type\": \"company announcement\", \"locator\": \"Novi Intelligence product\", \"review_status\": \"full text reviewed\", \"supports\": [\"Novi Intelligence joins operator-sourced data, machine learning and forecasts with proprietary metrics such as rock quality, completion intensity, depletion and well density.\", \"Product exposes underlying data for client validation; accuracy claims are vendor claims.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-006\"}], \"field_sources\": {\"business_problem\": [\"RS077\"], \"workflow\": [\"RS077\"], \"technology_data\": [\"RS077\"], \"integration_human_role\": [\"RS077\"], \"deployment_scope\": [\"RS077\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-005\", \"source_ids\": [\"RS077\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-163\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-163", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-164", "Company": "Applied Computing", "Company_primary": "Applied Computing", "Segment": "Cross-segment / enterprise", "Theme": "AI products", "Initiative": "Orbital on-edge process-industry AI and seed funding", "Extracted_finding": "Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.", "Stage_at_disclosure": "£9m seed announced 2025-05-28; company says one refinery live", "Reported_metric": "Refinery live sites: One large refinery (company says; unnamed) (Company-reported); Cloud AI cost comparison: 75% lower (company claim) (Company-reported); Payback: Under 10 weeks (company claim) (Company-reported)", "Interpretation_limit": "AI-127 appears to be later $20m Series A; no primary funding source verified in this review. No refinery/customer named, no model audit or safety case.", "Source_IDs": "RS081", "Source_date": "2025-05-28", "Source_type": "company announcement", "Source_locator": "Orbital architecture, financing and initial refinery", "Source_URL": "https://appliedcomputing.com/press-release-28052025", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-05-28; see chronology for operating dates.", "Context": "Process engineers seek process recommendations using sensor, engineering and chemical context while protecting sensitive plant data and limiting cloud compute costs.", "Implementation": "Orbital is presented as combining language models, time-series, physics and chemical-engineering models to help industrial teams analyze/optimize plant processes; deployment includes on-edge processing.", "Assessment": "Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-011-1", "Initiative_ID": "AI-164", "Title": "Orbital on-edge process-industry AI and seed funding", "Date": "2025-05-28", "Finding": "Applied Computing announced £9m seed investment and Orbital overview.", "Metric": "", "Source_IDs": "RS081", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-05-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-011"], "Research_search": " {\"key\": \"orbital-seed\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-127\"], \"company\": \"Applied Computing\", \"title\": \"Orbital on-edge process-industry AI and seed funding\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"£9m seed announced 2025-05-28; company says one refinery live\", \"practical_relevance\": \"Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.\", \"business_problem\": \"Process engineers seek process recommendations using sensor, engineering and chemical context while protecting sensitive plant data and limiting cloud compute costs.\", \"workflow\": \"Orbital is presented as combining language models, time-series, physics and chemical-engineering models to help industrial teams analyze/optimize plant processes; deployment includes on-edge processing.\", \"technology_data\": \"Multi-model system combining language/time-series/physics/chemistry, edge deployment. Detailed data interfaces, controls and named refinery not provided.\", \"integration_human_role\": \"Human decision process, safety interlocks and scope of closed-loop authority not disclosed.\", \"deployment_scope\": \"Company says one large refinery is live; it is unidentified and unverified externally.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Refinery live sites\", \"value\": \"One large refinery (company says; unnamed)\", \"period\": \"2025-05-28\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS081\"]}, {\"metric\": \"Cloud AI cost comparison\", \"value\": \"75% lower (company claim)\", \"period\": \"Not specified\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS081\"]}, {\"metric\": \"Payback\", \"value\": \"Under 10 weeks (company claim)\", \"period\": \"Not specified\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS081\"]}], \"economics\": \"£9m seed. Company claims 75% lower cost than cloud AI and sub-10-week payback; potential energy reductions are forward-looking, not audited.\", \"timeline\": [{\"date\": \"2025-05-28\", \"event\": \"Applied Computing announced £9m seed investment and Orbital overview.\", \"source_ids\": [\"RS081\"]}], \"limitations\": [\"AI-127 appears to be later $20m Series A; no primary funding source verified in this review.\", \"No refinery/customer named, no model audit or safety case.\"], \"assessment\": \"Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.\", \"sources\": [{\"id\": \"RS081\", \"title\": \"Applied Computing raises £9m seed to build Orbital AI\", \"url\": \"https://appliedcomputing.com/press-release-28052025\", \"publisher\": \"Applied Computing\", \"date\": \"2025-05-28\", \"type\": \"company announcement\", \"locator\": \"Orbital architecture, financing and initial refinery\", \"review_status\": \"full text reviewed\", \"supports\": [\"Applied Computing raised £9m seed funding for Orbital, a process-industry model using language, time-series, physics and chemical engineering models.\", \"Company says it is on edge, can work on sensitive process data, and reports one large refinery live.\", \"Cost savings, energy reduction and payback cited are company claims/potential, not independently supported.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-014\"}], \"field_sources\": {\"business_problem\": [\"RS081\"], \"workflow\": [\"RS081\"], \"technology_data\": [\"RS081\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS081\"], \"economics\": [\"RS081\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-011\", \"source_ids\": [\"RS081\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-164\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-164", "company": "Applied Computing", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://appliedcomputing.com/press-release-28052025"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-165", "Company": "Baker Hughes", "Company_primary": "Baker Hughes", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "AI sourcing optimization across procurement data", "Extracted_finding": "Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.", "Stage_at_disclosure": "Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.", "Reported_metric": "Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.", "Interpretation_limit": "Opportunity acceptance is not a savings percentage. Benefit amount, implementation costs, attribution method and publication date remain undisclosed.", "Source_IDs": "RS084", "Source_date": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "Source_type": "Vendor-hosted named customer case study", "Source_locator": "C3 AI customer page: Introduction, The Solution, and Project Highlights (web lines 147-214); page says initial 3-site pilot scaled to 10, then enterprise; CY2025 value in lines 162-164", "Source_URL": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "Evidence_status": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Fragmented procurement data and manual spreadsheet reconciliation.", "Implementation": "Vendor describes growth from a three-site pilot in 2019 to enterprise coverage, five source systems, weekly reviews and buyer feedback. The page remains undated.", "Assessment": "Data coverage is substantial, but a sourcing benefit should be tested against price movements, negotiated contracts and implementation costs before attributing it entirely to AI.", "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-015-1", "Initiative_ID": "AI-165", "Title": "AI sourcing optimization across procurement data", "Date": "2025; CY2025 result in undated case page", "Finding": "C3 AI case page reports Baker Hughes sourcing optimization scope and benefit.", "Metric": "", "Source_IDs": "RS084", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "M29-D-AI-165", "Initiative_ID": "AI-165", "Title": "AI sourcing optimization across enterprise procurement (implementation-detail enrichment)", "Date": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "Finding": "Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.", "Metric": "Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.", "Source_IDs": "RS084", "Evidence": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-165", "software_capital-015"], "Research_search": "{\"key\": \"missing-sept29-ai-165\", \"packet_id\": \"missing-sept29-ai-165\", \"record_ids\": [\"AI-165\"], \"existing_ids\": [\"AI-165\"], \"related_existing_ids\": [], \"company\": \"Baker Hughes\", \"title\": \"AI sourcing optimization across enterprise procurement (implementation-detail enrichment)\", \"department\": \"Supply chain & logistics\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.\", \"practical_relevance\": \"AI-165 already held a concise profile, but omitted the operational mechanics, feedback/validation controls, scale path, data footprint, user count, and the distinction between an opportunity acceptance rate and realized dollar benefit.\", \"business_problem\": \"\", \"workflow\": \"Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.\", \"technology_data\": \"Vendor describes growth from a three-site pilot in 2019 to enterprise coverage, five source systems, weekly reviews and buyer feedback. The page remains undated.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.\", \"period\": \"Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"RS084\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.\", \"event\": \"Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.\", \"source_ids\": [\"RS084\"]}], \"limitations\": [\"Opportunity acceptance is not a savings percentage. Benefit amount, implementation costs, attribution method and publication date remain undisclosed.\"], \"assessment\": \"AI-165 already held a concise profile, but omitted the operational mechanics, feedback/validation controls, scale path, data footprint, user count, and the distinction between an opportunity acceptance rate and realized dollar benefit.\", \"sources\": [{\"id\": \"RS084\", \"title\": \"Baker Hughes Achieves Bottom Line Savings with AI Sourcing Optimization\", \"url\": \"https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization\", \"publisher\": \"Baker Hughes\", \"date\": \"Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.\", \"type\": \"Vendor-hosted named customer case study\", \"locator\": \"C3 AI customer page: Introduction, The Solution, and Project Highlights (web lines 147-214); page says initial 3-site pilot scaled to 10, then enterprise; CY2025 value in lines 162-164\", \"review_status\": \"Full C3 AI HTML case study opened and read; all cited detail visible in page text.\", \"supports\": [\"Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.\", \"Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"RS084\"], \"field_sources\": {\"workflow\": [\"RS084\"], \"technology_data\": [\"RS084\"], \"deployment_scope\": [\"RS084\"], \"practical_relevance\": [\"RS084\"]}, \"verification\": \"verified partner\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Updated · 29 September 2026", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-165", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "200+ users", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-166", "Company": "Encino Energy", "Company_primary": "Encino Energy", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "Invoice AI and Snowflake-backed production-data agent rollout", "Extracted_finding": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "Stage_at_disclosure": "Invoice workflow in use; production-data agents rolling out in April 2025.", "Reported_metric": "Customer reports hundreds of administrative hours saved annually for one employee by the invoice application; not disaggregated to AI extraction alone. No production-agent outcome measured.", "Interpretation_limit": "Vendor/customer account. Maintenance-app savings are ordinary workflow automation, not AI-attributable. Snowflake supplies data; this is not proof of Snowflake-native inference.", "Source_IDs": "RS085", "Source_date": "2025-04-29", "Source_type": "Vendor customer case", "Source_locator": "Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83.", "Source_URL": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "Evidence_status": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "Microsoft story published 2025-04-29; production-agent rollout incomplete then. Invoice app is described as operating.", "Context": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "Implementation": "SCADA data enters Snowflake, then SQL virtual tables and Dataverse. A beta Snowflake connector was planned to remove the SQL step; AI uses Microsoft components.", "Assessment": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-017-1", "Initiative_ID": "AI-166", "Title": "Encino production-data and invoice AI workflows using Microsoft Power Platform", "Date": "2025-04-29", "Finding": "Microsoft published Encino Power Platform case; production agent rollout and invoice processing documented.", "Metric": "", "Source_IDs": "RS085", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "SW7-D-AI-166", "Initiative_ID": "AI-166", "Title": "Invoice AI and Snowflake-backed production-data agent rollout", "Date": "2025-04-29", "Finding": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "Metric": "Customer reports hundreds of administrative hours saved annually for one employee by the invoice application; not disaggregated to AI extraction alone. No production-agent outcome measured.", "Source_IDs": "RS085", "Evidence": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2025-04-29", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-166", "software_capital-017"], "Research_search": "{\"key\": \"swarm7-ai-166\", \"packet_id\": \"swarm7-ai-166\", \"record_ids\": [\"AI-166\"], \"existing_ids\": [\"AI-166\"], \"related_existing_ids\": [], \"company\": \"Encino Energy\", \"title\": \"Invoice AI and Snowflake-backed production-data agent rollout\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Invoice workflow in use; production-data agents rolling out in April 2025.\", \"practical_relevance\": \"AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.\", \"business_problem\": \"\", \"workflow\": \"AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.\", \"technology_data\": \"SCADA data enters Snowflake, then SQL virtual tables and Dataverse. A beta Snowflake connector was planned to remove the SQL step; AI uses Microsoft components.\", \"integration_human_role\": \"Invoice reviewer checks confidence; approvals remain human.\", \"deployment_scope\": \"Invoice workflow in use; production-data agents rolling out in April 2025.\", \"partners\": [\"Snowflake\", \"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Customer reports hundreds of administrative hours saved annually for one employee by the invoice application; not disaggregated to AI extraction alone. No production-agent outcome measured.\", \"period\": \"Annual customer estimate; measurement method undisclosed.\", \"attribution\": \"Source-reported claims; no independent validation.\", \"source_ids\": [\"RS085\"]}], \"economics\": \"Reported administrative time capacity; no AI project cost or cash payback.\", \"timeline\": [{\"date\": \"2025-04-29\", \"event\": \"Microsoft story published 2025-04-29; production-agent rollout incomplete then. Invoice app is described as operating.\", \"source_ids\": [\"RS085\"]}], \"limitations\": [\"Vendor/customer account. Maintenance-app savings are ordinary workflow automation, not AI-attributable. Snowflake supplies data; this is not proof of Snowflake-native inference.\"], \"assessment\": \"AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.\", \"sources\": [{\"title\": \"Encino Energy streamlines operations with Microsoft Power Platform\", \"url\": \"https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps\", \"date\": \"2025-04-29 (visible on page)\", \"locator\": \"Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83.\", \"review_status\": \"Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.\", \"type\": \"Vendor customer case\", \"id\": \"RS085\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"RS085\"], \"field_sources\": {\"workflow\": [\"RS085\"], \"technology_data\": [\"RS085\"], \"deployment_scope\": [\"RS085\"], \"practical_relevance\": [\"RS085\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Source-reported claims; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [{"claim": "Second-pass review", "reason": "Microsoft’s partner case gives no independent audit and does not disaggregate the administrative-time saving by AI versus workflow automation.", "source_url": ""}], "Audit_review": {"id": "AI-166", "company": "Encino Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps"], "findings": [{"severity": "medium", "field": "workflow / metric attribution", "issue": "Microsoft source explicitly identifies a Copilot Studio production-data agent and AI Builder invoice extraction with confidence review. However, “hundreds of hours” saved is attributed to the broader Power Platform invoice app, not isolated to AI extraction.", "correction": "Retain AI classification for the named Copilot/AI Builder components; do not attribute all Power Apps process savings to AI. Keep agent rollout/planning status distinct from the deployed invoice app.", "source_url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps"}], "residual_gaps": ["Microsoft’s partner case gives no independent audit and does not disaggregate the administrative-time saving by AI versus workflow automation."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Asset_location": "Not disclosed", "Followup_review": {"date": "2026-10-01", "decision": "Enriched", "note": "Deduplicated and original-source reviewed; specific cross-checks described in research methodology."}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-167", "Company": "TotalEnergies", "Company_primary": "TotalEnergies", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "MARGE maintenance report classification and BuyerCompanion procurement agent", "Extracted_finding": "The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.", "Stage_at_disclosure": "Customer story dated 2025-10-22; MARGE operational, BuyerCompanion France operational", "Reported_metric": "Fault cause identification: 25%-45% improvement (company-reported) (Vendor-reported customer statement); Savings target for selected procurement: 10% target, not realized result (Company-reported target)", "Interpretation_limit": "Model and data provenance are not disclosed. Separate MARGE measured identification improvement from BuyerCompanion savings target.", "Source_IDs": "RS086", "Source_date": "2025-10-22", "Source_type": "partner case study", "Source_locator": "MARGE and BuyerCompanion", "Source_URL": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-10-22; see chronology for operating dates.", "Context": "Maintenance intervention reports require manual fault-cause classification; procurement requests need supplier/specification/framework agreement guidance under purchasing rules.", "Implementation": "MARGE analyses and classifies intervention reports; BuyerCompanion drafts specifications, identifies suppliers/frameworks and recommends compliant purchases up to €50,000. TotalEnergies also rolled out 30,000 Microsoft 365 Copilot licenses with training and agent committee governance.", "Assessment": "The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-018-1", "Initiative_ID": "AI-167", "Title": "MARGE maintenance report classification and BuyerCompanion procurement agent", "Date": "2025-10-22", "Finding": "Microsoft customer story described 30k licenses and named MARGE and BuyerCompanion agents.", "Metric": "", "Source_IDs": "RS086", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2025-10-22", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-018"], "Research_search": " {\"key\": \"microsoft-totalenergies-agents\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"TotalEnergies\", \"title\": \"MARGE maintenance report classification and BuyerCompanion procurement agent\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Customer story dated 2025-10-22; MARGE operational, BuyerCompanion France operational\", \"practical_relevance\": \"The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.\", \"business_problem\": \"Maintenance intervention reports require manual fault-cause classification; procurement requests need supplier/specification/framework agreement guidance under purchasing rules.\", \"workflow\": \"MARGE analyses and classifies intervention reports; BuyerCompanion drafts specifications, identifies suppliers/frameworks and recommends compliant purchases up to €50,000. TotalEnergies also rolled out 30,000 Microsoft 365 Copilot licenses with training and agent committee governance.\", \"technology_data\": \"MARGE and BuyerCompanion use Microsoft Copilot Studio; Witivio supports BuyerCompanion. Underlying models/data sources not disclosed.\", \"integration_human_role\": \"Human decision-making retained; governance committee reviews value, architecture, cybersecurity, compliance, redundancy. BuyerCompanion produces recommendations, not stated to autonomously award purchases.\", \"deployment_scope\": \"Copilot rollout across about 30,000 employees; MARGE deployment scope not quantified; BuyerCompanion operational in France with potential wider staff reach.\", \"partners\": [\"Microsoft\", \"Witivio\"], \"disclosed_results\": [{\"metric\": \"Fault cause identification\", \"value\": \"25%-45% improvement (company-reported)\", \"period\": \"As of customer story, 2025-10-22\", \"attribution\": \"Vendor-reported customer statement\", \"source_ids\": [\"RS086\"]}, {\"metric\": \"Savings target for selected procurement\", \"value\": \"10% target, not realized result\", \"period\": \"Future target as of 2025-10-22\", \"attribution\": \"Company-reported target\", \"source_ids\": [\"RS086\"]}], \"economics\": \"TotalEnergies reports MARGE 25-45% improved fault-cause identification; BuyerCompanion targets 10% savings on certain purchases, not realized savings. Adoption metrics (70% recommending/using weekly) are self-reported.\", \"timeline\": [{\"date\": \"2025-10-22\", \"event\": \"Microsoft customer story described 30k licenses and named MARGE and BuyerCompanion agents.\", \"source_ids\": [\"RS086\"]}], \"limitations\": [\"Model and data provenance are not disclosed.\", \"Separate MARGE measured identification improvement from BuyerCompanion savings target.\"], \"assessment\": \"The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.\", \"sources\": [{\"id\": \"RS086\", \"title\": \"TotalEnergies from Microsoft 365 Copilot to Copilot Studio\", \"url\": \"https://www.microsoft.com/en/customers/story/25505-totalenergies-agents\", \"publisher\": \"Microsoft\", \"date\": \"2025-10-22\", \"type\": \"partner case study\", \"locator\": \"MARGE and BuyerCompanion\", \"review_status\": \"full text reviewed\", \"supports\": [\"MARGE uses Copilot Studio to classify intervention reports; company reports 25-45% improvement in fault-cause identification.\", \"BuyerCompanion streamlines purchases up to €50,000, operational in France, with company target of 10% savings on certain purchases.\", \"30,000 Copilot licenses, training and internal agent-validation committee described.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-020\"}], \"field_sources\": {\"business_problem\": [\"RS086\"], \"workflow\": [\"RS086\"], \"technology_data\": [\"RS086\"], \"integration_human_role\": [\"RS086\"], \"deployment_scope\": [\"RS086\"], \"economics\": [\"RS086\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-018\", \"source_ids\": [\"RS086\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-167\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-167", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/25505-totalenergies-agents"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-168", "Company": "Encino Energy", "Company_primary": "Encino Energy", "Segment": "Upstream", "Theme": "AI products", "Initiative": "InfinityPL autonomous plunger-lift optimization at Encino", "Extracted_finding": "Named AI-driven artificial-lift deployment at 344 platform wells; disclosed impacts remain vendor-reported.", "Stage_at_disclosure": "Case study published 2025-08-05; project timeframe stated as 2023 to present.", "Reported_metric": "Vendor case study gives three decline measures: pilot group, annual production decline improved 7% vs manually optimized wells; full-scale (344 wells), exponential decline rate 10% to 9% (1 percentage point); straight-line decline 19% manual vs 9% auto (10 points). Venting events -62%. All vendor-reported/customer case study.", "Interpretation_limit": "Vendor case only; raw data, cohort sizes and validation method are not supplied. Pilot and full-scale metrics should not be combined.", "Source_IDs": "RS087", "Source_date": "2025-08-05", "Source_type": "partner case study", "Source_locator": "Customer background; solutions; key outcomes", "Source_URL": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-08-05; see chronology for operating dates.", "Context": "Limited field teams could not consistently tune Encino’s growing plunger-lift portfolio, especially late-life wells.", "Implementation": "InfinityPL monitors wells, detects anomalies and autonomously adjusts setpoints; operators watch dashboards and intervene on exceptions.", "Assessment": "Analyst inference: The workflow is a named, scaled closed-loop deployment. The source supports the use case and reports improvements, but cannot independently establish impact.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-019-1", "Initiative_ID": "AI-168", "Title": "InfinityPL autonomous plunger-lift optimization at Encino", "Date": "2023", "Finding": "Stated project start.", "Metric": "", "Source_IDs": "RS087", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "software_capital-019-2", "Initiative_ID": "AI-168", "Title": "InfinityPL autonomous plunger-lift optimization at Encino", "Date": "2025-08-05", "Finding": "Case reports deployment scale and outcomes.", "Metric": "", "Source_IDs": "RS087", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2025-08-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-019"], "Research_search": " {\"key\": \"ambyint-encino-infinitypl\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Encino Energy\", \"title\": \"InfinityPL autonomous plunger-lift optimization at Encino\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Case study published 2025-08-05; project timeframe stated as 2023 to present.\", \"practical_relevance\": \"Named AI-driven artificial-lift deployment at 344 platform wells; disclosed impacts remain vendor-reported.\", \"business_problem\": \"Limited field teams could not consistently tune Encino’s growing plunger-lift portfolio, especially late-life wells.\", \"workflow\": \"InfinityPL monitors wells, detects anomalies and autonomously adjusts setpoints; operators watch dashboards and intervene on exceptions.\", \"technology_data\": \"Vendor describes real-time surveillance, anomaly detection and autonomous setpoint control; model and sensor details are undisclosed.\", \"integration_human_role\": \"Operators monitor performance and intervene on complex cases; setpoints run in auto mode.\", \"deployment_scope\": \"Encino Appalachian Basin; case says 344 platform wells, project timeframe 2023–present; pilot/full-scale cohort sizes are not given.\", \"partners\": [\"Ambyint\"], \"disclosed_results\": [{\"metric\": \"Annual decline improvement\", \"value\": \"7% vs manually optimized pilot wells\", \"period\": \"Aug 2025 case snapshot\", \"attribution\": \"Vendor-reported/customer case study\", \"source_ids\": [\"RS087\"]}, {\"metric\": \"Venting events reduced\", \"value\": \"62%\", \"period\": \"Aug 2025 full-scale case snapshot\", \"attribution\": \"Vendor-reported/customer case study\", \"source_ids\": [\"RS087\"]}], \"economics\": \"No audited dollar savings or ROI disclosed.\", \"timeline\": [{\"date\": \"2023\", \"event\": \"Stated project start.\", \"source_ids\": [\"RS087\"]}, {\"date\": \"2025-08-05\", \"event\": \"Case reports deployment scale and outcomes.\", \"source_ids\": [\"RS087\"]}], \"limitations\": [\"Vendor case only; raw data, cohort sizes and validation method are not supplied.\", \"Pilot and full-scale metrics should not be combined.\"], \"assessment\": \"Analyst inference: The workflow is a named, scaled closed-loop deployment. The source supports the use case and reports improvements, but cannot independently establish impact.\", \"sources\": [{\"id\": \"RS087\", \"title\": \"Infinity PL Drives Efficiency Gains for Encino Energy\", \"url\": \"https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy\", \"publisher\": \"Ambyint\", \"date\": \"2025-08-05\", \"type\": \"partner case study\", \"locator\": \"Customer background; solutions; key outcomes\", \"review_status\": \"full text reviewed\", \"supports\": [\"Case reports 344 Appalachian Basin wells and 2023–present timeframe.\", \"InfinityPL uses surveillance, anomaly detection and autonomous setpoint control with operator monitoring/intervention.\", \"Vendor reports 7% decline improvement in pilot wells versus manual optimization and 62% fewer venting events; validation method is absent.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"RS087\"}], \"field_sources\": {\"business_problem\": [\"RS087\"], \"workflow\": [\"RS087\"], \"technology_data\": [\"RS087\"], \"integration_human_role\": [\"RS087\"], \"deployment_scope\": [\"RS087\"], \"economics\": [\"RS087\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"Source URL was previously only excerpt-reviewed. Reopened full case text adds 344-well scale and specific vendor-reported results. Metrics remain attributed and unverified independently.\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-019\", \"source_ids\": [\"RS087\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-168\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "Partner/supplier primary evidence only; no operator-side corroboration found in this source pass.", "source_url": ""}], "Audit_review": {"id": "AI-168", "company": "Encino Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Ambyint case is supplier-reported and offers no direct operator confirmation in this audit.", "correction": "Retain vendor attribution for deployment and benefits; avoid presenting as independent verification.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "344 wells", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Encino adoption of Ambyint InfinityPL, evidenced only by the vendor case study. Exhibits quoting '1 pp' use the full-scale exponential measure; '7%' is the pilot-group measure.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-169", "Company": "Preem", "Company_primary": "Preem", "Segment": "Refining, LNG & processing", "Theme": "AI products", "Initiative": "Closed-loop AI optimization of Preem Lysekil FCC", "Extracted_finding": "Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.", "Stage_at_disclosure": "Partnership announced 2025-02-19; outcome not reported", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No post-startup result or deployment completion reported in source.", "Source_IDs": "RS088", "Source_date": "2025-02-19", "Source_type": "company announcement", "Source_locator": "Lysekil FCC deployment", "Source_URL": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "Evidence_status": "partially verified partner (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-02-19; see chronology for operating dates.", "Context": "Preem seeks improved FCC yield/efficiency and lower fuel use and emissions at its Lysekil refinery.", "Implementation": "Imubit Optimizing Brain is to optimize FCC process variables, improve yield profile and stabilize fuel-gas balance; announcement anticipates lower emissions/fuel consumption.", "Assessment": "Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "software_capital-020-1", "Initiative_ID": "AI-169", "Title": "Closed-loop AI optimization of Preem Lysekil FCC", "Date": "2025-02-19", "Finding": "Imubit and Preem announced Lysekil FCC partnership.", "Metric": "", "Source_IDs": "RS088", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2025-02-19", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["software_capital-020"], "Research_search": " {\"key\": \"imubit-preem-fcc\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Preem\", \"title\": \"Closed-loop AI optimization of Preem Lysekil FCC\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Partnership announced 2025-02-19; outcome not reported\", \"practical_relevance\": \"Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.\", \"business_problem\": \"Preem seeks improved FCC yield/efficiency and lower fuel use and emissions at its Lysekil refinery.\", \"workflow\": \"Imubit Optimizing Brain is to optimize FCC process variables, improve yield profile and stabilize fuel-gas balance; announcement anticipates lower emissions/fuel consumption.\", \"technology_data\": \"Imubit describes closed-loop AI optimization; release provides no variables/sensor details, model architecture, operator console, fallback controls or integration design.\", \"integration_human_role\": \"Closed-loop capability is stated; operating authority bounds and human approval are not disclosed.\", \"deployment_scope\": \"Single named refinery and FCC unit; partnership/deployment intent disclosed, not completion status or broad rollout.\", \"partners\": [\"Preem\"], \"disclosed_results\": [], \"economics\": \"No quantified savings or emissions results; outcomes are stated as intended.\", \"timeline\": [{\"date\": \"2025-02-19\", \"event\": \"Imubit and Preem announced Lysekil FCC partnership.\", \"source_ids\": [\"RS088\"]}], \"limitations\": [\"No post-startup result or deployment completion reported in source.\"], \"assessment\": \"Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.\", \"sources\": [{\"id\": \"RS088\", \"title\": \"Preem partners with Imubit for closed-loop AI optimization\", \"url\": \"https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization\", \"publisher\": \"Imubit\", \"date\": \"2025-02-19\", \"type\": \"company announcement\", \"locator\": \"Lysekil FCC deployment\", \"review_status\": \"full text reviewed\", \"supports\": [\"Imubit and Preem announced Optimizing Brain at Lysekil refinery FCC to optimize process variables, yields and fuel-gas balance.\", \"Fuel savings and emissions reduction were anticipated; no measured post-deployment results were given.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"sc-022\"}], \"field_sources\": {\"business_problem\": [\"RS088\"], \"workflow\": [\"RS088\"], \"technology_data\": [\"RS088\"], \"integration_human_role\": [\"RS088\"], \"deployment_scope\": [\"RS088\"], \"economics\": [\"RS088\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"\", \"cohort\": \"software_capital\", \"packet_id\": \"software_capital-020\", \"source_ids\": [\"RS088\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-169\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "partially verified partner", "Audit_flags": [{"claim": "Second-pass review", "reason": "Partner/supplier primary evidence only; no operator-side corroboration found in this source pass.", "source_url": ""}], "Audit_review": {"id": "AI-169", "company": "Preem", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization"], "findings": [{"severity": "medium", "field": "stage", "issue": "Imubit/Preem announcement says Optimizing Brain “will be deployed”; claimed emissions/fuel benefits are anticipated.", "correction": "Keep deployment prospective and metrics as expected, not achieved.", "source_url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified partner", "claim_status": "partially verified partner (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-170", "Company": "TC Energy", "Company_primary": "TC Energy", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "Autonomous Pipeline commercial decision support", "Extracted_finding": "Supports daily/weekly gas transportation and Park-and-Loan (PAL) decisions by giving commercial teams real-time operational insight across assets.", "Stage_at_disclosure": "In use; scale and launch date not disclosed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Report is primary but does not give publication month or initial deployment date. No financial attribution or evidence that model outputs directly execute trades. Distinguish from RAMP predictive maintenance: the report calls RAMP “advanced algorithms” but does not specify ML/AI, so it is not counted as an AI initiative.", "Source_IDs": "RS093; RS094", "Source_date": "2026-03-05", "Source_type": "annual report", "Source_locator": null, "Source_URL": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-03-05; see chronology for operating dates.", "Context": "TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities.", "Implementation": "TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities", "Assessment": "A practical internal commercial workflow because it connects asset telemetry to recurring daily/weekly transport decisions; disclosed detail is thin on data inputs, realized uplift and governance.", "Primary_department": "Trading & marketing", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-002-1", "Initiative_ID": "AI-170", "Title": "Autonomous Pipeline commercial decision support", "Date": "2025", "Finding": "2025 Sustainability Report describes the tool as an operational capability; launch date not given.", "Metric": "", "Source_IDs": "RS093", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-03-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-002"], "Research_search": " {\"key\": \"tc-autonomous-pipeline-ml\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"TC Energy\", \"title\": \"Autonomous Pipeline commercial decision support\", \"department\": \"Trading & marketing\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"In use; scale and launch date not disclosed\", \"practical_relevance\": \"Supports daily/weekly gas transportation and Park-and-Loan (PAL) decisions by giving commercial teams real-time operational insight across assets.\", \"business_problem\": \"TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities.\", \"workflow\": \"TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities\", \"technology_data\": \"Company identifies data analytics and machine learning; specific model, feeds, system architecture and measured accuracy are not disclosed.\", \"integration_human_role\": \"Commercial teams use the real-time insights to inform strategy and decisions; the report does not state autonomous execution.\", \"deployment_scope\": \"Company-wide commercial use is suggested by plural “commercial teams,” but assets, users, geographies and adoption rate are not disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"Incremental transportation/PAL opportunities are described as a value pathway; no quantified revenue, margin or cost outcome is reported.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"2025 Sustainability Report describes the tool as an operational capability; launch date not given.\", \"source_ids\": [\"RS093\"]}], \"limitations\": [\"Report is primary but does not give publication month or initial deployment date.\", \"No financial attribution or evidence that model outputs directly execute trades.\", \"Distinguish from RAMP predictive maintenance: the report calls RAMP “advanced algorithms” but does not specify ML/AI, so it is not counted as an AI initiative.\"], \"assessment\": \"A practical internal commercial workflow because it connects asset telemetry to recurring daily/weekly transport decisions; disclosed detail is thin on data inputs, realized uplift and governance.\", \"sources\": [{\"id\": \"RS093\", \"title\": \"TC Energy 2025 Report on Sustainability\", \"url\": \"https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf\", \"publisher\": \"TC Energy\", \"date\": \"2025\", \"type\": \"annual report\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Autonomous Pipeline analytics and machine learning for commercial daily/weekly asset strategies and transport/PAL opportunities; RAMP described as advanced algorithms only.\"], \"original_id\": \"tc25\"}, {\"id\": \"RS094\", \"title\": \"TC Energy 2026 Management Information Circular\", \"url\": \"https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf\", \"publisher\": \"TC Energy\", \"date\": \"2026-03-05\", \"type\": \"proxy statement\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"AI governance, training, board oversight; governance does not evidence operating deployment.\"], \"original_id\": \"tc26mic\"}], \"field_sources\": {\"business_problem\": [\"RS093\"], \"workflow\": [\"RS093\"], \"technology_data\": [\"RS093\"], \"integration_human_role\": [\"RS093\"], \"deployment_scope\": [\"RS093\"], \"economics\": [\"RS093\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-002\", \"source_ids\": [\"RS093\", \"RS094\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-170\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-170", "company": "TC Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and AI passages were located in the source.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-171", "Company": "TC Energy", "Company_primary": "TC Energy", "Segment": "Midstream", "Theme": "Enterprise AI", "Initiative": "Pipeline IQ engineering-document verification agents", "Extracted_finding": "PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.", "Stage_at_disclosure": "Operating use described by the source; launch date and full rollout scope not established.", "Reported_metric": "AWS reports 85% less document-review time and more than 100 PIPER users. Early tests on documents up to 120 pages reported 95–100% accuracy. Over $18.7m in maintenance capital avoidance is expected over five years for PIPER and SCOUT combined; it is a forecast.", "Interpretation_limit": "Undated page: launch and measurement dates are unknown. The test sample and accuracy definition are unstated. Engineer review remains required. Expected capital avoidance is not achieved savings; full costs and realized cash are absent.", "Source_IDs": "RS095; OCT5-S037", "Source_date": "Undated source; see chronology", "Source_type": "AWS customer story", "Source_locator": "PIPER, SCOUT and outcome sections", "Source_URL": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "Evidence_status": "Original source text reviewed on 6 October 2026", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "PIPER compares engineering documentation and material test reports with weld maps and pressure-testing documentation, including heat-number traceability and completeness checks. SCOUT reportedly searches/discovers information across millions of pages. Engineers make the final compliance review.", "Implementation": "PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.", "Assessment": "Strong practical AI workflow and concrete engineering control, but implementation period is undated and all performance/economic metrics are vendor/partner-reported; do not count the metrics as independently verified.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-003-1", "Initiative_ID": "AI-171", "Title": "Pipeline IQ engineering-document verification agents", "Date": "undated", "Finding": "AWS case study describes active users and current platform; publication and launch dates are not provided.", "Metric": "", "Source_IDs": "RS095", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "OCT5-AI-171-1", "Initiative_ID": "AI-171", "Title": "Pipeline IQ engineering-document verification agents", "Date": "Undated source; see chronology", "Finding": "PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.", "Metric": "AWS reports 85% less document-review time and more than 100 PIPER users. Early tests on documents up to 120 pages reported 95–100% accuracy. Over $18.7m in maintenance capital avoidance is expected over five years for PIPER and SCOUT combined; it is a forecast.", "Source_IDs": "OCT5-S037", "Source_URL": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Latest_disclosure": "Undated source; see chronology", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-gap-ai-171", "midstream-003"], "Research_search": "{\"packet_id\": \"oct6-gap-ai-171\", \"review_date\": \"2026-10-06\", \"title\": \"Pipeline IQ engineering-document verification agents\", \"workflow\": \"PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.\", \"deployment_scope\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"result_label\": \"Reported result and measurement scope\", \"attribution\": \"Vendor/customer account\", \"economics\": \"Full implementation costs and realized net cash are not disclosed.\", \"limitations\": [\"Undated page: launch and measurement dates are unknown. The test sample and accuracy definition are unstated. Engineer review remains required. Expected capital avoidance is not achieved savings; full costs and realized cash are absent.\"], \"verification\": \"verified partner\", \"source_review\": \"Full original document reviewed on 6 October 2026\", \"source_locator\": \"PIPER, SCOUT and outcome sections\", \"timeline_date\": \"Undated source; see chronology\", \"timeline_event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"stage\": \"Operating use described by the source; launch date and full rollout scope not established.\", \"key\": \"oct6-gap-ai-171\", \"record_ids\": [\"AI-171\"], \"company\": \"TC Energy\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"source_ids\": [\"RS095\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"RS095\"], \"workflow\": [\"RS095\"], \"technology_data\": [\"RS095\"], \"deployment_scope\": [\"RS095\"], \"practical_relevance\": [\"RS095\"], \"economics\": [\"RS095\"], \"integration_human_role\": [\"RS095\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and measurement scope\", \"value\": \"AWS reports 85% less document-review time and more than 100 PIPER users. Early tests on documents up to 120 pages reported 95–100% accuracy. Over $18.7m in maintenance capital avoidance is expected over five years for PIPER and SCOUT combined; it is a forecast.\", \"period\": \"undated\", \"attribution\": \"Vendor/customer account\", \"source_ids\": [\"RS095\"]}], \"timeline\": [{\"date\": \"Undated source; see chronology\", \"event\": \"New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.\", \"source_ids\": [\"RS095\"]}], \"sources\": [{\"id\": \"RS095\", \"title\": \"TC Energy Pipeline IQ on AWS\", \"url\": \"https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/\", \"date\": \"undated\", \"type\": \"partner case study\", \"publisher\": \"Amazon Web Services\", \"locator\": \"PIPER, SCOUT and outcome sections\", \"review_status\": \"Full original document reviewed on 6 October 2026\"}]}", "Research_status": "Expanded source review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-171", "company": "TC Energy", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, "Audit_outcome": "supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Undated page: launch and measurement dates are unknown. The test sample and accuracy definition are unstated. Engineer review remains required. Expected capital avoidance is not achieved savings; full costs and realized cash are absent.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Original source text reviewed on 6 October 2026", "independent_validation": "Not established by this collection"}}, {"ID": "AI-172", "Company": "Enbridge", "Company_primary": "Enbridge", "Segment": "Midstream", "Theme": "Enterprise AI", "Initiative": "Machine-learning prioritization of pipeline deformation features", "Extracted_finding": "Helps integrity teams triage large volumes of in-line inspection and field observations to focus engineering review on potentially injurious deformations.", "Stage_at_disclosure": "Development disclosed in 2025; subsequent operational use not established", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "2025 status is development; 2026 handbook supports ML use in assessment but does not establish that it is the identical tool. No external validation, model performance or savings disclosed. Annual integrity spend is not attributable to this ML workflow.", "Source_IDs": "RS096; RS097; RS098", "Source_date": "2026", "Source_type": "sustainability report", "Source_locator": null, "Source_URL": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026; see chronology for operating dates.", "Context": "The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation.", "Implementation": "The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation", "Assessment": "Good operational fit because prediction narrows an established inspection-to-engineering workflow; evidence supports development and later general ML use, but public material does not link the same named model to measured outcomes.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-004-1", "Initiative_ID": "AI-172", "Title": "Machine-learning prioritization of pipeline deformation features", "Date": "2025", "Finding": "Sustainability report describes tool as under development.", "Metric": "", "Source_IDs": "RS096", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "midstream-004-2", "Initiative_ID": "AI-172", "Title": "Machine-learning prioritization of pipeline deformation features", "Date": "2026", "Finding": "Customer handbook describes ML predictive analysis being used in complex defect assessment to prioritize repairs/resource allocation.", "Metric": "", "Source_IDs": "RS097", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-004"], "Research_search": " {\"key\": \"enbridge-ml-integrity-prioritization\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Enbridge\", \"title\": \"Machine-learning prioritization of pipeline deformation features\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Development disclosed in 2025; subsequent operational use not established\", \"practical_relevance\": \"Helps integrity teams triage large volumes of in-line inspection and field observations to focus engineering review on potentially injurious deformations.\", \"business_problem\": \"The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation.\", \"workflow\": \"The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation\", \"technology_data\": \"Machine learning applied to ILI/field inspection data; model design, thresholds and validation not disclosed.\", \"integration_human_role\": \"Engineers assess prioritized features and determine mitigation; model supports triage rather than replacing integrity decisions.\", \"deployment_scope\": \"Company pipeline integrity program; locations, asset coverage, production status and user count not disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No AI-attributable cost, risk reduction or integrity outcome quantified.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"Sustainability report describes tool as under development.\", \"source_ids\": [\"RS096\"]}, {\"date\": \"2026\", \"event\": \"Customer handbook describes ML predictive analysis being used in complex defect assessment to prioritize repairs/resource allocation.\", \"source_ids\": [\"RS097\"]}], \"limitations\": [\"2025 status is development; 2026 handbook supports ML use in assessment but does not establish that it is the identical tool.\", \"No external validation, model performance or savings disclosed.\", \"Annual integrity spend is not attributable to this ML workflow.\"], \"assessment\": \"Good operational fit because prediction narrows an established inspection-to-engineering workflow; evidence supports development and later general ML use, but public material does not link the same named model to measured outcomes.\", \"sources\": [{\"id\": \"RS096\", \"title\": \"Enbridge 2025 Sustainability Report\", \"url\": \"https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31\", \"publisher\": \"Enbridge\", \"date\": \"2025\", \"type\": \"sustainability report\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Describes developing machine-learning prioritization of potentially injurious deformation features from in-line inspection and field data; engineers assess and mitigate priority features.\"], \"original_id\": \"enb25\"}, {\"id\": \"RS097\", \"title\": \"Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability\", \"url\": \"https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability\", \"publisher\": \"Enbridge\", \"date\": \"2026\", \"type\": \"company technical publication\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"States ML predictive analysis in complex pipeline defect assessment to prioritize repairs and allocate resources; no quantified outcome.\"], \"original_id\": \"enb26\"}, {\"id\": \"RS098\", \"title\": \"Enbridge 2026 AI Policy\", \"url\": \"https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152\", \"publisher\": \"Enbridge\", \"date\": \"2026\", \"type\": \"company policy\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Company policy describes AI governance and general AI uses; does not by itself date or quantify named operating workflow.\"], \"original_id\": \"enbmic\"}], \"field_sources\": {\"business_problem\": [\"RS096\", \"RS097\"], \"workflow\": [\"RS096\", \"RS097\"], \"technology_data\": [\"RS096\"], \"integration_human_role\": [\"RS096\", \"RS097\"], \"deployment_scope\": [\"RS096\", \"RS097\"], \"economics\": [\"RS097\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-004\", \"source_ids\": [\"RS096\", \"RS097\", \"RS098\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-172\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-172", "company": "Enbridge", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Audit logged the primary PDF inaccessible on 27 Sep 2026; re-opened 5 Oct 2026 (HTTP 200) and AI passages were located in the source.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-173", "Company": "ONEOK", "Company_primary": "ONEOK", "Segment": "Midstream", "Theme": "Enterprise AI", "Initiative": "AI virtual river gauges for extreme-weather response", "Extracted_finding": "Aims to help pipeline operators locate where flooding or water-related exposure may threaten pipeline assets when physical river-gauge coverage is sparse.", "Stage_at_disclosure": "Deployment reported; initial date not disclosed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Page is undated; accessed 2026-09-27 does not prove an in-window launch or update. Gauge count and method are company-reported, provider and validation absent. No disclosed operational outcomes or comparison to physical gauges.", "Source_IDs": "RS099; RS100", "Source_date": "2026", "Source_type": "company sustainability webpage", "Source_locator": null, "Source_URL": "https://sustainability.oneok.com/innovation-spotlight", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026; see chronology for operating dates.", "Context": "The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response.", "Implementation": "The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response", "Assessment": "One of the clearest upstream/midstream-adjacent practical AI cases, with explicit physics-informed method, scale and human triage; time-window status remains uncertain because only an undated current page was reviewed.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-005-1", "Initiative_ID": "AI-173", "Title": "AI virtual river gauges for extreme-weather response", "Date": "undated", "Finding": "Current innovation page reports deployed gauge count and weather-response use; initial deployment date not given.", "Metric": "", "Source_IDs": "RS099", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-005"], "Research_search": " {\"key\": \"oneok-virtual-river-gauges\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"ONEOK\", \"title\": \"AI virtual river gauges for extreme-weather response\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Deployment reported; initial date not disclosed\", \"practical_relevance\": \"Aims to help pipeline operators locate where flooding or water-related exposure may threaten pipeline assets when physical river-gauge coverage is sparse.\", \"business_problem\": \"The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response.\", \"workflow\": \"The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response\", \"technology_data\": \"ONEOK describes physics-based machine learning combining gravity, river force, pipeline characteristics and terrain/water information; provider and detailed data sources not named.\", \"integration_human_role\": \"AI output is a risk-triage layer followed by field assessment; no autonomous operational control described.\", \"deployment_scope\": \"Company reports 48,872 virtual gauges and interaction with over 100,000 terrain/water locations; in-service footprint and geographic distribution not described.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No avoided outage, response-time or financial metric disclosed.\", \"timeline\": [{\"date\": \"undated\", \"event\": \"Current innovation page reports deployed gauge count and weather-response use; initial deployment date not given.\", \"source_ids\": [\"RS099\"]}], \"limitations\": [\"Page is undated; accessed 2026-09-27 does not prove an in-window launch or update.\", \"Gauge count and method are company-reported, provider and validation absent.\", \"No disclosed operational outcomes or comparison to physical gauges.\"], \"assessment\": \"One of the clearest upstream/midstream-adjacent practical AI cases, with explicit physics-informed method, scale and human triage; time-window status remains uncertain because only an undated current page was reviewed.\", \"sources\": [{\"id\": \"RS099\", \"title\": \"ONEOK Innovation Spotlight\", \"url\": \"https://sustainability.oneok.com/innovation-spotlight\", \"publisher\": \"ONEOK\", \"date\": \"undated\", \"type\": \"company sustainability webpage\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Reports AI-powered virtual river gauges and 48,872 deployed; weather-response risk triage. Undated page does not establish launch date or confirm in-window deployment.\"], \"publication_date\": \"undated; not stated on page\", \"original_id\": \"oneok\"}, {\"id\": \"RS100\", \"title\": \"ONEOK Q4 2025 earnings release\", \"url\": \"https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638\", \"publisher\": \"ONEOK\", \"date\": \"2026\", \"type\": \"earnings release\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Reviewed alongside company materials; no specific AI workflow found in release.\"], \"original_id\": \"oneokearn\"}], \"field_sources\": {\"business_problem\": [\"RS099\"], \"workflow\": [\"RS099\"], \"technology_data\": [\"RS099\"], \"integration_human_role\": [\"RS099\"], \"deployment_scope\": [\"RS099\"], \"economics\": [\"RS099\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Existing in-window deployment date could not be verified; do not imply the 48,872 gauge count is a 2026 increase.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-005\", \"source_ids\": [\"RS099\", \"RS100\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-173\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-173", "company": "ONEOK", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://sustainability.oneok.com/innovation-spotlight", "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-174", "Company": "Energy Transfer", "Company_primary": "Energy Transfer", "Segment": "AI-demand infrastructure", "Theme": "AI infrastructure", "Initiative": "Long-term gas transportation for Meta Louisiana data center", "Extracted_finding": "Dedicated pipeline transportation to a planned data-center project reflects gas infrastructure demand from AI/data-center expansion, not ET internal AI.", "Stage_at_disclosure": "Agreement announced; facilities planned for future service", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company says the project supports AI/data-center demand; no evidence Energy Transfer runs the data-center AI or has operating AI use. Future service, not current deliveries.", "Source_IDs": "RS104", "Source_date": "2025-11-04", "Source_type": "company announcement", "Source_locator": null, "Source_URL": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-11-04; see chronology for operating dates.", "Context": "Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service.", "Implementation": "Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service", "Assessment": "Separate infrastructure case to preserve demand exposure without overstating it as AI adoption.", "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "AI-demand infrastructure", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-011-1", "Initiative_ID": "AI-174", "Title": "Long-term gas transportation for Meta Louisiana data center", "Date": "2025-11-04", "Finding": "Long-term service agreement announced.", "Metric": "", "Source_IDs": "RS104", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-11-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-011"], "Research_search": " {\"key\": \"et-meta-ai-campus\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-062\"], \"company\": \"Energy Transfer\", \"title\": \"Long-term gas transportation for Meta Louisiana data center\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Agreement announced; facilities planned for future service\", \"practical_relevance\": \"Dedicated pipeline transportation to a planned data-center project reflects gas infrastructure demand from AI/data-center expansion, not ET internal AI.\", \"business_problem\": \"Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service.\", \"workflow\": \"Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service\", \"technology_data\": \"Gas transportation assets and contract; no AI methods or digital operations described.\", \"integration_human_role\": \"Energy Transfer provides transport; Entergy serves the end customer.\", \"deployment_scope\": \"Future site and lateral construction; disclosed target service date 2028, outside or beyond in-window completion.\", \"partners\": [\"Entergy Louisiana\", \"Meta Platforms\"], \"disclosed_results\": [], \"economics\": \"Contract terms are not fully disclosed beyond service volume/term in company release; no AI-derived economics.\", \"timeline\": [{\"date\": \"2025-11-04\", \"event\": \"Long-term service agreement announced.\", \"source_ids\": [\"RS104\"]}], \"limitations\": [\"Company says the project supports AI/data-center demand; no evidence Energy Transfer runs the data-center AI or has operating AI use.\", \"Future service, not current deliveries.\"], \"assessment\": \"Separate infrastructure case to preserve demand exposure without overstating it as AI adoption.\", \"sources\": [{\"id\": \"RS104\", \"title\": \"Entergy Louisiana and Energy Transfer sign agreement supporting Meta data center\", \"url\": \"https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports\", \"publisher\": \"Energy Transfer\", \"date\": \"2025-11-04\", \"type\": \"company announcement\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Long-term gas transportation infrastructure for future Meta site; separate AI-demand category, not internal AI.\"], \"original_id\": \"etmeta\"}], \"field_sources\": {\"business_problem\": [\"RS104\"], \"workflow\": [\"RS104\"], \"technology_data\": [\"RS104\"], \"integration_human_role\": [\"RS104\"], \"deployment_scope\": [\"RS104\"], \"economics\": [\"RS104\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"Not the same specific CloudBurst campus; exact existing AI-062 only maps CloudBurst so this separate Meta disclosure is related, not exact same initiative.\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-011\", \"source_ids\": [\"RS104\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-174\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-174", "company": "Energy Transfer", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "infrastructure", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-175", "Company": "Energy Transfer", "Company_primary": "Energy Transfer", "Segment": "AI-demand infrastructure", "Theme": "AI infrastructure", "Initiative": "Nexus Hubbard Campus gas transport for behind-the-meter AI campus", "Extracted_finding": "Gas transportation agreement supports a planned AI campus and on-site generation, creating an infrastructure link to compute demand rather than an internal AI workflow.", "Stage_at_disclosure": "Agreement reported in Q1 2026 update; operating status not disclosed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The source confirms an agreement but detailed project timing and volumes are not disclosed in the reviewed passage. AI demand does not indicate internal use of AI by ET.", "Source_IDs": "RS105", "Source_date": "2026", "Source_type": "earnings release", "Source_locator": null, "Source_URL": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026; see chronology for operating dates.", "Context": "Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation.", "Implementation": "Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation", "Assessment": "Track as commercial infrastructure opportunity only; details are less complete than the CloudBurst case.", "Primary_department": "Commercial, power & infrastructure", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "AI-demand infrastructure", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-012-1", "Initiative_ID": "AI-175", "Title": "Nexus Hubbard Campus gas transport for behind-the-meter AI campus", "Date": "2026", "Finding": "Q1 2026 earnings update reports transport agreements.", "Metric": "", "Source_IDs": "RS105", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-012"], "Research_search": " {\"key\": \"et-nexus-ai-campus\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Energy Transfer\", \"title\": \"Nexus Hubbard Campus gas transport for behind-the-meter AI campus\", \"department\": \"Commercial, power & infrastructure\", \"secondary_departments\": [], \"category\": \"AI-demand infrastructure\", \"stage\": \"Agreement reported in Q1 2026 update; operating status not disclosed\", \"practical_relevance\": \"Gas transportation agreement supports a planned AI campus and on-site generation, creating an infrastructure link to compute demand rather than an internal AI workflow.\", \"business_problem\": \"Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation.\", \"workflow\": \"Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation\", \"technology_data\": \"Natural-gas transportation; no AI model or operational-AI system described.\", \"integration_human_role\": \"ET is pipeline service provider; Nexus campus uses gas for power.\", \"deployment_scope\": \"Planned central Texas campus; volume, term and service date not reviewed in detail in the earnings update.\", \"partners\": [\"Nexus Data Centers\"], \"disclosed_results\": [], \"economics\": \"Financial contribution not disclosed.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"Q1 2026 earnings update reports transport agreements.\", \"source_ids\": [\"RS105\"]}], \"limitations\": [\"The source confirms an agreement but detailed project timing and volumes are not disclosed in the reviewed passage.\", \"AI demand does not indicate internal use of AI by ET.\"], \"assessment\": \"Track as commercial infrastructure opportunity only; details are less complete than the CloudBurst case.\", \"sources\": [{\"id\": \"RS105\", \"title\": \"Energy Transfer Q1 2026 results and updates\", \"url\": \"https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates\", \"publisher\": \"Energy Transfer\", \"date\": \"2026\", \"type\": \"earnings release\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Reports agreements to transport gas to Nexus Hubbard Campus, an AI data center campus with on-site gas generation; infrastructure supporting AI demand.\"], \"original_id\": \"etnexus\"}], \"field_sources\": {\"business_problem\": [\"RS105\"], \"workflow\": [\"RS105\"], \"technology_data\": [\"RS105\"], \"integration_human_role\": [\"RS105\"], \"deployment_scope\": [\"RS105\"], \"economics\": [\"RS105\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-012\", \"source_ids\": [\"RS105\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-175\"]}", "Research_status": "Expanded source review", "Research_category": "AI-demand infrastructure", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-175", "company": "Energy Transfer", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "infrastructure", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-176", "Company": "Targa Resources", "Company_primary": "Targa Resources", "Segment": "Midstream", "Theme": "Enterprise AI", "Initiative": "AI and emerging-technology operational deployment role", "Extracted_finding": "Shows an organizational role intended to evaluate, pilot and integrate AI into asset-heavy operating work, but does not establish that any particular AI workflow has been deployed.", "Stage_at_disclosure": "Recruiting / capability-building signal", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Job listing is a strategy/workforce signal, not a deployed initiative. No named use case or operating result.", "Source_IDs": "RS108", "Source_date": "2026-05-07", "Source_type": "job posting", "Source_locator": null, "Source_URL": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-05-07; see chronology for operating dates.", "Context": "Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance.", "Implementation": "Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance", "Assessment": "Useful leading indicator of planned operational AI capability, but insufficient evidence to claim production adoption.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Strategy only", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "midstream-015-1", "Initiative_ID": "AI-176", "Title": "AI and emerging-technology operational deployment role", "Date": "2026-05-07", "Finding": "Job posting published.", "Metric": "", "Source_IDs": "RS108", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-05-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["midstream-015"], "Research_search": " {\"key\": \"targa-ai-principal-engineer-role\", \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Targa Resources\", \"title\": \"AI and emerging-technology operational deployment role\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Strategy only\", \"stage\": \"Recruiting / capability-building signal\", \"practical_relevance\": \"Shows an organizational role intended to evaluate, pilot and integrate AI into asset-heavy operating work, but does not establish that any particular AI workflow has been deployed.\", \"business_problem\": \"Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance.\", \"workflow\": \"Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance\", \"technology_data\": \"Posting requests AI/ML deployment experience and familiarity with operational systems; no named vendor/model/data source disclosed.\", \"integration_human_role\": \"Role coordinates with operations, reliability, engineering and commercial stakeholders; hiring intent is the only verified implementation signal.\", \"deployment_scope\": \"One role/job posting across the described organization; no rollout count or facilities disclosed.\", \"partners\": [\"None disclosed\"], \"disclosed_results\": [], \"economics\": \"No economics reported.\", \"timeline\": [{\"date\": \"2026-05-07\", \"event\": \"Job posting published.\", \"source_ids\": [\"RS108\"]}], \"limitations\": [\"Job listing is a strategy/workforce signal, not a deployed initiative.\", \"No named use case or operating result.\"], \"assessment\": \"Useful leading indicator of planned operational AI capability, but insufficient evidence to claim production adoption.\", \"sources\": [{\"id\": \"RS108\", \"title\": \"Targa Principal Engineer, AI & Emerging Technology job posting\", \"url\": \"https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925\", \"publisher\": \"Targa Resources\", \"date\": \"2026-05-07\", \"type\": \"job posting\", \"locator\": null, \"review_status\": \"full text reviewed\", \"accessed\": \"2026-09-27\", \"supports\": [\"Role coordinates AI/technology selection, pilots and adoption across operations/reliability/SCADA/integrity/commercial; intent/hiring signal, not proof of deployment.\"], \"original_id\": \"targa\"}], \"field_sources\": {\"business_problem\": [\"RS108\"], \"workflow\": [\"RS108\"], \"technology_data\": [\"RS108\"], \"integration_human_role\": [\"RS108\"], \"deployment_scope\": [\"RS108\"], \"economics\": [\"RS108\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"midstream\", \"packet_id\": \"midstream-015\", \"source_ids\": [\"RS108\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-176\"]}", "Research_status": "Expanded source review", "Research_category": "Strategy only", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-176", "company": "Targa Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925"], "findings": [{"severity": "medium", "field": "stage", "issue": "Targa job description shows hiring intent and target responsibilities only.", "correction": "Do not infer a live operational AI system from the role.", "source_url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "strategy", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-177", "Company": "Aramco", "Company_primary": "Aramco", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "MetaBrain for Plants process-modeling system", "Extracted_finding": "Reusable industrial models could reduce the effort required to build plant-specific analytics.", "Stage_at_disclosure": "Announced/development-stage process-modeling system; future-tense benefits, facility deployment not established", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No measured accuracy, economic impact or independently evaluated results disclosed. Capability statements do not establish plant-wide autonomous control.", "Source_IDs": "RS111", "Source_date": "2025-02-09", "Source_type": "Company announcement / remarks", "Source_locator": "MetaBrain and MetaBrain for Plants paragraphs", "Source_URL": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-02-09; see chronology for operating dates.", "Context": "Interpreting process behavior and reducing repeated model-development work.", "Implementation": "A time-series transformer models plant processes to inform operators, engineers and scientists. Intended tasks include demand anticipation, process optimization and product-quality prediction.", "Assessment": "Ambitious model reuse across process operations, with limited public implementation evidence. A pilot-to-production study would be needed to evaluate transferability across units.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "root_backfill-005-1", "Initiative_ID": "AI-177", "Title": "MetaBrain for Plants process-modeling system", "Date": "2025-02-09", "Finding": "MetaBrain for Plants introduced in executive remarks.", "Metric": "", "Source_IDs": "RS111", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-02-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["root_backfill-005"], "Research_search": " {\"key\": \"aramco-metabrain-plants\", \"existing_ids\": [], \"company\": \"Aramco\", \"title\": \"MetaBrain for Plants process-modeling system\", \"category\": \"Internal adoption\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"stage\": \"Introduced at LEAP 2025; rollout scale undisclosed\", \"verification\": \"verified primary\", \"partners\": [], \"sources\": [{\"id\": \"RS111\", \"title\": \"Remarks by Ahmad O. Al-Khowaiter at LEAP 2025\", \"url\": \"https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025\", \"date\": \"2025-02-09\", \"locator\": \"MetaBrain and MetaBrain for Plants paragraphs\", \"publisher\": \"Aramco\", \"type\": \"Company announcement / remarks\", \"review_status\": \"Full primary text reviewed\", \"supports\": [], \"accessed\": \"2026-09-27\", \"original_id\": \"AR5\"}], \"field_sources\": {\"business_problem\": [\"RS111\"], \"workflow\": [\"RS111\"], \"technology_data\": [\"RS111\"], \"integration_human_role\": [\"RS111\"], \"deployment_scope\": [\"RS111\"], \"economics\": [\"RS111\"]}, \"disclosed_results\": [], \"timeline\": [{\"date\": \"2025-02-09\", \"event\": \"MetaBrain for Plants introduced in executive remarks.\", \"source_ids\": [\"RS111\"]}], \"limitations\": [\"No measured accuracy, economic impact or independently evaluated results disclosed.\", \"Capability statements do not establish plant-wide autonomous control.\"], \"economics\": \"Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.\", \"practical_relevance\": \"Reusable industrial models could reduce the effort required to build plant-specific analytics.\", \"business_problem\": \"Interpreting process behavior and reducing repeated model-development work.\", \"workflow\": \"A time-series transformer models plant processes to inform operators, engineers and scientists. Intended tasks include demand anticipation, process optimization and product-quality prediction.\", \"technology_data\": \"Large time-series datasets; distinct from the company’s industrial language model.\", \"integration_human_role\": \"Experts use model insights. The remarks do not establish autonomous execution or safety-system integration.\", \"deployment_scope\": \"No named plant, user count or deployment schedule disclosed.\", \"assessment\": \"Ambitious model reuse across process operations, with limited public implementation evidence. A pilot-to-production study would be needed to evaluate transferability across units.\", \"corrections_to_existing\": \"New workflow; related to the broader AI portfolio rather than a replacement for its record.\", \"related_existing_ids\": [\"AI-038\"], \"cohort\": \"root_backfill\", \"packet_id\": \"root_backfill-005\", \"source_ids\": [\"RS111\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-177\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-177", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025"], "findings": [{"severity": "medium", "field": "stage", "issue": "LEAP remarks use future-tense capability language for MetaBrain for Plants; no facility rollout/results established by this source.", "correction": "Stage as introduced/developing capability, not a verified plant deployment.", "source_url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-178", "Company": "ExxonMobil", "Company_primary": "ExxonMobil", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI monitoring of production bottlenecks at operating sites", "Extracted_finding": "Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.", "Stage_at_disclosure": "Company source dated 2025-10-23", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Company page describes practice at high level and does not verify legacy Q3 2025 transcript’s “Permian” scope. No data on bottlenecks addressed or realized output.", "Source_IDs": "RS112", "Source_date": "2025-10-23", "Source_type": "company webpage", "Source_locator": "Production bottleneck monitoring", "Source_URL": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-10-23; see chronology for operating dates.", "Context": "A bottleneck or mechanical issue in one production area can constrain connected processes and reduce throughput.", "Implementation": "ExxonMobil India says its AI teams work with production personnel to monitor equipment in real time, identify current or prospective bottlenecks, then coordinate with operators and production engineers to remove them.", "Assessment": "Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors_backfill-008-1", "Initiative_ID": "AI-178", "Title": "AI monitoring of production bottlenecks at operating sites", "Date": "2025-10-23", "Finding": "AI monitoring of production bottlenecks at operating sites documented or disclosed.", "Metric": "", "Source_IDs": "RS112", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-10-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-008"], "Research_search": " {\"key\": \"exxon-ops-monitoring\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-022\"], \"company\": \"ExxonMobil\", \"title\": \"AI monitoring of production bottlenecks at operating sites\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company source dated 2025-10-23\", \"practical_relevance\": \"Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.\", \"business_problem\": \"A bottleneck or mechanical issue in one production area can constrain connected processes and reduce throughput.\", \"workflow\": \"ExxonMobil India says its AI teams work with production personnel to monitor equipment in real time, identify current or prospective bottlenecks, then coordinate with operators and production engineers to remove them.\", \"technology_data\": \"AI applied to operational/equipment data; company content also describes connecting sensor and image data, but this source does not specify specific model, field, system or sensor architecture.\", \"integration_human_role\": \"Engineers and asset operators investigate AI flags against physical operating context; source explicitly pairs AI and people.\", \"deployment_scope\": \"ExxonMobil India describes company operations at a general level; no number of sites or assets stated.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No measured production uplift or cost result.\", \"timeline\": [{\"date\": \"2025-10-23\", \"event\": \"AI monitoring of production bottlenecks at operating sites documented or disclosed.\", \"source_ids\": [\"RS112\"]}], \"limitations\": [\"Company page describes practice at high level and does not verify legacy Q3 2025 transcript’s “Permian” scope.\", \"No data on bottlenecks addressed or realized output.\"], \"assessment\": \"Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.\", \"sources\": [{\"id\": \"RS112\", \"title\": \"ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence\", \"url\": \"https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence\", \"publisher\": \"ExxonMobil\", \"date\": \"2025-10-23\", \"type\": \"company webpage\", \"locator\": \"Production bottleneck monitoring\", \"review_status\": \"full text reviewed\", \"supports\": [\"ExxonMobil India describes using AI with production staff to monitor equipment in real time and identify current or prospective bottlenecks affecting production.\", \"Engineers and asset operators use domain knowledge to investigate and remove bottlenecks; no basin/site counts or metrics.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-010\"}], \"field_sources\": {\"business_problem\": [\"RS112\"], \"workflow\": [\"RS112\"], \"technology_data\": [\"RS112\"], \"integration_human_role\": [\"RS112\"], \"deployment_scope\": [\"RS112\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-008\", \"source_ids\": [\"RS112\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-178\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-178", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-179", "Company": "PETRONAS / AIngineer / Microsoft", "Company_primary": "PETRONAS", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "Industrial AI copilot integrated with engineering technical references", "Extracted_finding": "Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.", "Stage_at_disclosure": "Collaboration announced 2025-06-10; source says technical use cases and prior deployments, without naming sites", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "MoU and commercialization collaboration does not establish the extent of PETRONAS operational rollout. No named technical use case or measured outcome.", "Source_IDs": "RS113", "Source_date": "2025-06-10", "Source_type": "company announcement", "Source_locator": "AIngineer industrial engineer copilot", "Source_URL": "https://www.petronas.com/bm/node/1668", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-06-10; see chronology for operating dates.", "Context": "Engineers in complex process industries need rapid access to technical standards, databases and accumulated know-how when diagnosing reliability and maintenance issues.", "Implementation": "AIngineer copilot is embedded in everyday engineering workflow and connects technical databases/standards to provide answers and insight; collaboration with Microsoft supports cloud scaling/commercialization. PETRONAS says it has collaborated on multiple technical use cases.", "Assessment": "Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "majors_backfill-010-1", "Initiative_ID": "AI-179", "Title": "Industrial AI copilot integrated with engineering technical references", "Date": "2025-06-10", "Finding": "Industrial AI copilot integrated with engineering technical references documented or disclosed.", "Metric": "", "Source_IDs": "RS113", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-06-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["majors_backfill-010"], "Research_search": " {\"key\": \"petronas-aingineer\", \"existing_ids\": [], \"related_existing_ids\": [\"AI-079\"], \"company\": \"PETRONAS / AIngineer / Microsoft\", \"title\": \"Industrial AI copilot integrated with engineering technical references\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Collaboration announced 2025-06-10; source says technical use cases and prior deployments, without naming sites\", \"practical_relevance\": \"Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.\", \"business_problem\": \"Engineers in complex process industries need rapid access to technical standards, databases and accumulated know-how when diagnosing reliability and maintenance issues.\", \"workflow\": \"AIngineer copilot is embedded in everyday engineering workflow and connects technical databases/standards to provide answers and insight; collaboration with Microsoft supports cloud scaling/commercialization. PETRONAS says it has collaborated on multiple technical use cases.\", \"technology_data\": \"LLM plus machine learning algorithms; integration with technical databases/standards; Microsoft cloud infrastructure; named model versions/security architecture not disclosed.\", \"integration_human_role\": \"Designed to augment engineers’ decision-making and preserve institutional knowledge; release does not specify approval or safety controls for recommendations.\", \"deployment_scope\": \"Energy/heavy industry product; PETRONAS use cases mentioned but no asset, user count or independent customer deployment details.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [], \"economics\": \"No realized cost/diagnostic-time outcome quantified; lower maintenance costs and faster diagnosis are stated aims.\", \"timeline\": [{\"date\": \"2025-06-10\", \"event\": \"Industrial AI copilot integrated with engineering technical references documented or disclosed.\", \"source_ids\": [\"RS113\"]}], \"limitations\": [\"MoU and commercialization collaboration does not establish the extent of PETRONAS operational rollout.\", \"No named technical use case or measured outcome.\"], \"assessment\": \"Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.\", \"sources\": [{\"id\": \"RS113\", \"title\": \"PETRONAS-incubated AIngineer joins forces with Microsoft\", \"url\": \"https://www.petronas.com/bm/node/1668\", \"publisher\": \"PETRONAS\", \"date\": \"2025-06-10\", \"type\": \"company announcement\", \"locator\": \"AIngineer industrial engineer copilot\", \"review_status\": \"full text reviewed\", \"supports\": [\"AIngineer is a PETRONAS-incubated industrial AI copilot using LLMs and ML and integrating technical databases and standards.\", \"PETRONAS says it has technical use cases and prior deployments in energy/government; this release does not name a customer asset or quantify results.\", \"Microsoft collaboration includes cloud scaling and prospective marketplace commercialization.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"mb-013\"}], \"field_sources\": {\"business_problem\": [\"RS113\"], \"workflow\": [\"RS113\"], \"technology_data\": [\"RS113\"], \"integration_human_role\": [\"RS113\"], \"deployment_scope\": [\"RS113\"], \"economics\": [\"RS113\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"majors_backfill\", \"packet_id\": \"majors_backfill-010\", \"source_ids\": [\"RS113\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-179\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "source_url": ""}, {"claim": "Second-pass review", "reason": "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "source_url": ""}], "Audit_review": {"id": "AI-179", "company": "PETRONAS / AIngineer / Microsoft", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/bm/node/1668"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-180", "Company": "Indian Oil Corporation (IndianOil)", "Company_primary": "Indian Oil Corporation (IndianOil)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI/ML predictive maintenance across refineries and operating locations", "Extracted_finding": "A concrete reliability workflow is identified, but the public report does not name facilities, equipment classes or achieved downtime reduction.", "Stage_at_disclosure": "IndianOil’s FY2024-25 annual report says it is deploying AI and machine learning for predictive maintenance; site-specific rollout status is not disclosed.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The report gives a broad ongoing-deployment statement with no named site, asset class, model, status by asset or outcome data.", "Source_IDs": "RS116", "Source_date": "Undated source; see chronology", "Source_type": "annual report webpage", "Source_locator": "Leveraging Data and AI for Operational Efficiency", "Source_URL": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Improve asset reliability and reduce downtime at refineries and operating locations.", "Implementation": "IndianOil says it is deploying AI and machine learning for predictive maintenance. The report does not describe specific alarms, predictions or maintenance decisions.", "Assessment": "Analyst inference: predictive maintenance is operationally relevant, but disclosure is too broad to evaluate technical maturity or results.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Refining, LNG & process operations"], "Department_assignment_basis": "Workflow-based analyst classification: predictive maintenance for asset reliability and downtime reduction. IndianOil’s source does not name the responsible organizational department, asset owner, or buyer.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-001-1", "Initiative_ID": "AI-180", "Title": "AI/ML predictive maintenance across refineries and operating locations", "Date": "2025 (FY2024-25 annual report)", "Finding": "IndianOil reported ongoing deployment of AI/ML predictive maintenance.", "Metric": "", "Source_IDs": "RS116", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-001"], "Research_search": "{\"secondary_departments\": [\"Refining, LNG & process operations\"], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"indianoil-vishleshan-2\", \"existing_ids\": [], \"company\": \"Indian Oil Corporation (IndianOil)\", \"title\": \"AI/ML predictive maintenance across refineries and operating locations\", \"department\": \"Maintenance & reliability\", \"stage\": \"IndianOil’s FY2024-25 annual report says it is deploying AI and machine learning for predictive maintenance; site-specific rollout status is not disclosed.\", \"practical_relevance\": \"A concrete reliability workflow is identified, but the public report does not name facilities, equipment classes or achieved downtime reduction.\", \"business_problem\": \"Improve asset reliability and reduce downtime at refineries and operating locations.\", \"workflow\": \"IndianOil says it is deploying AI and machine learning for predictive maintenance. The report does not describe specific alarms, predictions or maintenance decisions.\", \"technology_data\": \"AI/ML is stated; data sources, model types and interfaces are not disclosed.\", \"integration_human_role\": \"Human review and work-order integration are not disclosed.\", \"deployment_scope\": \"Refineries and operating locations; site count and rollout stage are not disclosed.\", \"economics\": \"No measured uptime, downtime or financial impact disclosed.\", \"timeline\": [{\"date\": \"2025 (FY2024-25 annual report)\", \"event\": \"IndianOil reported ongoing deployment of AI/ML predictive maintenance.\", \"source_ids\": [\"RS116\"]}], \"limitations\": [\"The report gives a broad ongoing-deployment statement with no named site, asset class, model, status by asset or outcome data.\"], \"assessment\": \"Analyst inference: predictive maintenance is operationally relevant, but disclosure is too broad to evaluate technical maturity or results.\", \"sources\": [{\"id\": \"RS116\", \"title\": \"IndianOil – Intellectual Capital, Integrated Annual Report 2024-25\", \"url\": \"https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html\", \"publisher\": \"Indian Oil Corporation\", \"date\": \"2025 (FY2024-25)\", \"type\": \"annual report webpage\", \"locator\": \"Leveraging Data and AI for Operational Efficiency\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"IndianOil says it is deploying AI and machine learning for predictive maintenance in refineries and operating locations to improve reliability and reduce downtime.\"], \"original_id\": \"regional-001\"}], \"field_sources\": {\"business_problem\": [\"RS116\"], \"workflow\": [\"RS116\"], \"technology_data\": [\"RS116\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS116\"], \"economics\": []}, \"cohort\": \"regional\", \"packet_id\": \"regional-001\", \"source_ids\": [\"RS116\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-180\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-180", "company": "Indian Oil Corporation (IndianOil)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html"], "findings": [{"severity": "medium", "field": "detail", "issue": "IndianOil annual report supports AI/ML predictive maintenance across refineries/operating locations, but gives no site, asset, performance metric or spend.", "correction": "Keep description broad and do not infer a named tool, site or return.", "source_url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "Source describes predictive maintenance at refineries and “operating locations” without naming assets or sites. Refining is supported by the stated refinery scope, but broader coverage cannot be assigned without more detail.", "Segment_classification_confidence": "medium", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Department assigned from the disclosed predictive-maintenance workflow; IndianOil does not identify the organizational owner or buyer.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-181", "Company": "Bharat Petroleum Corporation (BPCL)", "Company_primary": "Bharat Petroleum Corporation (BPCL)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Refinery predictive alerts and AI video safety analytics", "Extracted_finding": "The projects put AI into two concrete refinery workflows: early warning for critical-equipment degradation and video-based safety/compliance monitoring. BPCL distinguishes implemented video coverage from the predictive system’s implementation stage.", "Stage_at_disclosure": "FY2024-25: video analytics implemented at all three refineries; predictive analytics initiated across all three, extending a successful Kochi trial", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "No measured reliability, safety or financial outcome is given. Predictive analytics was described as an initiated implementation, so full operational rollout is not established.", "Source_IDs": "RS117", "Source_date": "Undated source; see chronology", "Source_type": "annual report", "Source_locator": "Digital initiatives, p. 135", "Source_URL": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Undated evidence; an in-window disclosure date is not established.", "Context": "Reduce the chance that critical equipment failure or performance degradation causes lost production, and detect field safety hazards/non-compliance.", "Implementation": "A cloud-based predictive-analytics project is being implemented across Mumbai, Kochi and Bina refineries to issue early alerts for identified critical equipment; this extended a successful Kochi trial. Separately, Ultra Critical Video Analytics applies AI algorithms to field video for non-compliance and safety-hazard detection and integrates with BPCL’s Work Permit System.", "Assessment": "Analyst inference: combining asset alerts with existing work-permit controls is a practical maintenance-and-safety application, but the public report supports deployment status more strongly than achieved impact.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-002-1", "Initiative_ID": "AI-181", "Title": "Refinery predictive alerts and AI video safety analytics", "Date": "FY2024-25", "Finding": "Predictive analytics initiated across three refineries, extending Kochi trial; UCVA implemented across all three.", "Metric": "", "Source_IDs": "RS117", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-002"], "Research_search": " {\"secondary_departments\": [\"HSE & emissions\"], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"bpcl-refinery-reliability-and-safety-ai\", \"existing_ids\": [], \"company\": \"Bharat Petroleum Corporation (BPCL)\", \"title\": \"Refinery predictive alerts and AI video safety analytics\", \"department\": \"Maintenance & reliability\", \"stage\": \"FY2024-25: video analytics implemented at all three refineries; predictive analytics initiated across all three, extending a successful Kochi trial\", \"practical_relevance\": \"The projects put AI into two concrete refinery workflows: early warning for critical-equipment degradation and video-based safety/compliance monitoring. BPCL distinguishes implemented video coverage from the predictive system’s implementation stage.\", \"business_problem\": \"Reduce the chance that critical equipment failure or performance degradation causes lost production, and detect field safety hazards/non-compliance.\", \"workflow\": \"A cloud-based predictive-analytics project is being implemented across Mumbai, Kochi and Bina refineries to issue early alerts for identified critical equipment; this extended a successful Kochi trial. Separately, Ultra Critical Video Analytics applies AI algorithms to field video for non-compliance and safety-hazard detection and integrates with BPCL’s Work Permit System.\", \"technology_data\": \"Cloud analytics and predictive alerts are disclosed for the equipment workflow; video-feed algorithms and Work Permit System integration for UCVA. Models, sensors, alarm thresholds and cybersecurity architecture are not disclosed.\", \"integration_human_role\": \"The source describes alerts and work-permit integration but not automated control actions, escalation rules or human review responsibilities.\", \"deployment_scope\": \"Three refineries (Mumbai, Kochi, Bina). UCVA was implemented across all three; predictive reliability analytics was initiated across all three after a Kochi trial.\", \"economics\": \"No AI-attributed cost, uptime, incident or return metrics disclosed. BPCL also describes conventional Operations Driven Reliability, excluded from the AI workflow here.\", \"timeline\": [{\"date\": \"FY2024-25\", \"event\": \"Predictive analytics initiated across three refineries, extending Kochi trial; UCVA implemented across all three.\", \"source_ids\": [\"RS117\"]}], \"limitations\": [\"No measured reliability, safety or financial outcome is given. Predictive analytics was described as an initiated implementation, so full operational rollout is not established.\"], \"assessment\": \"Analyst inference: combining asset alerts with existing work-permit controls is a practical maintenance-and-safety application, but the public report supports deployment status more strongly than achieved impact.\", \"sources\": [{\"id\": \"RS117\", \"title\": \"Annual Report 2024-25\", \"url\": \"https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf\", \"publisher\": \"Bharat Petroleum\", \"date\": \"2025 (FY2024-25)\", \"type\": \"annual report\", \"locator\": \"Digital initiatives, p. 135\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"AI predictive analytics for asset reliability/performance was initiated across Mumbai, Kochi and Bina refineries, extending a successful Kochi trial; cloud predictive alerts target failures/degradation in identified critical equipment.\", \"AI video algorithms detect non-compliance and safety hazards from field video; Ultra Critical Video Analytics was implemented across all three refineries and integrated with the in-house Work Permit System.\"], \"original_id\": \"regional-002\"}], \"field_sources\": {\"business_problem\": [\"RS117\"], \"workflow\": [\"RS117\"], \"technology_data\": [\"RS117\"], \"integration_human_role\": [\"RS117\"], \"deployment_scope\": [\"RS117\"], \"economics\": []}, \"cohort\": \"regional\", \"packet_id\": \"regional-002\", \"source_ids\": [\"RS117\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-181\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-181", "company": "Bharat Petroleum Corporation (BPCL)", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, "Audit_outcome": "unverified", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "scaled", "Scale_note": "3 refineries", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. Re-opened 5 Oct 2026 (HTTP 200) but the AI passage could not be text-extracted; claims not re-read.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-182", "Company": "Hindustan Petroleum Corporation (HPCL)", "Company_primary": "Hindustan Petroleum Corporation (HPCL)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI-enabled LC-Max digital suite at Visakh residue-upgradation facility", "Extracted_finding": "This is a named AI-enabled suite tied to a major residue-hydrocracking unit. It combines monitoring, predictive analytics and optimization; public disclosure does not separate the suite’s contribution from the physical process technology’s performance.", "Stage_at_disclosure": "Deployed with the facility commissioning announced January 2026", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "AI contribution to process performance is not isolated; model/data/control details and quantified AI-specific outcomes are absent.", "Source_IDs": "RS118", "Source_date": "2026-01-05", "Source_type": "company announcement", "Source_locator": "paragraphs 24-28", "Source_URL": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-01-05; see chronology for operating dates.", "Context": "Stabilize and optimize operation of the high-complexity residue-upgradation process.", "Implementation": "At commissioning of the Visakh RUF, HPCL said the digital suite integrates real-time monitoring, predictive analytics based on proprietary thermodynamic models and AI-driven optimization.", "Assessment": "Analyst inference: the integration of predictive analytics with process optimization in a newly commissioned complex unit is operationally relevant, though public evidence does not establish autonomy or incremental AI value.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-003-1", "Initiative_ID": "AI-182", "Title": "AI-enabled LC-Max digital suite at Visakh residue-upgradation facility", "Date": "2026-01-05", "Finding": "HPCL announced RUF commissioning and deployment of its LC-Max Digital Suite.", "Metric": "", "Source_IDs": "RS118", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-01-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-003"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"hpcl-visakh-lcmax-digital-suite\", \"existing_ids\": [], \"company\": \"Hindustan Petroleum Corporation (HPCL)\", \"title\": \"AI-enabled LC-Max digital suite at Visakh residue-upgradation facility\", \"department\": \"Refining, LNG & process operations\", \"stage\": \"Deployed with the facility commissioning announced January 2026\", \"practical_relevance\": \"This is a named AI-enabled suite tied to a major residue-hydrocracking unit. It combines monitoring, predictive analytics and optimization; public disclosure does not separate the suite’s contribution from the physical process technology’s performance.\", \"business_problem\": \"Stabilize and optimize operation of the high-complexity residue-upgradation process.\", \"workflow\": \"At commissioning of the Visakh RUF, HPCL said the digital suite integrates real-time monitoring, predictive analytics based on proprietary thermodynamic models and AI-driven optimization.\", \"technology_data\": \"Proprietary thermodynamic models and AI are named. Sensor coverage, model architecture, control integration and operating mode are not described.\", \"integration_human_role\": \"Human approval, safeguards, alert routing and whether optimization is advisory or closed-loop are not disclosed.\", \"deployment_scope\": \"The RUF LC-Max Digital Suite at HPCL’s Visakh Refinery.\", \"economics\": \"The announcement reports physical-unit capacity and conversion/yield characteristics, but does not attribute these outcomes to AI. AI-specific economics are not disclosed.\", \"timeline\": [{\"date\": \"2026-01-05\", \"event\": \"HPCL announced RUF commissioning and deployment of its LC-Max Digital Suite.\", \"source_ids\": [\"RS118\"]}], \"limitations\": [\"AI contribution to process performance is not isolated; model/data/control details and quantified AI-specific outcomes are absent.\"], \"assessment\": \"Analyst inference: the integration of predictive analytics with process optimization in a newly commissioned complex unit is operationally relevant, though public evidence does not establish autonomy or incremental AI value.\", \"sources\": [{\"id\": \"RS118\", \"title\": \"HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery\", \"url\": \"https://www.hindustanpetroleum.com/NewsroomDetails/449\", \"publisher\": \"Hindustan Petroleum\", \"date\": \"2026-01-05\", \"type\": \"company announcement\", \"locator\": \"paragraphs 24-28\", \"review_status\": \"full text reviewed\", \"supports\": [\"At commissioning of the Visakh residue-upgradation facility, HPCL said its LC-Max Digital Suite integrates real-time monitoring, predictive analytics using proprietary thermodynamic models and AI-driven optimization.\", \"The source reports RUF physical-process conversion/yield figures separately; it does not attribute those outcomes to AI.\"], \"original_id\": \"regional-003\"}], \"field_sources\": {\"business_problem\": [\"RS118\"], \"workflow\": [\"RS118\"], \"technology_data\": [\"RS118\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS118\"], \"economics\": [\"RS118\"]}, \"cohort\": \"regional\", \"packet_id\": \"regional-003\", \"source_ids\": [\"RS118\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-182\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-182", "company": "Hindustan Petroleum Corporation (HPCL)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.hindustanpetroleum.com/NewsroomDetails/449"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-183", "Company": "Reliance Industries Limited (RIL)", "Company_primary": "Reliance Industries Limited (RIL)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI-driven crude-feedstock blend selection for O2C", "Extracted_finding": "Crude selection and blending are directly linked to refinery economics and operating constraints. Reliance names a deployed decision-support tool but does not disclose its operating scope or measured value.", "Stage_at_disclosure": "Described by the chairman as an existing proprietary tool in the June 2026 shareholder statement", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The public statement lacks technical detail, deployment scale, quantified results and attribution methodology.", "Source_IDs": "RS119", "Source_date": "2026-06", "Source_type": "shareholder statement", "Source_locator": "O2C business performance", "Source_URL": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-06; see chronology for operating dates.", "Context": "Select efficient crude blends for the Oil-to-Chemicals system.", "Implementation": "Reliance says its proprietary AI-driven feedstock-optimization tool helps select efficient crude blends; a broader statement says AI process optimization improves O2C yield and reduces energy use.", "Assessment": "Analyst inference: blend selection is a high-value, repeatable refinery planning use case; public disclosure establishes existence of the tool but not its effect size or integration depth.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-004-1", "Initiative_ID": "AI-183", "Title": "AI-driven crude-feedstock blend selection for O2C", "Date": "2026-06", "Finding": "Chairman’s statement describes AI feedstock optimization and O2C process optimization.", "Metric": "", "Source_IDs": "RS119", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-06", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-004"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"ril-crude-feedstock-optimizer\", \"existing_ids\": [], \"company\": \"Reliance Industries Limited (RIL)\", \"title\": \"AI-driven crude-feedstock blend selection for O2C\", \"department\": \"Refining, LNG & process operations\", \"stage\": \"Described by the chairman as an existing proprietary tool in the June 2026 shareholder statement\", \"practical_relevance\": \"Crude selection and blending are directly linked to refinery economics and operating constraints. Reliance names a deployed decision-support tool but does not disclose its operating scope or measured value.\", \"business_problem\": \"Select efficient crude blends for the Oil-to-Chemicals system.\", \"workflow\": \"Reliance says its proprietary AI-driven feedstock-optimization tool helps select efficient crude blends; a broader statement says AI process optimization improves O2C yield and reduces energy use.\", \"technology_data\": \"AI and proprietary tool are identified, but model type, crude assays, unit constraints, planning integration and data refresh are not disclosed.\", \"integration_human_role\": \"The source says the tool helps select blends; final decision authority or operator approval is not described.\", \"deployment_scope\": \"Reliance O2C operations are referenced, with no site or asset count disclosed.\", \"economics\": \"Company states the broader O2C AI optimization is improving yield and reducing energy, but gives no quantity, period-specific baseline or AI-specific financial impact.\", \"timeline\": [{\"date\": \"2026-06\", \"event\": \"Chairman’s statement describes AI feedstock optimization and O2C process optimization.\", \"source_ids\": [\"RS119\"]}], \"limitations\": [\"The public statement lacks technical detail, deployment scale, quantified results and attribution methodology.\"], \"assessment\": \"Analyst inference: blend selection is a high-value, repeatable refinery planning use case; public disclosure establishes existence of the tool but not its effect size or integration depth.\", \"sources\": [{\"id\": \"RS119\", \"title\": \"Chairman’s Statement, 49th Annual General Meeting\", \"url\": \"https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf\", \"publisher\": \"Reliance Industries\", \"date\": \"2026-06\", \"type\": \"shareholder statement\", \"locator\": \"O2C business performance\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Reliance says its proprietary AI-driven feedstock-optimization tool helps select efficient crude blends.\", \"The same statement says AI-driven O2C process optimization improves yield and reduces energy; no quantified result, control design or site scope is given.\"], \"original_id\": \"regional-004\"}], \"field_sources\": {\"business_problem\": [\"RS119\"], \"workflow\": [\"RS119\"], \"technology_data\": [\"RS119\"], \"integration_human_role\": [\"RS119\"], \"deployment_scope\": [\"RS119\"], \"economics\": [\"RS119\"]}, \"cohort\": \"regional\", \"packet_id\": \"regional-004\", \"source_ids\": [\"RS119\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-183\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-183", "company": "Reliance Industries Limited (RIL)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf"], "findings": [{"severity": "medium", "field": "economics", "issue": "RIL chairman statement names the AI feedstock-blend tool but discloses no method, scale or measured benefit.", "correction": "Treat workflow as company-reported and economics/outcome as undisclosed.", "source_url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-184", "Company": "PTT Global Chemical (PTTGC / GC)", "Company_primary": "PTT Global Chemical (PTTGC / GC)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Smart Plant AI: plant-wide optimization and predictive maintenance", "Extracted_finding": "Analyst inference: these workflows target plant-wide operating efficiency and equipment reliability.", "Stage_at_disclosure": "GC reported the Smart Plant workflows in July 2026; site-level rollout is unspecified.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Rollout counts and measured outcomes are not disclosed.", "Source_IDs": "RS120", "Source_date": "2026-07-20", "Source_type": "company announcement", "Source_locator": "Smart Plant / Supply Chain & Operations Optimization", "Source_URL": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-07-20; see chronology for operating dates.", "Context": "Improve production efficiency and equipment reliability.", "Implementation": "Plant-wide optimization uses AI and digital plant models in real-time operating optimization. AI/ML analyze equipment-condition data for anomaly/failure prediction; AI vision, drones and robotics support inspection.", "Assessment": "Analyst inference: coupling complex-wide optimization with condition-based failure prediction is a relatively mature operational portfolio; the evidence is company-reported and lacks measurable technical/financial proof.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-005-1", "Initiative_ID": "AI-184", "Title": "Smart Plant AI: plant-wide optimization and predictive maintenance", "Date": "2026-07-20", "Finding": "GC detailed Smart Plant AI initiatives in a company announcement.", "Metric": "", "Source_IDs": "RS120", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-07-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-005"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"pttgc-smart-plant-ai\", \"existing_ids\": [], \"company\": \"PTT Global Chemical (PTTGC / GC)\", \"title\": \"Smart Plant AI: plant-wide optimization and predictive maintenance\", \"department\": \"Maintenance & reliability\", \"stage\": \"GC reported the Smart Plant workflows in July 2026; site-level rollout is unspecified.\", \"practical_relevance\": \"Analyst inference: these workflows target plant-wide operating efficiency and equipment reliability.\", \"business_problem\": \"Improve production efficiency and equipment reliability.\", \"workflow\": \"Plant-wide optimization uses AI and digital plant models in real-time operating optimization. AI/ML analyze equipment-condition data for anomaly/failure prediction; AI vision, drones and robotics support inspection.\", \"technology_data\": \"AI, digital plant models, equipment analytics, computer vision and robotics are named.\", \"integration_human_role\": \"Approval, maintenance-system integration and closed-loop control are not disclosed.\", \"deployment_scope\": \"GC production complex; site count not stated.\", \"economics\": \"No AI-workflow-specific savings or results disclosed.\", \"timeline\": [{\"date\": \"2026-07-20\", \"event\": \"GC detailed Smart Plant AI initiatives in a company announcement.\", \"source_ids\": [\"RS120\"]}], \"limitations\": [\"Rollout counts and measured outcomes are not disclosed.\"], \"assessment\": \"Analyst inference: coupling complex-wide optimization with condition-based failure prediction is a relatively mature operational portfolio; the evidence is company-reported and lacks measurable technical/financial proof.\", \"sources\": [{\"id\": \"RS120\", \"title\": \"GC Advances Holistic Optimization Through Digital Technology\", \"url\": \"https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness\", \"publisher\": \"PTT Global Chemical\", \"date\": \"2026-07-20\", \"type\": \"company announcement\", \"locator\": \"Smart Plant / Supply Chain & Operations Optimization\", \"review_status\": \"full text reviewed\", \"supports\": [\"GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale.\"], \"original_id\": \"regional-005\"}], \"field_sources\": {\"business_problem\": [\"RS120\"], \"workflow\": [\"RS120\"], \"technology_data\": [\"RS120\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS120\"], \"economics\": [\"RS120\"]}, \"cohort\": \"regional\", \"packet_id\": \"regional-005\", \"source_ids\": [\"RS120\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-184\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-184", "company": "PTT Global Chemical (PTTGC / GC)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-185", "Company": "PTT Global Chemical (PTTGC / GC)", "Company_primary": "PTT Global Chemical (PTTGC / GC)", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "HERO AI-enabled heat-exchanger-network energy optimization", "Extracted_finding": "Analyst inference: a defined AI/simulation application to process energy efficiency.", "Stage_at_disclosure": "Industrial-scale deployment reported by GC in July 2026.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Deployment and “first in Southeast Asia” claims are company-reported; no baseline or quantified result.", "Source_IDs": "RS120", "Source_date": "2026-07-20", "Source_type": "company announcement", "Source_locator": "Smart Plant / Supply Chain & Operations Optimization", "Source_URL": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "Evidence_status": "partially verified primary (audit: source supports some, not all, profile claims)", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-07-20; see chronology for operating dates.", "Context": "Optimize heat integration to reduce process energy use.", "Implementation": "HERO uses AI and process simulation to optimize heat integration across production processes.", "Assessment": "Analyst inference: AI plus process simulation applied to heat integration is a clearly scoped industrial workflow, but effect size and replication remain unverified.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-006-1", "Initiative_ID": "AI-185", "Title": "HERO AI-enabled heat-exchanger-network energy optimization", "Date": "2026-07-20", "Finding": "GC reported industrial-scale deployment of HERO.", "Metric": "", "Source_IDs": "RS120", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-07-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-006"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"pttgc-hero-heat-integration\", \"existing_ids\": [], \"company\": \"PTT Global Chemical (PTTGC / GC)\", \"title\": \"HERO AI-enabled heat-exchanger-network energy optimization\", \"department\": \"Refining, LNG & process operations\", \"stage\": \"Industrial-scale deployment reported by GC in July 2026.\", \"practical_relevance\": \"Analyst inference: a defined AI/simulation application to process energy efficiency.\", \"business_problem\": \"Optimize heat integration to reduce process energy use.\", \"workflow\": \"HERO uses AI and process simulation to optimize heat integration across production processes.\", \"technology_data\": \"AI and process simulation; data and control details are undisclosed.\", \"integration_human_role\": \"Approval and automatic-control status are undisclosed.\", \"deployment_scope\": \"GC says deployed at industrial scale; sites are not named.\", \"economics\": \"No HERO-specific energy or financial result disclosed.\", \"timeline\": [{\"date\": \"2026-07-20\", \"event\": \"GC reported industrial-scale deployment of HERO.\", \"source_ids\": [\"RS120\"]}], \"limitations\": [\"Deployment and “first in Southeast Asia” claims are company-reported; no baseline or quantified result.\"], \"assessment\": \"Analyst inference: AI plus process simulation applied to heat integration is a clearly scoped industrial workflow, but effect size and replication remain unverified.\", \"sources\": [{\"id\": \"RS120\", \"title\": \"GC Advances Holistic Optimization Through Digital Technology\", \"url\": \"https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness\", \"publisher\": \"PTT Global Chemical\", \"date\": \"2026-07-20\", \"type\": \"company announcement\", \"locator\": \"Smart Plant / Supply Chain & Operations Optimization\", \"review_status\": \"full text reviewed\", \"supports\": [\"GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale.\"], \"original_id\": \"regional-005\"}], \"field_sources\": {\"business_problem\": [\"RS120\"], \"workflow\": [\"RS120\"], \"technology_data\": [\"RS120\"], \"integration_human_role\": [], \"deployment_scope\": [\"RS120\"], \"economics\": []}, \"cohort\": \"regional\", \"packet_id\": \"regional-006\", \"source_ids\": [\"RS120\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-185\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "partially verified primary", "Audit_flags": [{"claim": "Second-pass review", "reason": "Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim.", "source_url": ""}], "Audit_review": {"id": "AI-185", "company": "PTT Global Chemical (PTTGC / GC)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, "Audit_outcome": "partially supported", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "partially verified primary", "claim_status": "partially verified primary (audit: source supports some, not all, profile claims)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-186", "Company": "Rosneft", "Company_primary": "Rosneft", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "EchoTools AI well-diagnostic interpretation and operating-mode recommendations", "Extracted_finding": "EchoTools converts acoustic well-survey interpretation into faster well diagnostics and candidate operating-mode changes. The 2026 disclosure indicates use on thousands of wells and a ranked optimization queue; Rosneft reports production gains but no independent validation.", "Stage_at_disclosure": "Introduced July 2025; rollout and operating results updated in August 2026", "Reported_metric": "well indicators interpreted: more than 10,000 in about 2 minutes, versus more than 3 days manually (Company-reported); additional production per well: average 1 tonne of oil/day per well (Company-reported); wells analyzed / optimization candidates: more than 2,000 analyzed; over 500 ranked for planned operating-mode optimization (Company-reported)", "Interpretation_limit": "All performance and rollout figures are company-reported; methods, baselines and independent validation are not disclosed. The average daily-per-well uplift is not tied to a stated measurement period or cohort definition.", "Source_IDs": "RS121; RS122", "Source_date": "2026-08-20", "Source_type": "company announcement", "Source_locator": "EchoTools", "Source_URL": "https://rosneft.ru/press/news/item/222577", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-08-20; see chronology for operating dates.", "Context": "Speed up interpretation of well annulus fluid-level/acoustic data and identify operating-mode changes that may increase production.", "Implementation": "AI determines fluid level and sound speed in the well annulus; these values feed bottomhole-pressure calculations and recommendations for well operating mode. The August 2026 update describes screening over 2,000 wells and ranking over 500 for planned optimization.", "Assessment": "Analyst inference: among screened regional cases, EchoTools provides one of the clearest workflows and implementation metrics; evidence supports decision assistance and prioritization, not autonomous operation.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-007-1", "Initiative_ID": "AI-186", "Title": "EchoTools AI well-diagnostic interpretation and operating-mode recommendations", "Date": "2025-07-17", "Finding": "EchoTools AI workflow publicly described; this is the initial disclosure identified in the screen.", "Metric": "", "Source_IDs": "RS121", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "regional-007-2", "Initiative_ID": "AI-186", "Title": "EchoTools AI well-diagnostic interpretation and operating-mode recommendations", "Date": "2026-08-20", "Finding": "Company reported first-half 2026 rollout/update metrics.", "Metric": "", "Source_IDs": "RS122", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-08-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-007"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [{\"metric\": \"well indicators interpreted\", \"value\": \"more than 10,000 in about 2 minutes, versus more than 3 days manually\", \"period\": \"July 2025 disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS121\"]}, {\"metric\": \"additional production per well\", \"value\": \"average 1 tonne of oil/day per well\", \"period\": \"July 2025 disclosure; evaluation basis not given\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS121\"]}, {\"metric\": \"wells analyzed / optimization candidates\", \"value\": \"more than 2,000 analyzed; over 500 ranked for planned operating-mode optimization\", \"period\": \"1H 2026\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS122\"]}], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"rosneft-echotools\", \"existing_ids\": [], \"company\": \"Rosneft\", \"title\": \"EchoTools AI well-diagnostic interpretation and operating-mode recommendations\", \"department\": \"Production & field operations\", \"stage\": \"Introduced July 2025; rollout and operating results updated in August 2026\", \"practical_relevance\": \"EchoTools converts acoustic well-survey interpretation into faster well diagnostics and candidate operating-mode changes. The 2026 disclosure indicates use on thousands of wells and a ranked optimization queue; Rosneft reports production gains but no independent validation.\", \"business_problem\": \"Speed up interpretation of well annulus fluid-level/acoustic data and identify operating-mode changes that may increase production.\", \"workflow\": \"AI determines fluid level and sound speed in the well annulus; these values feed bottomhole-pressure calculations and recommendations for well operating mode. The August 2026 update describes screening over 2,000 wells and ranking over 500 for planned optimization.\", \"technology_data\": \"AI software; model details, telemetry integration, uncertainty handling and production-system interfaces are not disclosed.\", \"integration_human_role\": \"The software recommends operating adjustments and identifies candidate wells; the public source does not describe approval, field execution or safety-control responsibilities.\", \"deployment_scope\": \"Rosneft reports over 2,000 wells analyzed in 1H 2026 and over 500 shortlisted for planned mode optimization; broader rollout scope is not stated.\", \"economics\": \"Rosneft reports additional oil output but gives no realized financial value, deployment cost or independent verification.\", \"timeline\": [{\"date\": \"2025-07-17\", \"event\": \"EchoTools AI workflow publicly described; this is the initial disclosure identified in the screen.\", \"source_ids\": [\"RS121\"]}, {\"date\": \"2026-08-20\", \"event\": \"Company reported first-half 2026 rollout/update metrics.\", \"source_ids\": [\"RS122\"]}], \"limitations\": [\"All performance and rollout figures are company-reported; methods, baselines and independent validation are not disclosed.\", \"The average daily-per-well uplift is not tied to a stated measurement period or cohort definition.\"], \"assessment\": \"Analyst inference: among screened regional cases, EchoTools provides one of the clearest workflows and implementation metrics; evidence supports decision assistance and prioritization, not autonomous operation.\", \"sources\": [{\"id\": \"RS121\", \"title\": \"Rosneft improves well research methods (EchoTools)\", \"url\": \"https://rosneft.ru/press/news/item/222577\", \"publisher\": \"Rosneft\", \"date\": \"2025-07-17\", \"type\": \"company announcement\", \"locator\": \"EchoTools\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"AI interprets fluid level and sound speed in the well annulus; the derived values support bottomhole-pressure calculation and selection of an operating mode.\", \"Rosneft says the software processes over 10,000 well indicators in about two minutes versus over three days manually, and recommendations yield an average additional 1 tonne/day per well.\"], \"original_id\": \"regional-006\"}, {\"id\": \"RS122\", \"title\": \"Rosneft increases production through innovative software (EchoTools update)\", \"url\": \"https://www.rosneft.ru/press/news/item/224545/\", \"publisher\": \"Rosneft\", \"date\": \"2026-08-20\", \"type\": \"company announcement\", \"locator\": \"EchoTools rollout update\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Rosneft reports that in 1H 2026 EchoTools analyzed information from more than 2,000 wells and generated a ranking of over 500 candidate wells for operating-mode optimization.\", \"This is a later rollout update to the July 2025 named workflow, not its launch date.\"], \"original_id\": \"regional-007\"}], \"field_sources\": {\"business_problem\": [\"RS121\"], \"workflow\": [\"RS121\", \"RS122\"], \"technology_data\": [\"RS121\"], \"integration_human_role\": [\"RS121\"], \"deployment_scope\": [\"RS122\"], \"economics\": [\"RS121\"]}, \"cohort\": \"regional\", \"packet_id\": \"regional-007\", \"source_ids\": [\"RS121\", \"RS122\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-186\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-186", "company": "Rosneft", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, "Audit_outcome": "unverified", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. rosneft.ru did not respond on 5 Oct 2026.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-187", "Company": "Rosneft", "Company_primary": "Rosneft", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "LLM virtual assistant for technical text search and documentation", "Extracted_finding": "This is a named internal GenAI workflow applied to domain work from geology through field-development design. The disclosed functionality remains knowledge-work assistance rather than autonomous engineering decisions.", "Stage_at_disclosure": "Company reported active use in July 2025", "Reported_metric": "routine workload reduction: 70-80% (Company-reported)", "Interpretation_limit": "Company-reported workload reduction has no method or independent validation; use volume and controls are undisclosed.", "Source_IDs": "RS123", "Source_date": "2025-07-28", "Source_type": "company announcement", "Source_locator": "LLM virtual assistant", "Source_URL": "https://rosneft.ru/press/news/item/222604/", "Evidence_status": "Cited source inaccessible on audit; claims not reverified", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-07-28; see chronology for operating dates.", "Context": "Reduce routine effort in finding, checking and formatting technical text information.", "Implementation": "Employees submit simple text queries to a company-developed LLM assistant used for tasks ranging from geology to field-development design; Rosneft says it helps search, verify and format text information.", "Assessment": "Analyst inference: useful internal productivity application, but less operationally consequential and less auditable than the company’s well-diagnostic deployment.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "regional-008-1", "Initiative_ID": "AI-187", "Title": "LLM virtual assistant for technical text search and documentation", "Date": "2025-07-28", "Finding": "Rosneft announced the deployed LLM-based virtual assistant.", "Metric": "", "Source_IDs": "RS123", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-07-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["regional-008"], "Research_search": " {\"secondary_departments\": [], \"category\": \"Internal adoption\", \"disclosed_results\": [{\"metric\": \"routine workload reduction\", \"value\": \"70-80%\", \"period\": \"July 2025 company disclosure\", \"attribution\": \"Company-reported\", \"source_ids\": [\"RS123\"]}], \"partners\": [], \"verification\": \"verified primary\", \"key\": \"rosneft-llm-geology-assistant\", \"existing_ids\": [], \"company\": \"Rosneft\", \"title\": \"LLM virtual assistant for technical text search and documentation\", \"department\": \"Reservoir engineering & field development\", \"stage\": \"Company reported active use in July 2025\", \"practical_relevance\": \"This is a named internal GenAI workflow applied to domain work from geology through field-development design. The disclosed functionality remains knowledge-work assistance rather than autonomous engineering decisions.\", \"business_problem\": \"Reduce routine effort in finding, checking and formatting technical text information.\", \"workflow\": \"Employees submit simple text queries to a company-developed LLM assistant used for tasks ranging from geology to field-development design; Rosneft says it helps search, verify and format text information.\", \"technology_data\": \"Large language models are stated. Model, corpus, retrieval architecture, access controls and validation process are not disclosed.\", \"integration_human_role\": \"The assistant responds to user text queries; final engineering decisions remain unspecified. Human review should be assumed only as an analyst inference, not an explicit company control statement.\", \"deployment_scope\": \"Used across unspecified company tasks including geology and field-development design; user count and business-unit rollout are not disclosed.\", \"economics\": \"No spend, productivity baseline, headcount effect or financial return disclosed.\", \"timeline\": [{\"date\": \"2025-07-28\", \"event\": \"Rosneft announced the deployed LLM-based virtual assistant.\", \"source_ids\": [\"RS123\"]}], \"limitations\": [\"Company-reported workload reduction has no method or independent validation; use volume and controls are undisclosed.\"], \"assessment\": \"Analyst inference: useful internal productivity application, but less operationally consequential and less auditable than the company’s well-diagnostic deployment.\", \"sources\": [{\"id\": \"RS123\", \"title\": \"Rosneft introduces new technologies using artificial intelligence\", \"url\": \"https://rosneft.ru/press/news/item/222604/\", \"publisher\": \"Rosneft\", \"date\": \"2025-07-28\", \"type\": \"company announcement\", \"locator\": \"LLM virtual assistant\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"Company-developed LLM-based virtual assistant is used for text search, verification and formatting tasks, with applications from geology to field-development design.\", \"Rosneft claims 70-80% lower routine workload for these tasks; no independent validation or financial return is disclosed.\"], \"original_id\": \"regional-008\"}], \"field_sources\": {\"business_problem\": [\"RS123\"], \"workflow\": [\"RS123\"], \"technology_data\": [\"RS123\"], \"integration_human_role\": [\"RS123\"], \"deployment_scope\": [\"RS123\"], \"economics\": [\"RS123\"]}, \"cohort\": \"regional\", \"packet_id\": \"regional-008\", \"source_ids\": [\"RS123\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-187\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "source inaccessible on audit", "Audit_flags": [{"claim": "Second-pass review", "reason": "One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state.", "source_url": ""}], "Audit_review": {"id": "AI-187", "company": "Rosneft", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, "Audit_outcome": "unverified", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The project's 27 September 2026 audit logged the primary source as inaccessible; claims are not reverified. rosneft.ru did not respond on 5 Oct 2026.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "source inaccessible on audit", "claim_status": "Cited source inaccessible on audit; claims not reverified", "independent_validation": "Not established by this collection"}}, {"ID": "AI-188", "Company": "Permian Resources", "Company_primary": "Permian Resources", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI-driven subsurface insight and precision wellbore placement", "Extracted_finding": "A concrete subsurface-to-well-placement workflow intended to expand play boundaries. The source does not establish its model design, deployment extent, or causal contribution to well outcomes.", "Stage_at_disclosure": "Described in Q3 2025 earnings presentation; operating scale and asset-specific deployment are not quantified.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Single company presentation disclosure, with no model, data, validation, or deployment-count detail. Do not attribute adjacent Haley outperformance to AI.", "Source_IDs": "RS124", "Source_date": "2025-11-05", "Source_type": "investor presentation", "Source_locator": "Slide 5, “Leading Execution, Cutting Edge Technology Drive Robust Results”", "Source_URL": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-11-05; see chronology for operating dates.", "Context": "Expanding productive play boundaries requires subsurface interpretation and precise wellbore placement.", "Implementation": "Permian Resources says it uses AI-driven subsurface insights with precision wellbore placement. The presentation places the claim in a discussion of technical execution, next to its Haley development update, but does not specify a decision sequence, geological inputs, or whether AI recommendations are advisory or automated.", "Assessment": "Analyst inference: the described workflow is decision-support for subsurface interpretation and wellbore placement, potentially useful for testing play extensions. Practical value and operational maturity cannot be established from the slide alone.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["Exploration & geoscience", "Drilling & completions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-1-001-1", "Initiative_ID": "AI-188", "Title": "AI-driven subsurface insight and precision wellbore placement", "Date": "2025-11-05", "Finding": "Q3 2025 investor presentation disclosed AI-driven subsurface insights and precision wellbore placement.", "Metric": "", "Source_IDs": "RS124", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-11-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-1-001"], "Research_search": " {\"key\": \"permian-resources-ai-subsurface-placement\", \"existing_ids\": [], \"company\": \"Permian Resources\", \"title\": \"AI-driven subsurface insight and precision wellbore placement\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [\"Exploration & geoscience\", \"Drilling & completions\"], \"category\": \"Internal adoption\", \"stage\": \"Described in Q3 2025 earnings presentation; operating scale and asset-specific deployment are not quantified.\", \"practical_relevance\": \"A concrete subsurface-to-well-placement workflow intended to expand play boundaries. The source does not establish its model design, deployment extent, or causal contribution to well outcomes.\", \"business_problem\": \"Expanding productive play boundaries requires subsurface interpretation and precise wellbore placement.\", \"workflow\": \"Permian Resources says it uses AI-driven subsurface insights with precision wellbore placement. The presentation places the claim in a discussion of technical execution, next to its Haley development update, but does not specify a decision sequence, geological inputs, or whether AI recommendations are advisory or automated.\", \"technology_data\": \"The company labels the insight AI-driven. It does not disclose model class, training data, geophysical/seismic inputs, validation approach, software, or external technology partners in the cited slide.\", \"integration_human_role\": \"Not disclosed. The source does not say who reviews the insight, how geoscientists or drilling engineers approve placement, or whether the system controls execution.\", \"deployment_scope\": \"Company presentation refers to play boundary expansion and discusses Permian operations. Number of wells/acreage or specific AI-assisted placement outcomes is not disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No AI-specific economics disclosed. Haley wells were reported to outperform offsets by ~45%, but the presentation does not attribute this result to the AI workflow; it is excluded from AI results.\", \"timeline\": [{\"date\": \"2025-11-05\", \"event\": \"Q3 2025 investor presentation disclosed AI-driven subsurface insights and precision wellbore placement.\", \"source_ids\": [\"RS124\"]}], \"limitations\": [\"Single company presentation disclosure, with no model, data, validation, or deployment-count detail.\", \"Do not attribute adjacent Haley outperformance to AI.\"], \"assessment\": \"Analyst inference: the described workflow is decision-support for subsurface interpretation and wellbore placement, potentially useful for testing play extensions. Practical value and operational maturity cannot be established from the slide alone.\", \"sources\": [{\"id\": \"RS124\", \"title\": \"Q3 2025 Earnings Presentation\", \"url\": \"https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf\", \"publisher\": \"Permian Resources\", \"date\": \"2025-11-05\", \"type\": \"investor presentation\", \"locator\": \"Slide 5, “Leading Execution, Cutting Edge Technology Drive Robust Results”\", \"review_status\": \"full text reviewed\", \"supports\": [\"Permian Resources said it was leveraging AI-driven subsurface insights and precision wellbore placement to expand play boundaries.\", \"The slide places this statement alongside its Haley development discussion; it separately reports Haley early-Q3 2025 startup and approximately 45% offset outperformance, without attributing that result to AI.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"a1-001\"}], \"field_sources\": {\"business_problem\": [\"RS124\"], \"workflow\": [\"RS124\"], \"technology_data\": [\"RS124\"], \"integration_human_role\": [\"RS124\"], \"deployment_scope\": [\"RS124\"], \"economics\": [\"RS124\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"\", \"cohort\": \"round2/additions-1\", \"packet_id\": \"round2-additions-1-001\", \"source_ids\": [\"RS124\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-188\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-188", "company": "Permian Resources", "key": "permian-resources-ai-subsurface-placement", "outcome": "Independent review completed", "reviewer_outcome": "supported with disclosure limits", "findings": [{"field": "Independent review", "issue": "The November 5, 2025 presentation, printed slide 5, explicitly describes AI-driven subsurface insight and precision wellbore placement. Adjacent Haley outperformance is not attributable to AI. Corrected department to reservoir engineering with exploration and drilling secondary links.", "correction": "", "source_url": ""}], "residual_gaps": ["Single company presentation disclosure, with no model, data, validation, or deployment-count detail.", "Do not attribute adjacent Haley outperformance to AI."], "source_urls_reviewed": ["https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-189", "Company": "KazMunayGas", "Company_primary": "KazMunayGas", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI predictive-maintenance rollout across Kazakhstan refineries", "Extracted_finding": "This is a named asset-health workflow with subsidiary-level equipment examples and a parent-company rollout plan. It is more concrete than a generic digital program, but outcome data remain limited to an identified anomaly rather than verified avoided downtime or savings.", "Stage_at_disclosure": "2025 pilots at Atyrau and Pavlodar; a full-scale rollout wave began in 2026; ANPZ subsidiary detail describes a pilot launched in October (year not stated; its article is dated December 2025)", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "KMG and its subsidiary are the reporting parties; no independent performance validation located. ANPZ article omits the year of the October pilot start; December 2025 publication does not justify a stronger date claim. ANPZ’s 28-unit initial circuit and KMG’s >100-unit phased plans are different scopes and should not be summed or conflated. No model accuracy, avoided downtime, savings or spend disclosed.", "Source_IDs": "RS125; RS126", "Source_date": "2026-05-22", "Source_type": "company announcement", "Source_locator": "Russian-language release; refinery predictive analytics program", "Source_URL": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-05-22; see chronology for operating dates.", "Context": "Detect early equipment anomalies from operating data so refinery teams can investigate before failures cause unplanned shutdowns.", "Implementation": "AI models analyze equipment/process parameters from sensor streams to identify abnormal behavior and alert staff. At ANPZ, the pilot covered equipment around catalytic cracking and CCR; the refinery reported a compressor anomaly was found in time to address it during scheduled work.", "Assessment": "Analyst inference: The initiative appears operationally relevant because it couples anomaly detection with refinery maintenance systems and has asset-level examples. Public evidence supports pilot activity and planned expansion, not quantified reliability economics.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Refining, LNG & process operations", "Refining, LNG & process operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-2-001-1", "Initiative_ID": "AI-189", "Title": "AI predictive-maintenance rollout across Kazakhstan refineries", "Date": "2025", "Finding": "KMG reports pilot projects at Atyrau and Pavlodar; its May 2026 release says an early-warning system for delayed-coking reactors was already functioning at Atyrau.", "Metric": "", "Source_IDs": "RS125", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-2-001-2", "Initiative_ID": "AI-189", "Title": "AI predictive-maintenance rollout across Kazakhstan refineries", "Date": "October (year omitted; article dated 2025-12-25)", "Finding": "ANPZ says a pilot with Honeywell launched in October and covered catalytic-cracking compressor and CCR equipment; it reports a compressor anomaly addressed during planned maintenance.", "Metric": "", "Source_IDs": "RS126", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-2-001-3", "Initiative_ID": "AI-189", "Title": "AI predictive-maintenance rollout across Kazakhstan refineries", "Date": "2026-04", "Finding": "KMG reports the initial predictive-analytics stage for critical compressors began at Shymkent refinery.", "Metric": "", "Source_IDs": "RS125", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-2-001-4", "Initiative_ID": "AI-189", "Title": "AI predictive-maintenance rollout across Kazakhstan refineries", "Date": "2026", "Finding": "KMG reports a full-scale rollout wave began at Atyrau and Pavlodar; >100 assets at each site remain a three-stage plan.", "Metric": "", "Source_IDs": "RS125", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-2-001-5", "Initiative_ID": "AI-189", "Title": "AI predictive-maintenance rollout across Kazakhstan refineries", "Date": "2026", "Finding": "ANPZ says 28 critical assets are listed for its initial monitoring circuit.", "Metric": "", "Source_IDs": "RS126", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-05-22", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-2-001"], "Research_search": " {\"key\": \"kmg-refinery-ai-predictive-maintenance\", \"existing_ids\": [], \"company\": \"KazMunayGas\", \"title\": \"AI predictive-maintenance rollout across Kazakhstan refineries\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"Refining, LNG & process operations\", \"Refining, LNG & process operations\"], \"category\": \"Internal adoption\", \"stage\": \"2025 pilots at Atyrau and Pavlodar; a full-scale rollout wave began in 2026; ANPZ subsidiary detail describes a pilot launched in October (year not stated; its article is dated December 2025)\", \"practical_relevance\": \"This is a named asset-health workflow with subsidiary-level equipment examples and a parent-company rollout plan. It is more concrete than a generic digital program, but outcome data remain limited to an identified anomaly rather than verified avoided downtime or savings.\", \"business_problem\": \"Detect early equipment anomalies from operating data so refinery teams can investigate before failures cause unplanned shutdowns.\", \"workflow\": \"AI models analyze equipment/process parameters from sensor streams to identify abnormal behavior and alert staff. At ANPZ, the pilot covered equipment around catalytic cracking and CCR; the refinery reported a compressor anomaly was found in time to address it during scheduled work.\", \"technology_data\": \"KMG describes AI analysis of high-volume operating data from thousands of installed sensors for early anomaly detection. ANPZ says its wider plant digital ecosystem includes Honeywell PHD, MES and equipment monitoring systems; the release does not explicitly state the predictive model was integrated with each platform.\", \"integration_human_role\": \"The release describes alerts and recommendations for personnel to check anomalies and take action; human maintenance response remains part of the workflow.\", \"deployment_scope\": \"KMG says pilots were implemented at Atyrau and Pavlodar in 2025. Its May 2026 release says a full-scale wave began at both sites, with >100 critical units per refinery planned across three stages; Shymkent began initial compressor coverage in April 2026; 10 critical units are connected at the polypropylene plant. A separate ANPZ December 2025 article says 28 critical assets were listed for its first monitoring circuit in 2026. These are different rollout scopes/stages, not completed fleetwide deployment.\", \"partners\": [\"Honeywell (ANPZ pilot)\"], \"disclosed_results\": [], \"economics\": \"No monetary savings, avoided downtime, predictive accuracy or failure-reduction result disclosed. KMG’s only concrete operating example is an anomaly detected before scheduled maintenance; do not convert that example into quantified avoided loss.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"KMG reports pilot projects at Atyrau and Pavlodar; its May 2026 release says an early-warning system for delayed-coking reactors was already functioning at Atyrau.\", \"source_ids\": [\"RS125\"]}, {\"date\": \"October (year omitted; article dated 2025-12-25)\", \"event\": \"ANPZ says a pilot with Honeywell launched in October and covered catalytic-cracking compressor and CCR equipment; it reports a compressor anomaly addressed during planned maintenance.\", \"source_ids\": [\"RS126\"]}, {\"date\": \"2026-04\", \"event\": \"KMG reports the initial predictive-analytics stage for critical compressors began at Shymkent refinery.\", \"source_ids\": [\"RS125\"]}, {\"date\": \"2026\", \"event\": \"KMG reports a full-scale rollout wave began at Atyrau and Pavlodar; >100 assets at each site remain a three-stage plan.\", \"source_ids\": [\"RS125\"]}, {\"date\": \"2026\", \"event\": \"ANPZ says 28 critical assets are listed for its initial monitoring circuit.\", \"source_ids\": [\"RS126\"]}], \"limitations\": [\"KMG and its subsidiary are the reporting parties; no independent performance validation located.\", \"ANPZ article omits the year of the October pilot start; December 2025 publication does not justify a stronger date claim.\", \"ANPZ’s 28-unit initial circuit and KMG’s >100-unit phased plans are different scopes and should not be summed or conflated.\", \"No model accuracy, avoided downtime, savings or spend disclosed.\"], \"assessment\": \"Analyst inference: The initiative appears operationally relevant because it couples anomaly detection with refinery maintenance systems and has asset-level examples. Public evidence supports pilot activity and planned expansion, not quantified reliability economics.\", \"sources\": [{\"id\": \"RS125\", \"title\": \"КМГ внедряет предиктивную аналитику на НПЗ\", \"url\": \"https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/\", \"publisher\": \"KazMunayGas\", \"date\": \"2026-05-22\", \"type\": \"company announcement\", \"locator\": \"Russian-language release; refinery predictive analytics program\", \"review_status\": \"full text reviewed\", \"supports\": [\"KazMunayGas reports that predictive-maintenance pilots ran in 2025 at Atyrau and Pavlodar refineries; its Atyrau delayed-coking reactor early-warning system was operating at disclosure.\", \"KMG says 2026 deployment waves target more than 100 critical equipment units each at Atyrau and Pavlodar; Shymkent began an initial compressor-monitoring stage in April 2026, and the polypropylene plant connected 10 equipment units.\", \"The release describes sensor data and AI-based anomaly detection to alert staff before failure; no accuracy, avoided-downtime or financial return is quantified.\"], \"original_id\": \"R2-001\"}, {\"id\": \"RS126\", \"title\": \"AI control of equipment at Atyrau refinery: predictive analytics in action\", \"url\": \"https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/\", \"publisher\": \"Atyrau Oil Refinery (ANPZ)\", \"date\": \"2025-12-25\", \"type\": \"company announcement\", \"locator\": \"AI predictive analytics pilot; connected to Honeywell PHD/MES\", \"review_status\": \"full text reviewed\", \"supports\": [\"ANPZ says a predictive-failure-analysis pilot began in October (year omitted; article is dated 2025-12-25) with Honeywell; it applies AI to operating parameters and flags anomalies for staff follow-up.\", \"The refinery names equipment around the catalytic-cracking and CCR units; it reports the system surfaced a compressor issue in time for planned maintenance and a separate prime-pump pressure anomaly.\", \"The release describes 28 critical assets as a planned 2026 monitoring scope; it does not report accuracy, avoided downtime or savings.\"], \"original_id\": \"R2-002\"}], \"field_sources\": {\"business_problem\": [\"RS125\", \"RS126\"], \"workflow\": [\"RS125\", \"RS126\"], \"technology_data\": [\"RS125\", \"RS126\"], \"integration_human_role\": [\"RS126\"], \"deployment_scope\": [\"RS125\", \"RS126\"], \"economics\": [\"RS125\", \"RS126\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"New KMG/ANPZ workflow supported by primary sources; year of ANPZ October pilot start is unstated.\", \"related_existing_ids\": [], \"cohort\": \"round2/additions-2\", \"packet_id\": \"round2-additions-2-001\", \"source_ids\": [\"RS125\", \"RS126\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-189\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-189", "company": "KazMunayGas", "key": "kmg-refinery-ai-predictive-maintenance", "outcome": "Independent review completed", "reviewer_outcome": "supported with stage caveats", "findings": [{"field": "stage/timeline", "issue": "The May 22, 2026 KMG release distinguishes 2025 pilots at Atyrau and Pavlodar from 2026 rollout waves. It says Shymkent began compressor monitoring in April 2026, Atyrau/Pavlodar each planned >100 critical units over three stages, and 10 units were connected at the polypropylene plant. The packet preserves these distinct scopes and does not claim full rollout complete.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "asset-level workflow / date", "issue": "The Atyrau refinery’s Dec. 25, 2025 release says its Honeywell pilot launched in October (in context, October 2025), reports compressor anomaly detection before scheduled maintenance, and cites a separate pump-inlet pressure anomaly. The packet describes the compressor case without asserting quantified avoided downtime.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["KMG and its subsidiary are the reporting parties; no independent performance validation located.", "ANPZ article omits the year of the October pilot start; December 2025 publication does not justify a stronger date claim.", "ANPZ’s 28-unit initial circuit and KMG’s >100-unit phased plans are different scopes and should not be summed or conflated.", "No model accuracy, avoided downtime, savings or spend disclosed."], "source_urls_reviewed": ["https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-190", "Company": "QatarEnergy LNG", "Company_primary": "QatarEnergy LNG", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "AI pilot for rotating-equipment fleet predictive maintenance", "Extracted_finding": "The disclosure identifies a practical reliability problem and proposed asset-wide scaling, but omits pilot location, vendor, deployment date and measured impact.", "Stage_at_disclosure": "A successful pilot is described in the 2025 strategy material; expansion across assets is a stated plan", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The document indicates a pilot and future expansion, not confirmed asset-wide deployment. Partner, location, AI technique, economics and operating metrics are not disclosed. The PDF is available and searchable in full; the relevant strategy passage is brief and does not disclose technical or outcome detail.", "Source_IDs": "RS127", "Source_date": "2025-07", "Source_type": "company magazine / strategy document", "Source_locator": "The Pioneer 167, PDF p. 5, Corporate Strategic Initiatives: AI Solutions for Machinery Fleet", "Source_URL": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-07; see chronology for operating dates.", "Context": "Use legacy equipment data to anticipate rotating-equipment maintenance needs and expand a successful AI pilot across the LNG asset fleet.", "Implementation": "QatarEnergy LNG describes AI solutions for machinery-fleet predictive use based on legacy rotating-equipment data. Detailed alert, work-order or maintenance-decision steps are not disclosed.", "Assessment": "Analyst inference: This is a credible but low-detail proof of a specific LNG machinery-fleet use case. It signals a test and intent to scale, not evidence of an operational fleet-wide system.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Refining, LNG & process operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-2-002-1", "Initiative_ID": "AI-190", "Title": "AI pilot for rotating-equipment fleet predictive maintenance", "Date": "2025-07", "Finding": "Company magazine lists pilot expansion across assets as a strategic initiative.", "Metric": "", "Source_IDs": "RS127", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-2-002"], "Research_search": " {\"key\": \"qatarenergy-lng-machinery-fleet-ai\", \"existing_ids\": [], \"company\": \"QatarEnergy LNG\", \"title\": \"AI pilot for rotating-equipment fleet predictive maintenance\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"Refining, LNG & process operations\"], \"category\": \"Internal adoption\", \"stage\": \"A successful pilot is described in the 2025 strategy material; expansion across assets is a stated plan\", \"practical_relevance\": \"The disclosure identifies a practical reliability problem and proposed asset-wide scaling, but omits pilot location, vendor, deployment date and measured impact.\", \"business_problem\": \"Use legacy equipment data to anticipate rotating-equipment maintenance needs and expand a successful AI pilot across the LNG asset fleet.\", \"workflow\": \"QatarEnergy LNG describes AI solutions for machinery-fleet predictive use based on legacy rotating-equipment data. Detailed alert, work-order or maintenance-decision steps are not disclosed.\", \"technology_data\": \"AI and historical/legacy rotating-equipment data are named. Model type, sensors, integrations and system architecture are not disclosed.\", \"integration_human_role\": \"Not disclosed beyond an intended predictive-maintenance use; whether maintenance personnel review alerts or systems generate work orders is not stated.\", \"deployment_scope\": \"The 2025 magazine calls this a successful pilot and proposes expansion across all assets; it does not name pilot asset(s), number of assets, or rollout milestones.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"No financial, reliability, avoided-failure or downtime metrics disclosed.\", \"timeline\": [{\"date\": \"2025-07\", \"event\": \"Company magazine lists pilot expansion across assets as a strategic initiative.\", \"source_ids\": [\"RS127\"]}], \"limitations\": [\"The document indicates a pilot and future expansion, not confirmed asset-wide deployment.\", \"Partner, location, AI technique, economics and operating metrics are not disclosed.\", \"The PDF is available and searchable in full; the relevant strategy passage is brief and does not disclose technical or outcome detail.\"], \"assessment\": \"Analyst inference: This is a credible but low-detail proof of a specific LNG machinery-fleet use case. It signals a test and intent to scale, not evidence of an operational fleet-wide system.\", \"sources\": [{\"id\": \"RS127\", \"title\": \"The Pioneer 167 — QatarEnergy LNG\", \"url\": \"https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf\", \"publisher\": \"QatarEnergy LNG\", \"date\": \"2025-07\", \"type\": \"company magazine / strategy document\", \"locator\": \"The Pioneer 167, PDF p. 5, Corporate Strategic Initiatives: AI Solutions for Machinery Fleet\", \"review_status\": \"full text reviewed\", \"supports\": [\"QatarEnergy LNG’s 2025 strategic-plan page says the company will expand a successful AI pilot across all assets.\", \"It identifies legacy data use for predictive purposes in rotating equipment.\", \"The passage provides no pilot location, vendor, asset count, rollout milestones or measured results.\"], \"original_id\": \"R2-003\"}], \"field_sources\": {\"business_problem\": [\"RS127\"], \"workflow\": [\"RS127\"], \"technology_data\": [\"RS127\"], \"integration_human_role\": [\"RS127\"], \"deployment_scope\": [\"RS127\"], \"economics\": [\"RS127\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"New workflow from the full 2025 QatarEnergy LNG magazine; the source discloses a successful pilot and planned expansion, but little implementation detail.\", \"related_existing_ids\": [], \"cohort\": \"round2/additions-2\", \"packet_id\": \"round2-additions-2-002\", \"source_ids\": [\"RS127\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-190\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-190", "company": "QatarEnergy LNG", "key": "qatarenergy-lng-machinery-fleet-ai", "outcome": "Independent review completed", "reviewer_outcome": "supported; review-status update needed", "findings": [{"field": "source review status", "issue": "The cited QatarEnergy LNG Issue 167 PDF opened as a complete 13-page primary-source document in this review. Page 5 explicitly says “expand the successful AI pilot across all assets” and “use of legacy data for predictive purposes in rotating equipment.” Packet limitations are accurate, but its statement “Only an excerpt of the magazine PDF was reviewed” is stale and should be updated to full-text reviewed.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "stage", "issue": "The primary source supports a completed/successful pilot and an expansion objective; it does not support completed asset-wide deployment, nor identify the pilot asset, dates, or results. Packet stage is correctly bounded.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The document indicates a pilot and future expansion, not confirmed asset-wide deployment.", "Partner, location, AI technique, economics and operating metrics are not disclosed.", "The PDF is available and searchable in full; the relevant strategy passage is brief and does not disclose technical or outcome detail."], "source_urls_reviewed": ["https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-191", "Company": "Ambyint", "Company_primary": "Ambyint", "Segment": "Upstream services & technology", "Theme": "AI products", "Initiative": "InfinityPL methane-mitigation upgrade for plunger-lift wells", "Extracted_finding": "A funded, domain-specific AI upgrade targets plunger-lift inefficiency and methane venting. Disclosure is unusually concrete about algorithms, partner, target workflow and prospective scale, but does not establish achieved emissions reductions.", "Stage_at_disclosure": "2025 ERA-funded development/upgrade; Tourmaline named as planned deployment partner; field deployment status of the upgraded capabilities not verified.", "Reported_metric": "ERA funding: $4.1 million (ERA-reported grant amount); Potential methane venting reduction: up to 90% (Company- and funder-reported prospective target)", "Interpretation_limit": "The up-to-90% figure is an expected reduction, not a verified result. No Tourmaline deployment date, number of wells, baseline, measurement protocol, or post-deployment result is disclosed. Technical and safety-control details are not disclosed. Sources do not establish autonomous setpoint management or describe operator approval/override controls for the upgrade.", "Source_IDs": "RS128; RS129", "Source_date": "2025-05-09", "Source_type": "company announcement", "Source_locator": "Full page", "Source_URL": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-05-09; see chronology for operating dates.", "Context": "Plunger-lift wells can vent methane when operation is inefficient; continuous manual monitoring and intervention limit coverage.", "Implementation": "Upgrade Ambyint’s cloud InfinityPL platform to predict inefficiencies, detect complex anomalies and reduce continuous manual intervention. Tourmaline plans to deploy the upgraded technology across a substantial number of Alberta plunger-lift sites.", "Assessment": "Analyst inference: A specific practical AI workstream with third-party funding and a named prospective operator deployment. Evidence supports development and intent, not live rollout or quantified impact.", "Primary_department": "Production & field operations", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-3-001-1", "Initiative_ID": "AI-191", "Title": "InfinityPL methane-mitigation upgrade for plunger-lift wells", "Date": "2025-04-16", "Finding": "ERA announced project support and Tourmaline’s planned deployment partnership.", "Metric": "", "Source_IDs": "RS129", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-3-001-2", "Initiative_ID": "AI-191", "Title": "InfinityPL methane-mitigation upgrade for plunger-lift wells", "Date": "2025-05-09", "Finding": "Ambyint described the funded InfinityPL upgrade.", "Metric": "", "Source_IDs": "RS128", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-05-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-3-001"], "Research_search": " {\"key\": \"ambyint-infinitypl-methane-upgrade\", \"existing_ids\": [], \"company\": \"Ambyint\", \"title\": \"InfinityPL methane-mitigation upgrade for plunger-lift wells\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"HSE & emissions\"], \"category\": \"Commercial AI product\", \"stage\": \"2025 ERA-funded development/upgrade; Tourmaline named as planned deployment partner; field deployment status of the upgraded capabilities not verified.\", \"practical_relevance\": \"A funded, domain-specific AI upgrade targets plunger-lift inefficiency and methane venting. Disclosure is unusually concrete about algorithms, partner, target workflow and prospective scale, but does not establish achieved emissions reductions.\", \"business_problem\": \"Plunger-lift wells can vent methane when operation is inefficient; continuous manual monitoring and intervention limit coverage.\", \"workflow\": \"Upgrade Ambyint’s cloud InfinityPL platform to predict inefficiencies, detect complex anomalies and reduce continuous manual intervention. Tourmaline plans to deploy the upgraded technology across a substantial number of Alberta plunger-lift sites.\", \"technology_data\": \"ERA describes advanced machine-learning and reinforcement-learning algorithms added to the existing AI-driven cloud platform. Specific input data, model validation and control guardrails are not disclosed.\", \"integration_human_role\": \"The project aims to reduce continuous manual intervention. Sources do not specify whether the upgraded system changes setpoints autonomously or describe operator approval, override, or safety controls.\", \"deployment_scope\": \"Existing platform reported deployed by Ambyint at 17 North American producers; new ERA-funded project identifies Tourmaline plunger-lift well sites in Alberta as planned deployment area. Existing footprint is not equivalent to deployment of upgraded methane functionality.\", \"partners\": [\"Emissions Reduction Alberta\", \"Tourmaline Oil\"], \"disclosed_results\": [{\"metric\": \"ERA funding\", \"value\": \"$4.1 million\", \"period\": \"2025 award\", \"attribution\": \"ERA-reported grant amount\", \"source_ids\": [\"RS128\", \"RS129\"]}, {\"metric\": \"Potential methane venting reduction\", \"value\": \"up to 90%\", \"period\": \"project target/estimate, not measured outcome\", \"attribution\": \"Company- and funder-reported prospective target\", \"source_ids\": [\"RS128\", \"RS129\"]}], \"economics\": \"Funding is for technology development; no customer savings or realized methane reduction for the upgraded system is reported.\", \"timeline\": [{\"date\": \"2025-04-16\", \"event\": \"ERA announced project support and Tourmaline’s planned deployment partnership.\", \"source_ids\": [\"RS129\"]}, {\"date\": \"2025-05-09\", \"event\": \"Ambyint described the funded InfinityPL upgrade.\", \"source_ids\": [\"RS128\"]}], \"limitations\": [\"The up-to-90% figure is an expected reduction, not a verified result.\", \"No Tourmaline deployment date, number of wells, baseline, measurement protocol, or post-deployment result is disclosed.\", \"Technical and safety-control details are not disclosed.\", \"Sources do not establish autonomous setpoint management or describe operator approval/override controls for the upgrade.\"], \"assessment\": \"Analyst inference: A specific practical AI workstream with third-party funding and a named prospective operator deployment. Evidence supports development and intent, not live rollout or quantified impact.\", \"sources\": [{\"id\": \"RS128\", \"title\": \"Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction\", \"url\": \"https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/\", \"publisher\": \"Ambyint\", \"date\": \"2025-05-09\", \"type\": \"company announcement\", \"locator\": \"Full page\", \"review_status\": \"full text reviewed\", \"supports\": [\"Ambyint says $4.1m from ERA supports a next-generation InfinityPL platform upgrade.\", \"Company says the existing platform is deployed by 17 producers, and the project is expected to reduce methane venting by up to 90%.\", \"Tourmaline is named as key planned deployment partner at plunger-lift sites in Alberta.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3A001\"}, {\"id\": \"RS129\", \"title\": \"Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector\", \"url\": \"https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/\", \"publisher\": \"Emissions Reduction Alberta / Government of Alberta\", \"date\": \"2025-04-16\", \"type\": \"public funding announcement\", \"locator\": \"Ambyint project description\", \"review_status\": \"full text reviewed\", \"supports\": [\"ERA confirms $4.1m funding for Ambyint’s upgrade to the InfinityPL AI platform using advanced ML and reinforcement learning.\", \"ERA describes prediction of inefficiencies, complex-anomaly detection and reduced manual intervention; up to-90% methane-venting reduction remains expected.\", \"Tourmaline plans deployment across a substantial number of Alberta plunger-lift well sites.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3A002\"}], \"field_sources\": {\"business_problem\": [\"RS128\", \"RS129\"], \"workflow\": [\"RS128\", \"RS129\"], \"technology_data\": [\"RS129\"], \"integration_human_role\": [\"RS128\", \"RS129\"], \"deployment_scope\": [\"RS128\", \"RS129\"], \"economics\": [\"RS128\", \"RS129\"]}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"No exact profile match found in the round-two baseline; addition is a distinct 2025 funded upgrade, not a claim that Ambyint’s established optimization platform was first launched in 2025. Public sources state $4.1 million and do not establish autonomous setpoint management.\", \"cohort\": \"round2/additions-3\", \"packet_id\": \"round2-additions-3-001\", \"source_ids\": [\"RS128\", \"RS129\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-191\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-191", "company": "Ambyint", "key": "ambyint-infinitypl-methane-upgrade", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; correct two overstatements", "findings": [{"field": "funding metric", "issue": "The two linked sources state $4.1 million, not $4.08 million. The more precise $4.08m value in disclosed_results is unsupported by this source pair; either use $4.1m or supply an official source that establishes $4.08m.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "workflow / human role", "issue": "Both linked sources support an existing InfinityPL platform upgrade to predict inefficiencies, detect complex anomalies and reduce continuous manual intervention. Neither linked source states that the new upgrade will autonomously manage well setpoints or describes operator approval/override controls. Remove or mark setpoint control as undisclosed unless a further source is cited.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "deployment stage / results", "issue": "The sources consistently describe a funded upgrade and Tourmaline as a planned deployment partner. Up-to-90% methane-venting reduction is prospective, not an achieved outcome; the packet handles this distinction correctly. The existing 17-producer footprint refers to the existing platform, not the upgraded methane feature.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The up-to-90% figure is an expected reduction, not a verified result.", "No Tourmaline deployment date, number of wells, baseline, measurement protocol, or post-deployment result is disclosed.", "Technical and safety-control details are not disclosed.", "Sources do not establish autonomous setpoint management or describe operator approval/override controls for the upgrade."], "source_urls_reviewed": ["https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-192", "Company": "Kpler", "Company_primary": "Kpler", "Segment": "Trading & shipping", "Theme": "AI products", "Initiative": "AI email parsing and vessel recommendation in chartering workflow", "Extracted_finding": "Targets a specific operational bottleneck in maritime energy trading: turning high-volume unstructured broker/cargo communications into searchable lists and matching vessel candidates to commercial and risk constraints.", "Stage_at_disclosure": "AI email-parsing and chartering-agent workflow presented in October 2025; source said AI chartering agent was planned to launch in Q4 2025. Current product materials describe AI email parsing, but dated production launch of the agent was not independently confirmed.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "The Q4 2025 agent was announced as planned; launch/commercial availability is not confirmed by dated primary evidence reviewed here. No evidence that recommendations execute trades or replace charterer approval. No email-parsing accuracy figure appears in the reviewed current product page; no independent accuracy or productivity metric was identified.", "Source_IDs": "RS130; RS131", "Source_date": "2025-10-16", "Source_type": "company announcement", "Source_locator": "Harness AI in Chartering", "Source_URL": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-10-16; see chronology for operating dates.", "Context": "Charterers and brokers manually process email, PDFs and spreadsheets containing cargo and vessel positions; delays and scattered records impede fixture decisions.", "Implementation": "AI parsing converts unstructured emails, PDFs and spreadsheets into structured cargo/position lists. A planned chartering agent recommends vessel candidates using freight economics, risk and user-defined filters. Present product materials also describe AI extraction of vessel names, IMO numbers, cargo details and laycan dates into tonnage/cargo/fixture lists.", "Assessment": "Analyst inference: A useful AI-enabled commercial workflow with visible data handoffs and a human decision-maker. Verification is strongest for announced features; rollout, accuracy and realized economics remain open.", "Primary_department": "Trading & marketing", "Secondary_departments": ["Supply chain & logistics"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-3-002-1", "Initiative_ID": "AI-192", "Title": "AI email parsing and vessel recommendation in chartering workflow", "Date": "2025-10-16", "Finding": "Kpler event described email parsing and previewed a chartering agent planned for Q4 launch.", "Metric": "", "Source_IDs": "RS130", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "round2-additions-3-002-2", "Initiative_ID": "AI-192", "Title": "AI email parsing and vessel recommendation in chartering workflow", "Date": null, "Finding": "Current product page describes natural-language analysis and AI extraction in chartering; page publication date not stated.", "Metric": "", "Source_IDs": "RS131", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2025-10-16", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-3-002"], "Research_search": " {\"key\": \"kpler-ai-chartering-email-extraction\", \"existing_ids\": [], \"company\": \"Kpler\", \"title\": \"AI email parsing and vessel recommendation in chartering workflow\", \"department\": \"Trading & marketing\", \"secondary_departments\": [\"Supply chain & logistics\"], \"category\": \"Commercial AI product\", \"stage\": \"AI email-parsing and chartering-agent workflow presented in October 2025; source said AI chartering agent was planned to launch in Q4 2025. Current product materials describe AI email parsing, but dated production launch of the agent was not independently confirmed.\", \"practical_relevance\": \"Targets a specific operational bottleneck in maritime energy trading: turning high-volume unstructured broker/cargo communications into searchable lists and matching vessel candidates to commercial and risk constraints.\", \"business_problem\": \"Charterers and brokers manually process email, PDFs and spreadsheets containing cargo and vessel positions; delays and scattered records impede fixture decisions.\", \"workflow\": \"AI parsing converts unstructured emails, PDFs and spreadsheets into structured cargo/position lists. A planned chartering agent recommends vessel candidates using freight economics, risk and user-defined filters. Present product materials also describe AI extraction of vessel names, IMO numbers, cargo details and laycan dates into tonnage/cargo/fixture lists.\", \"technology_data\": \"Kpler describes AI-powered parsing across email/PDF/spreadsheet inputs and market, freight and risk data; underlying model, precision validation dataset and operating controls are undisclosed. Current product page cites 10+ years of proprietary historical data and real-time market inputs, but its undated statements are not treated as verified 2025 launch evidence.\", \"integration_human_role\": \"Outputs support chartering professionals’ analysis and selection; the event description does not say the agent autonomously fixes vessels or executes trades.\", \"deployment_scope\": \"Commercial product directed to shipowners, charterers and brokers across commodity markets, including oil, refined products and LNG; number of customers/users of AI features not disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Kpler’s current product page promotes faster analysis but no attributable customer time-saved, cost or trading-return evidence was identified.\", \"timeline\": [{\"date\": \"2025-10-16\", \"event\": \"Kpler event described email parsing and previewed a chartering agent planned for Q4 launch.\", \"source_ids\": [\"RS130\"]}, {\"date\": null, \"event\": \"Current product page describes natural-language analysis and AI extraction in chartering; page publication date not stated.\", \"source_ids\": [\"RS131\"]}], \"limitations\": [\"The Q4 2025 agent was announced as planned; launch/commercial availability is not confirmed by dated primary evidence reviewed here.\", \"No evidence that recommendations execute trades or replace charterer approval.\", \"No email-parsing accuracy figure appears in the reviewed current product page; no independent accuracy or productivity metric was identified.\"], \"assessment\": \"Analyst inference: A useful AI-enabled commercial workflow with visible data handoffs and a human decision-maker. Verification is strongest for announced features; rollout, accuracy and realized economics remain open.\", \"sources\": [{\"id\": \"RS130\", \"title\": \"Navigating Risk & Harnessing AI Briefing\", \"url\": \"https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai\", \"publisher\": \"Kpler\", \"date\": \"2025-10-16\", \"type\": \"company announcement\", \"locator\": \"Harness AI in Chartering\", \"review_status\": \"full text reviewed\", \"supports\": [\"Kpler described AI-powered parsing of emails, PDFs and spreadsheets into structured cargo and position lists.\", \"Kpler previewed an AI chartering agent to recommend vessel candidates based on freight economics, risk and custom filters, and said it was launching Q4.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3K001\"}, {\"id\": \"RS131\", \"title\": \"Kpler AI\", \"url\": \"https://www.kpler.com/solutions/ai\", \"publisher\": \"Kpler\", \"date\": null, \"type\": \"product page\", \"locator\": \"Kpler Copilot / Kpler agents / FAQ\", \"review_status\": \"full text reviewed\", \"supports\": [\"Current Kpler page describes natural-language questions, AI-assisted market analysis, and chartering/cargo/compliance agent categories.\", \"Page describes sources including proprietary history, AIS, satellite tracking, customs/trade flows and refinery information.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3K002\"}], \"field_sources\": {\"business_problem\": [\"RS130\"], \"workflow\": [\"RS130\", \"RS131\"], \"technology_data\": [\"RS130\", \"RS131\"], \"integration_human_role\": [\"RS130\"], \"deployment_scope\": [\"RS130\", \"RS131\"], \"economics\": []}, \"verification\": \"verified primary\", \"corrections_to_existing\": \"No exact Kpler profile match found in the round-two baseline. Removed unsupported 99% accuracy claim; the dated event source supports the announced workflow and planned Q4 2025 launch, while current undated product material does not establish a numeric parsing-accuracy result.\", \"cohort\": \"round2/additions-3\", \"packet_id\": \"round2-additions-3-002\", \"source_ids\": [\"RS130\", \"RS131\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-192\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-192", "company": "Kpler", "key": "kpler-ai-chartering-email-extraction", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; remove unsupported current-page metric", "findings": [{"field": "disclosed_results", "issue": "The packet’s “99% email parsing accuracy” is not supported by the linked current Kpler AI page. Directly searched that page for “accuracy” and “99”; no accuracy metric appeared. Remove the 99% result unless an observed source supports it.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "stage / dates", "issue": "Kpler’s Oct. 16, 2025 event page supports email/PDF/spreadsheet parsing and describes the AI chartering agent as “launching Q4.” The packet correctly states that a dated launch or commercial availability was not independently confirmed; preserve Q4 as planned, not achieved.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "source scope", "issue": "The current solutions page is undated and supports general conversational Kpler AI/data foundations, not email parsing accuracy or the dated launch status of the agent. The packet correctly treats its date as unknown, but its support array should not use it to support the 99% result.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The Q4 2025 agent was announced as planned; launch/commercial availability is not confirmed by dated primary evidence reviewed here.", "No evidence that recommendations execute trades or replace charterer approval.", "No email-parsing accuracy figure appears in the reviewed current product page; no independent accuracy or productivity metric was identified."], "source_urls_reviewed": ["https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "https://www.kpler.com/solutions/ai"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-193", "Company": "Kodiak Gas Services", "Company_primary": "Kodiak Gas Services", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "Material Replenisher AI agent for compression-field technicians", "Extracted_finding": "A deployed conversational agent addresses the parts-search and ordering bottleneck for Kodiak field technicians. IFS reports annual ROI and technician hours returned after rollout; methodology and independent validation are not disclosed.", "Stage_at_disclosure": "Rolled out by 2026-04-23, according to IFS; rollout date and usage scale not specified.", "Reported_metric": "IFS's 27 November 2025 release quoted Kodiak's CIO on a conditional scenario: $3 million annual ROI and 90,000 hours returned if half the workforce engaged with the agent daily. In its 23 April 2026 release, IFS later described the same figures as the result after Kodiak rolled out the agent, without disclosing a measurement period or method. The latter remains vendor-reported and independently unverified.", "Interpretation_limit": "The April 2026 result is vendor-reported; no measurement method, baseline, timeframe or independent customer validation is provided. The 2025 release gives the same numbers as a conditional scenario and does not date the rollout. Warehouse count, nearby-stock search, data-system integration details, adoption rates and control boundaries are not disclosed in the linked sources.", "Source_IDs": "RS132; RS133", "Source_date": "2026-04-23", "Source_type": "Vendor press release", "Source_locator": "Paragraph describing Kodiak Material Replenisher rollout and vendor-reported annual ROI/hours; prior conditional scenario remains in November 2025 disclosure.", "Source_URL": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-04-23; see chronology for operating dates.", "Context": "Kodiak says its 800 technicians support 4.5 million horsepower of compression capacity and spent time searching for parts; the reviewed sources do not specify a 150+ warehouse network.", "Implementation": "Technicians find and order parts conversationally through the Material Replenisher agent. The sources do not specify nearby-stock search, exact warehouse-location inputs, or all fulfillment steps.", "Assessment": "Analyst inference: IFS reports a live customer workflow for conversational parts finding and ordering. Its 2026 release presents ROI and time returned as post-rollout results, but absent methodology these remain vendor-attributed rather than independently verified outcomes.", "Primary_department": "Supply chain & logistics", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-3-003-1", "Initiative_ID": "AI-193", "Title": "Material Replenisher AI agent for compression-field technicians", "Date": "2025-11-27", "Finding": "IFS published a conditional half-workforce daily-use scenario for $3m annual ROI and 90,000 hours; this release does not establish rollout by this date.", "Metric": "", "Source_IDs": "RS132", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "round2-additions-3-003-2", "Initiative_ID": "AI-193", "Title": "Material Replenisher AI agent for compression-field technicians", "Date": "2026-04-23", "Finding": "IFS said Kodiak had rolled out Material Replenisher and described $3m annual ROI / 90,000 hours returned as the result.", "Metric": "", "Source_IDs": "RS133", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2026-04-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-3-003"], "Research_search": " {\"key\": \"ifs-kodiak-material-replenisher\", \"existing_ids\": [], \"company\": \"Kodiak Gas Services\", \"title\": \"Material Replenisher AI agent for compression-field technicians\", \"department\": \"Supply chain & logistics\", \"secondary_departments\": [\"Maintenance & reliability\"], \"category\": \"Internal adoption\", \"stage\": \"Rolled out by 2026-04-23, according to IFS; rollout date and usage scale not specified.\", \"practical_relevance\": \"A deployed conversational agent addresses the parts-search and ordering bottleneck for Kodiak field technicians. IFS reports annual ROI and technician hours returned after rollout; methodology and independent validation are not disclosed.\", \"business_problem\": \"Kodiak says its 800 technicians support 4.5 million horsepower of compression capacity and spent time searching for parts; the reviewed sources do not specify a 150+ warehouse network.\", \"workflow\": \"Technicians find and order parts conversationally through the Material Replenisher agent. The sources do not specify nearby-stock search, exact warehouse-location inputs, or all fulfillment steps.\", \"technology_data\": \"IFS identifies Material Replenisher as an IFS Loops agent and Kodiak as an IFS Cloud customer. The reviewed sources do not specify the exact inventory or warehouse data integrations, model, or retrieval architecture.\", \"integration_human_role\": \"Technicians initiate conversational parts requests; approval thresholds, exception handling, fulfillment confirmation and other controls are not described.\", \"deployment_scope\": \"IFS reports rollout at Kodiak; it describes a 4.5-million-horsepower fleet and 800 technicians. Warehouse count, sites covered, user adoption and operational rollout date are not disclosed.\", \"partners\": [\"IFS\"], \"disclosed_results\": [{\"metric\": \"Potential annual ROI and technician hours returned\", \"value\": \"$3 million and 90,000 hours\", \"period\": \"2025 conditional scenario: if half of workforce engages daily for one year\", \"attribution\": \"IFS-reported customer estimate; conditional and not a measured result\", \"source_ids\": [\"RS132\"]}, {\"metric\": \"Annual ROI and technician hours returned\", \"value\": \"$3 million and 90,000 hours\", \"period\": \"IFS described these as the result following Kodiak rollout in a release published 2026-04-23; methodology/timeframe not provided\", \"attribution\": \"Vendor-reported result; not independently validated\", \"source_ids\": [\"RS133\"]}], \"economics\": \"The November 2025 IFS release framed $3m annual ROI and 90,000 hours as a conditional scenario if half the technicians used the agent daily. In April 2026, IFS said Kodiak had rolled out the agent and called the same figures the result. This later claim is vendor-reported and lacks measurement methodology, timeframe, and independent validation.\", \"timeline\": [{\"date\": \"2025-11-27\", \"event\": \"IFS published a conditional half-workforce daily-use scenario for $3m annual ROI and 90,000 hours; this release does not establish rollout by this date.\", \"source_ids\": [\"RS132\"]}, {\"date\": \"2026-04-23\", \"event\": \"IFS said Kodiak had rolled out Material Replenisher and described $3m annual ROI / 90,000 hours returned as the result.\", \"source_ids\": [\"RS133\"]}], \"limitations\": [\"The April 2026 result is vendor-reported; no measurement method, baseline, timeframe or independent customer validation is provided.\", \"The 2025 release gives the same numbers as a conditional scenario and does not date the rollout.\", \"Warehouse count, nearby-stock search, data-system integration details, adoption rates and control boundaries are not disclosed in the linked sources.\"], \"assessment\": \"Analyst inference: IFS reports a live customer workflow for conversational parts finding and ordering. Its 2026 release presents ROI and time returned as post-rollout results, but absent methodology these remain vendor-attributed rather than independently verified outcomes.\", \"sources\": [{\"id\": \"RS132\", \"title\": \"IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI\", \"url\": \"https://www.ifs.com/en/insights/news/ifs-cloud-25r2\", \"publisher\": \"IFS\", \"date\": \"2025-11-27\", \"type\": \"company announcement\", \"locator\": \"Kodiak customer disclosure / CIO quote\", \"review_status\": \"full text reviewed\", \"supports\": [\"IFS reported Kodiak rolled out a Material Replenisher agent for technicians to find and order parts conversationally.\", \"CIO’s $3m and 90,000-hour figures are conditional on half the workforce engaging once a day; the 2025 release does not establish rollout date.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3IFS001\"}, {\"id\": \"RS133\", \"title\": \"IFS Loops launches Agent Studio\", \"url\": \"https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio\", \"publisher\": \"IFS\", \"date\": \"2026-04-23\", \"type\": \"company announcement\", \"locator\": \"Kodiak paragraph\", \"review_status\": \"full text reviewed\", \"supports\": [\"IFS says Kodiak rolled out Material Replenisher; technicians find and order parts through conversation.\", \"IFS calls $3m annual ROI and 90,000 hours returned “the result” after describing rollout; this vendor claim lacks disclosed measurement methodology or independent validation.\"], \"accessed\": \"2026-09-27\", \"original_id\": \"P3IFS002\"}], \"field_sources\": {\"business_problem\": [\"RS132\", \"RS133\"], \"workflow\": [\"RS132\", \"RS133\"], \"technology_data\": [\"RS132\", \"RS133\"], \"integration_human_role\": [\"RS132\"], \"deployment_scope\": [\"RS132\", \"RS133\"], \"economics\": [\"RS132\", \"RS133\"]}, \"verification\": \"verified partner\", \"corrections_to_existing\": \"No exact Kodiak Material Replenisher record was found in the baseline. The 2025 release gives a conditional estimate; the 2026 release reports rollout and calls the figures the result. Scope and realized-value details remain vendor-reported; removed unsupported warehouse-count and system-integration specifics.\", \"cohort\": \"round2/additions-3\", \"packet_id\": \"round2-additions-3-003\", \"source_ids\": [\"RS132\", \"RS133\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-193\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-193", "company": "Kodiak Gas Services", "key": "ifs-kodiak-material-replenisher", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; deployment date and results attribution need correction", "findings": [{"field": "deployment date", "issue": "The Nov. 27, 2025 IFS Cloud release gives a conditional scenario (“If half of our workforce engages…”) but does not state that Kodiak had rolled out Material Replenisher by that date. The Apr. 23, 2026 IFS Loops release says Kodiak “rolled out” the agent. Change the deployment timeline to “by 2026-04-23 / date not specified”; keep Nov. 2025 only for the conditional estimate.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "disclosed_results", "issue": "The 2025 release attributes $3m/year and 90,000 hours to a conditional assumption of half the workforce using the agent daily. The 2026 release changes attribution language and calls $3m annual ROI and 90,000 hours returned “the result” after saying Kodiak rolled out the agent. The packet currently says the 2026 release merely repeats potential estimates and does not treat them as realized; that understates the later vendor claim. Separate the 2025 conditional estimate from the 2026 IFS-reported realized result, while making clear the latter remains vendor-reported and lacks methodology or independent validation.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "deployment scope / workflow", "issue": "The 2026 source confirms 4.5 million horsepower, 800 technicians, prior time spent searching for parts, and conversational find-and-order functionality. The linked two-source set does not substantiate the packet’s “150+ warehouses,” nearby-stock search, or specific IFS Cloud inventory/warehouse-location data integration. These may be true but require a cited source; otherwise remove or mark undisclosed.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The April 2026 result is vendor-reported; no measurement method, baseline, timeframe or independent customer validation is provided.", "The 2025 release gives the same numbers as a conditional scenario and does not date the rollout.", "Warehouse count, nearby-stock search, data-system integration details, adoption rates and control boundaries are not disclosed in the linked sources."], "source_urls_reviewed": ["https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-194", "Company": "Expand Energy", "Company_primary": "Expand Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Leucipa production optimization and Lucy assistant pilot", "Extracted_finding": "Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.", "Stage_at_disclosure": "Deployment beginning at Operations Support Center; Lucy pilot planned", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "A deployment start and assistant pilot are different stages. No quantified production uplift or cost saving is established. Do not equate the broader vendor platform footprint with Expand’s actual deployment.", "Source_IDs": "RS134", "Source_date": "2026-03-03", "Source_type": "company announcement", "Source_locator": "Leucipa deployment and Snowflake data-platform sections", "Source_URL": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-03-03; see chronology for operating dates.", "Context": "Production teams need usable operating insights and coordinated field decisions.", "Implementation": "Expand describes Leucipa for production management, optimization and forecasting, plus a Lucy generative-AI chat pilot for real-time production insight.", "Assessment": "Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-root-001-1", "Initiative_ID": "AI-194", "Title": "Leucipa production optimization and Lucy assistant pilot", "Date": "2026-03-03", "Finding": "Operator reports initial deployment and planned team expansion.", "Metric": "", "Source_IDs": "RS134", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-03-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-root-001"], "Research_search": " {\"key\": \"expand-leucipa-lucy\", \"existing_ids\": [], \"company\": \"Expand Energy\", \"title\": \"Leucipa production optimization and Lucy assistant pilot\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Deployment beginning at Operations Support Center; Lucy pilot planned\", \"practical_relevance\": \"Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.\", \"business_problem\": \"Production teams need usable operating insights and coordinated field decisions.\", \"workflow\": \"Expand describes Leucipa for production management, optimization and forecasting, plus a Lucy generative-AI chat pilot for real-time production insight.\", \"technology_data\": \"Leucipa and a conversational generative-AI interface; model architecture and validation metrics not disclosed.\", \"integration_human_role\": \"Initial use is at the Operations Support Center. Decision authority and control-system write permissions are not disclosed.\", \"deployment_scope\": \"Field Operations and Production Engineering expansion is planned; completed rollout is not established.\", \"partners\": [\"Baker Hughes\"], \"disclosed_results\": [], \"economics\": \"Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.\", \"timeline\": [{\"date\": \"2026-03-03\", \"event\": \"Operator reports initial deployment and planned team expansion.\", \"source_ids\": [\"RS134\"]}], \"limitations\": [\"A deployment start and assistant pilot are different stages.\", \"No quantified production uplift or cost saving is established.\", \"Do not equate the broader vendor platform footprint with Expand’s actual deployment.\"], \"assessment\": \"Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.\", \"sources\": [{\"id\": \"RS134\", \"title\": \"EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency\", \"url\": \"https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/\", \"publisher\": \"Expand Energy\", \"date\": \"2026-03-03\", \"type\": \"company announcement\", \"locator\": \"Leucipa deployment and Snowflake data-platform sections\", \"review_status\": \"full text reviewed\", \"supports\": [\"Operations Support Center deployment; Lucy pilot; Snowflake drilling ML and SCADA alerting.\"], \"original_id\": \"root2-exe\"}], \"verification\": \"verified primary\", \"corrections_to_existing\": \"New second-pass research entry; no earlier initiative ID duplicated.\", \"field_sources\": {\"business_problem\": [\"RS134\"], \"workflow\": [\"RS134\"], \"technology_data\": [\"RS134\"], \"integration_human_role\": [\"RS134\"], \"deployment_scope\": [\"RS134\"], \"economics\": [\"RS134\"]}, \"cohort\": \"round2/additions-root\", \"packet_id\": \"round2-additions-root-001\", \"source_ids\": [\"RS134\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-194\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-194", "company": "Expand Energy", "key": "expand-leucipa-lucy", "outcome": "Independent review completed", "reviewer_outcome": "supported with stage limitation", "findings": [{"field": "Independent review", "issue": "Leucipa is described as AI-powered and deployment begins through the Operations Support Center. Lucy is explicitly a pilot (“will also pilot”), so do not write that Lucy is already deployed.", "correction": "", "source_url": ""}], "residual_gaps": ["A deployment start and assistant pilot are different stages.", "No quantified production uplift or cost saving is established.", "Do not equate the broader vendor platform footprint with Expand’s actual deployment."], "source_urls_reviewed": ["https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-195", "Company": "Expand Energy", "Company_primary": "Expand Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Snowflake-hosted drilling ML and operational alerts", "Extracted_finding": "Analyst inference: This is a distinct drilling and alerting workflow alongside production optimization. Determine how alerts are validated and what happens when model confidence is low.", "Stage_at_disclosure": "Operating. Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Data integration benefits cannot be attributed entirely to AI. No independently measured drilling or failure-rate improvement is provided. Current platform/ML use and future dashboard/MLOps work have different stages; platform efficiency does not establish AI-only financial benefit. Same source and substantive claims already in AI-195. No additional adoption or numerical-result count. Site data coverage is not proof of automatic control at all 3,700 locations. Keep this platform separate from Leucipa in AI-194.", "Source_IDs": "RS134; SW7-013; OCT5-S041", "Source_date": "2026-03-03", "Source_type": "company announcement", "Source_locator": "Leucipa deployment and Snowflake data-platform sections", "Source_URL": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-03-03; see chronology for operating dates.", "Context": "Centralize engineering and operating data so forecasts, drilling models and alerts can be used in everyday work.", "Implementation": "Snowpark ML supports production/type-curve forecasting. Drilling ML and operational alerts use centralized data; Cortex Analyst supports natural-language queries. The case describes site data for 3,700 locations. A full drilling dashboard is under development and broader MLOps maturity is on the roadmap.", "Assessment": "The detailed Snowflake case reaches beyond drilling alerts into forecasting and integration. Its estimated labor benefit and M&A platform gains are recorded in the existing dossier, with no AI-only cash ROI established.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW7-D-AI-195", "Initiative_ID": "AI-195", "Title": "Snowflake forecasting ML, drilling analytics and site alerts", "Date": "Undated supplemental case", "Finding": "Snowpark supports ML-based type-curve forecasts; Snowflake hosts drilling ML and Cortex Analyst questions. SCADA/site data produces alerts for operations-center investigation.", "Metric": "Estimated 1,000+ labor hours saved annually in one business unit; data monitoring covers 3,700 sites. Neither proves production uplift. Separately, 40% less data-related M&A time is a platform result.", "Source_IDs": "SW7-013", "Evidence": "Full official Snowflake customer case read; customer executive quotes present. Undated, vendor-authored, estimates/claims not independently verified.", "Origin": "Seventh research swarm"}, {"ID": "OCT5-AI-195-1", "Initiative_ID": "AI-195", "Title": "Snowflake architecture: operating drilling ML versus dashboard development", "Date": "Undated; reviewed 2026-10-06", "Finding": "The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.", "Metric": "", "Source_IDs": "OCT5-S041", "Source_URL": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Latest_disclosure": "Undated; reviewed 2026-10-06", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-three-rounds-ai-195", "swarm7-ai-195", "round2-additions-root-002"], "Research_search": "{\"packet_id\": \"oct6-three-rounds-ai-195\", \"review_date\": \"2026-10-06\", \"title\": \"Snowflake architecture: operating drilling ML versus dashboard development\", \"workflow\": \"The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.\", \"technology_data\": \"The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.\", \"deployment_scope\": \"Operating drilling ML and SCADA alerts; a full-scale Cortex drilling dashboard under development and MLOps platform roadmap. Undated vendor page cannot establish launch dates.\", \"result_label\": \"Supplemental source: scope and limits\", \"attribution\": \"Vendor-published; not independently evaluated\", \"economics\": \"No project cost and measured cash-benefit pair disclosed.\", \"limitations\": [\"Current platform/ML use and future dashboard/MLOps work have different stages; platform efficiency does not establish AI-only financial benefit.\"], \"verification\": \"verified partner\", \"source_review\": \"Full original source reviewed in Luna verification and editor checks on 6 October 2026.\", \"source_locator\": \"Digging deeper into operational efficiency; Energizing the future\", \"timeline_date\": \"Undated; reviewed 2026-10-06\", \"timeline_event\": \"Snowflake architecture: operating drilling ML versus dashboard development\", \"stage\": \"Operating. Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.\", \"key\": \"oct6-three-rounds-ai-195\", \"record_ids\": [\"AI-195\"], \"company\": \"Expand Energy\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"source_ids\": [\"SW7-013\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"SW7-013\"], \"workflow\": [\"SW7-013\"], \"technology_data\": [\"SW7-013\"], \"deployment_scope\": [\"SW7-013\"], \"practical_relevance\": [\"SW7-013\"], \"economics\": [\"SW7-013\"], \"integration_human_role\": [\"SW7-013\"]}, \"disclosed_results\": [{\"metric\": \"Supplemental source: scope and limits\", \"value\": \"Near-zero maintenance is a qualitative platform description. The 40% M&A improvement and estimated >1,000 hours/year are not AI-specific realized results.\", \"period\": \"No publication/update date displayed; accessed 2026-10-01\", \"attribution\": \"Vendor-published; not independently evaluated\", \"source_ids\": [\"SW7-013\"]}], \"timeline\": [{\"date\": \"Undated; reviewed 2026-10-06\", \"event\": \"Snowflake architecture: operating drilling ML versus dashboard development\", \"source_ids\": [\"SW7-013\"]}], \"sources\": [{\"id\": \"SW7-013\", \"title\": \"Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter\", \"url\": \"https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/\", \"date\": \"No publication/update date displayed; accessed 2026-10-01\", \"type\": \"Vendor customer case\", \"publisher\": \"Expand Energy\", \"locator\": \"Digging deeper into operational efficiency; Energizing the future\", \"review_status\": \"Full original source reviewed in Luna verification and editor checks on 6 October 2026.\"}]}", "Research_status": "Supplemental source added · October gap review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-195", "company": "Expand Energy", "key": "expand-snowflake-drilling-ml", "outcome": "Independent review completed", "reviewer_outcome": "supported", "findings": [{"field": "Independent review", "issue": "Issuer source supports ML models for forecasting and real-time data/model hosting for drilling, plus SCADA alerts to the Operations Support Center. No result metric or specific model detail is disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Data integration benefits cannot be attributed entirely to AI.", "No independently measured drilling or failure-rate improvement is provided."], "source_urls_reviewed": ["https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-195", "stage": "Operating", "stage_basis": "Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.", "source_url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "source_date": "2026-03-03", "evidence_origin": "Operator", "review_correction": "Retain the dated primary source as AI-195’s anchor. Distinguish Snowflake workflows described as established/current from Leucipa/Lucy deployment beginning through OSC; not all AI workflows are fully rolled out.", "limits": "No quantified drilling/forecast lift, failure reduction, or alert precision disclosed. AI dashboard development is separately described as active development in the undated Snowflake customer story.", "source_id": "RS134"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Data integration benefits cannot be attributed entirely to AI. No independently measured drilling or failure-rate improvement is provided. Current platform/ML use and future dashboard/MLOps work have different stages; platform efficiency does not establish AI-only financial benefit. Same source and substantive claims already in AI-195. No additional adoption or numerical-result count. Site data coverage is not proof of automatic control at all 3,700 locations. Keep this platform separate from Leucipa in AI-194.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-196", "Company": "Viridien", "Company_primary": "Viridien", "Segment": "Upstream services & technology", "Theme": "AI products", "Initiative": "Evergreen AI-assisted subsurface data preparation", "Extracted_finding": "Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.", "Stage_at_disclosure": "Commercial solution launched", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Data-readiness claims are vendor statements. No time-saving benchmark, accuracy test or contract economics provided.", "Source_IDs": "RS135", "Source_date": "2026-06-10", "Source_type": "company announcement", "Source_locator": "Product launch description", "Source_URL": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-06-10; see chronology for operating dates.", "Context": "Fragmented geoscience information requires repeated manual preparation before interpretation.", "Implementation": "Evergreen classifies, curates, contextualizes and validates data, then ingests and standardizes it across files, records and applications.", "Assessment": "Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.", "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-root-003-1", "Initiative_ID": "AI-196", "Title": "Evergreen AI-assisted subsurface data preparation", "Date": "2026-06-10", "Finding": "Evergreen solution launch announced.", "Metric": "", "Source_IDs": "RS135", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-06-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-root-003"], "Research_search": " {\"key\": \"viridien-evergreen-data\", \"existing_ids\": [], \"company\": \"Viridien\", \"title\": \"Evergreen AI-assisted subsurface data preparation\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Commercial solution launched\", \"practical_relevance\": \"Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.\", \"business_problem\": \"Fragmented geoscience information requires repeated manual preparation before interpretation.\", \"workflow\": \"Evergreen classifies, curates, contextualizes and validates data, then ingests and standardizes it across files, records and applications.\", \"technology_data\": \"AI and domain data science cover structured, semi-structured and unstructured geoscience data; model identities are not disclosed.\", \"integration_human_role\": \"The product connects third-party applications across on-premise, cloud and hybrid environments. Human quality-control procedures are not specified.\", \"deployment_scope\": \"Marketed to subsurface asset teams; no named customer or installed footprint disclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.\", \"timeline\": [{\"date\": \"2026-06-10\", \"event\": \"Evergreen solution launch announced.\", \"source_ids\": [\"RS135\"]}], \"limitations\": [\"Data-readiness claims are vendor statements.\", \"No time-saving benchmark, accuracy test or contract economics provided.\"], \"assessment\": \"Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.\", \"sources\": [{\"id\": \"RS135\", \"title\": \"Viridien redefines subsurface data readiness at scale with Evergreen Data Solution\", \"url\": \"https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution\", \"publisher\": \"Viridien\", \"date\": \"2026-06-10\", \"type\": \"company announcement\", \"locator\": \"Product launch description\", \"review_status\": \"full text reviewed\", \"supports\": [\"AI-assisted classification, curation, contextualization and validation of subsurface data.\"], \"original_id\": \"root2-vir\"}], \"verification\": \"verified primary\", \"corrections_to_existing\": \"New second-pass research entry; no earlier initiative ID duplicated.\", \"field_sources\": {\"business_problem\": [\"RS135\"], \"workflow\": [\"RS135\"], \"technology_data\": [\"RS135\"], \"integration_human_role\": [\"RS135\"], \"deployment_scope\": [\"RS135\"], \"economics\": [\"RS135\"]}, \"cohort\": \"round2/additions-root\", \"packet_id\": \"round2-additions-root-003\", \"source_ids\": [\"RS135\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-196\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-196", "company": "Viridien", "key": "viridien-evergreen-data", "outcome": "Independent review completed", "reviewer_outcome": "supported as commercial product launch", "findings": [{"field": "Independent review", "issue": "Launch and AI-assisted subsurface data preparation capability supported. Label as a vendor product, not an operator deployment; no customer/results disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Data-readiness claims are vendor statements.", "No time-saving benchmark, accuracy test or contract economics provided."], "source_urls_reviewed": ["https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-197", "Company": "TGS", "Company_primary": "TGS", "Segment": "Upstream services & technology", "Theme": "AI products", "Initiative": "Seismic Foundation Model contract for an international operator", "Extracted_finding": "Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.", "Stage_at_disclosure": "Multiyear contract award; customer unnamed", "Reported_metric": "No quantified results disclosed", "Interpretation_limit": "Contract award establishes commercial adoption, not measured interpretation quality. No model size, benchmark, reserve impact or financial terms disclosed.", "Source_IDs": "RS136", "Source_date": "2026-08-18", "Source_type": "company announcement", "Source_locator": "Opening four paragraphs", "Source_URL": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "Evidence_status": "verified primary", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2026-08-18; see chronology for operating dates.", "Context": "Structural and stratigraphic interpretation is the target workflow.", "Implementation": "TGS will develop and deliver AI foundation-model capabilities for selected exploration and development interpretation tasks.", "Assessment": "Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.", "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Commercial AI product", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-root-004-1", "Initiative_ID": "AI-197", "Title": "Seismic Foundation Model contract for an international operator", "Date": "2026-08-18", "Finding": "Multiyear IOC award announced.", "Metric": "", "Source_IDs": "RS136", "Evidence": "verified primary", "Origin": "Research expansion"}], "Latest_disclosure": "2026-08-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-root-004"], "Research_search": " {\"key\": \"tgs-seismic-foundation-contract\", \"existing_ids\": [], \"company\": \"TGS\", \"title\": \"Seismic Foundation Model contract for an international operator\", \"department\": \"Exploration & geoscience\", \"secondary_departments\": [], \"category\": \"Commercial AI product\", \"stage\": \"Multiyear contract award; customer unnamed\", \"practical_relevance\": \"Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.\", \"business_problem\": \"Structural and stratigraphic interpretation is the target workflow.\", \"workflow\": \"TGS will develop and deliver AI foundation-model capabilities for selected exploration and development interpretation tasks.\", \"technology_data\": \"TGS SFM learns geological patterns from the company’s global multi-client seismic library.\", \"integration_human_role\": \"The announcement positions the model as interpreter support; review controls and model-acceptance criteria are not disclosed.\", \"deployment_scope\": \"An international oil company has contracted the capability; identity, asset locations and deployment schedule are undisclosed.\", \"partners\": [], \"disclosed_results\": [], \"economics\": \"Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.\", \"timeline\": [{\"date\": \"2026-08-18\", \"event\": \"Multiyear IOC award announced.\", \"source_ids\": [\"RS136\"]}], \"limitations\": [\"Contract award establishes commercial adoption, not measured interpretation quality.\", \"No model size, benchmark, reserve impact or financial terms disclosed.\"], \"assessment\": \"Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.\", \"sources\": [{\"id\": \"RS136\", \"title\": \"TGS Awarded Multi-Year AI Seismic Foundation Model Contract\", \"url\": \"https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract\", \"publisher\": \"TGS\", \"date\": \"2026-08-18\", \"type\": \"company announcement\", \"locator\": \"Opening four paragraphs\", \"review_status\": \"full text reviewed\", \"supports\": [\"Unnamed IOC contract; structural and stratigraphic interpretation; training on global multi-client seismic library.\"], \"original_id\": \"root2-tgs\"}], \"verification\": \"verified primary\", \"corrections_to_existing\": \"New second-pass research entry; no earlier initiative ID duplicated.\", \"field_sources\": {\"business_problem\": [\"RS136\"], \"workflow\": [\"RS136\"], \"technology_data\": [\"RS136\"], \"integration_human_role\": [\"RS136\"], \"deployment_scope\": [\"RS136\"], \"economics\": [\"RS136\"]}, \"cohort\": \"round2/additions-root\", \"packet_id\": \"round2-additions-root-004\", \"source_ids\": [\"RS136\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-197\"]}", "Research_status": "Expanded source review", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": {"id": "AI-197", "company": "TGS", "key": "tgs-seismic-foundation-contract", "outcome": "Independent review completed", "reviewer_outcome": "supported as vendor-reported contract award", "findings": [{"field": "Independent review", "issue": "Multi-year contract for an unnamed international operator and seismic interpretation scope supported. No value, customer identity or outcome is disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Contract award establishes commercial adoption, not measured interpretation quality.", "No model size, benchmark, reserve impact or financial terms disclosed."], "source_urls_reviewed": ["https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "verified primary", "independent_validation": "Not established by this collection"}}, {"ID": "AI-198", "Company": "Woodside", "Company_primary": "Woodside", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Trion integrated well-delivery contract with AI-enabled drilling", "Extracted_finding": "Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.", "Stage_at_disclosure": "Contract awarded; services scheduled for early2026 in the cited disclosure", "Reported_metric": "Contracted well-delivery scope: 18 ultra-deepwater wells (Vendor-reported)", "Interpretation_limit": "The cited April2025 release is an award and forward schedule, not a current progress certificate. No AI-attributable efficiency or well-quality result is disclosed.", "Source_IDs": "RS137", "Source_date": "2025-04-25", "Source_type": "earnings release", "Source_locator": "Core — Woodside Trion contract bullet", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "Evidence_status": "verified partner", "PE_relationship": null, "Coverage_bucket": "Research expansion", "Time_scope": "Dated source: 2025-04-25; see chronology for operating dates.", "Context": "The ultra-deepwater Trion project requires coordinated delivery of complex wells.", "Implementation": "SLB’s integrated contract includes AI-enabled drilling alongside directional services, logging, fluids, cementing and completions.", "Assessment": "Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-root-005-1", "Initiative_ID": "AI-198", "Title": "Trion integrated well-delivery contract with AI-enabled drilling", "Date": "2025-04-25", "Finding": "Award disclosed in quarterly results.", "Metric": "", "Source_IDs": "RS137", "Evidence": "verified partner", "Origin": "Research expansion"}, {"ID": "round2-additions-root-005-2", "Initiative_ID": "AI-198", "Title": "Trion integrated well-delivery contract with AI-enabled drilling", "Date": "2026 (planned)", "Finding": "Services scheduled to begin early in the year; completion not verified here.", "Metric": "", "Source_IDs": "RS137", "Evidence": "verified partner", "Origin": "Research expansion"}], "Latest_disclosure": "2025-04-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["round2-additions-root-005"], "Research_search": " {\"key\": \"woodside-trion-ai-drilling\", \"existing_ids\": [], \"company\": \"Woodside\", \"title\": \"Trion integrated well-delivery contract with AI-enabled drilling\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Contract awarded; services scheduled for early2026 in the cited disclosure\", \"practical_relevance\": \"Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.\", \"business_problem\": \"The ultra-deepwater Trion project requires coordinated delivery of complex wells.\", \"workflow\": \"SLB’s integrated contract includes AI-enabled drilling alongside directional services, logging, fluids, cementing and completions.\", \"technology_data\": \"The specific AI model and input design are not disclosed in the earnings bullet.\", \"integration_human_role\": \"Services are to be managed through SLB Performance Live delivery centers. Driller override and autonomous-control boundaries are unspecified.\", \"deployment_scope\": \"18 ultra-deepwater wells in Mexico; delivery completion is not established by the award disclosure.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"Contracted well-delivery scope\", \"value\": \"18 ultra-deepwater wells\", \"period\": \"Award disclosed April2025\", \"attribution\": \"Vendor-reported\", \"source_ids\": [\"RS137\"]}], \"economics\": \"Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.\", \"timeline\": [{\"date\": \"2025-04-25\", \"event\": \"Award disclosed in quarterly results.\", \"source_ids\": [\"RS137\"]}, {\"date\": \"2026 (planned)\", \"event\": \"Services scheduled to begin early in the year; completion not verified here.\", \"source_ids\": [\"RS137\"]}], \"limitations\": [\"The cited April2025 release is an award and forward schedule, not a current progress certificate.\", \"No AI-attributable efficiency or well-quality result is disclosed.\"], \"assessment\": \"Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.\", \"sources\": [{\"id\": \"RS137\", \"title\": \"SLB Announces First-Quarter 2025 Results\", \"url\": \"https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results\", \"publisher\": \"SLB\", \"date\": \"2025-04-25\", \"type\": \"earnings release\", \"locator\": \"Core — Woodside Trion contract bullet\", \"review_status\": \"excerpt reviewed\", \"supports\": [\"18-well integrated contract with AI-enabled drilling, planned early-2026 services and Performance Live delivery.\"], \"original_id\": \"root2-wood\"}], \"verification\": \"verified partner\", \"corrections_to_existing\": \"New second-pass research entry; no earlier initiative ID duplicated.\", \"field_sources\": {\"business_problem\": [\"RS137\"], \"workflow\": [\"RS137\"], \"technology_data\": [\"RS137\"], \"integration_human_role\": [\"RS137\"], \"deployment_scope\": [\"RS137\"], \"economics\": [\"RS137\"]}, \"cohort\": \"round2/additions-root\", \"packet_id\": \"round2-additions-root-005\", \"source_ids\": [\"RS137\"], \"report_date\": \"2026-09-27\", \"evidence_note\": \"Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.\", \"record_ids\": [\"AI-198\"]}", "Research_status": "Expanded source review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": {"id": "AI-198", "company": "Woodside", "key": "woodside-trion-ai-drilling", "outcome": "Independent review completed", "reviewer_outcome": "supported as planned contract scope", "findings": [{"field": "Independent review", "issue": "SLB states award covers 18 ultra-deepwater wells and includes AI-enabled drilling; services were scheduled to start in early 2026. Keep source attribution as SLB, and do not infer work actually commenced or that the contract produced measured results.", "correction": "", "source_url": ""}], "residual_gaps": ["The cited April2025 release is an award and forward schedule, not a current progress certificate.", "No AI-attributable efficiency or well-quality result is disclosed."], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Rests only on SLB's own earnings release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "verified partner", "independent_validation": "Not established by this collection"}}, {"ID": "AI-199", "Company": "Technip Energies", "Company_primary": "Technip Energies", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "Digital and AI transformation program — end-2025 progress", "Extracted_finding": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025.", "Stage_at_disclosure": "Mixed digital and AI program; end-2025 output counts", "Reported_metric": "The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.", "Interpretation_limit": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved. 2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.", "Source_IDs": "RS138; WATCH22PM-TEN", "Source_date": "2026 (exact publication day unverified)", "Source_type": "Indexed original annual-report excerpt", "Source_locator": "Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.", "Source_URL": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "Evidence_status": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "PE_relationship": null, "Coverage_bucket": "22 October evening targeted review", "Time_scope": "2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.", "Context": "Move useful digital and AI ideas into working tools and train staff to use them.", "Implementation": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established. Company transformation program; project and asset populations undisclosed.", "Assessment": "Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Workforce & learning"], "Department_assignment_basis": "Analyst workflow mapping; company organizational ownership not disclosed.", "Ambition_shortlist_category": "Not selected for original shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Luna research expansion", "22 October evening original-source review; previous evidence retained as dated history"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "round2-additions-root-006-1", "Initiative_ID": "AI-199", "Title": "Digital Acceleration Plan and AI for All workforce program", "Date": "2025", "Finding": "Shareholder message reports implementation of AI for All and progress on the digital plan.", "Metric": "", "Source_IDs": "RS138", "Evidence": "verified primary", "Origin": "Research expansion"}, {"ID": "WATCH22PM-AI-199", "Initiative_ID": "AI-199", "Title": "Technip Energies reports progress, without proving 55 AI deployments", "Date": "2026 (exact publication day unverified)", "Finding": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established.", "Metric": "The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.", "Source_IDs": "WATCH22PM-TEN", "Source_URL": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "Evidence": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved."}], "Latest_disclosure": "2026 (exact publication day unverified)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch22pm-ai-199", "round2-additions-root-006"], "Research_search": "{\"packet_id\": \"watch22pm-ai-199\", \"key\": \"watch22pm-ai-199\", \"record_ids\": [\"AI-199\"], \"company\": \"Technip Energies\", \"title\": \"Technip Energies reports progress, without proving 55 AI deployments\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-22\", \"verification\": \"verified primary excerpt\", \"stage\": \"Mixed digital and AI program; end-2025 output counts\", \"business_problem\": \"Move useful digital and AI ideas into working tools and train staff to use them.\", \"workflow\": \"Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established.\", \"technology_data\": \"Mixed digital and AI initiatives and staff upskilling; no inference of 55 AI production deployments.\", \"integration_human_role\": \"Program and upskilling outputs disclosed; project-level approval and execution boundaries unknown.\", \"deployment_scope\": \"Company transformation program; project and asset populations undisclosed. 2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.\", \"economics\": \"The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.\", \"practical_relevance\": \"Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?\", \"assessment\": \"Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?\", \"limitations\": [\"Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved.\", \"End-2025 program output counts, not a benefit baseline.\"], \"source_ids\": [\"WATCH22PM-TEN\"], \"partners\": [], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026 (exact publication day unverified)\", \"event\": \"Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025.\", \"source_ids\": [\"WATCH22PM-TEN\"]}], \"sources\": [{\"id\": \"WATCH22PM-TEN\", \"title\": \"Technip Energies 2025 Annual Report\", \"url\": \"https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf\", \"date\": \"2026 (exact publication day unverified)\", \"publisher\": \"Technip Energies\", \"locator\": \"Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.\", \"excerpt\": \"more than 55 digital and AI initiatives\", \"access\": \"Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.\", \"type\": \"Indexed original annual-report excerpt\", \"review_status\": \"Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.\"}], \"field_sources\": {\"business_problem\": [\"WATCH22PM-TEN\"], \"workflow\": [\"WATCH22PM-TEN\"], \"technology_data\": [\"WATCH22PM-TEN\"], \"integration_human_role\": [\"WATCH22PM-TEN\"], \"deployment_scope\": [\"WATCH22PM-TEN\"], \"economics\": [\"WATCH22PM-TEN\"]}}", "Research_status": "Updated 22 October 2026; indexed original excerpt with access limit", "Research_category": "Internal adoption", "Research_verification": "verified primary excerpt", "Audit_flags": [], "Audit_review": {"id": "AI-199", "company": "Technip Energies", "key": "technip-dap-ai-for-all", "outcome": "Independent review completed", "reviewer_outcome": "supported with economics/date caveat", "findings": [{"field": "Independent review", "issue": "Technip’s 2025 letter supports DAP launched in 2025, advanced analytics/LLM use, AI for All implementation and a €100m cost-savings-plus-incremental-revenue target by 2028. The target is not achieved savings; no workflow-level result or AI-only attribution is provided.", "correction": "", "source_url": ""}], "residual_gaps": ["The web page does not expose an exact publication date.", "Financial target combines digital cost savings and incremental revenue; it is not AI-only realized savings.", "Do not treat the workforce program as evidence all employees are active AI users."], "source_urls_reviewed": ["https://investors.technipenergies.com/shareholder-information"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, "Audit_outcome": "Independent review completed", "Segment_original": "Research expansion", "Segment_assignment_basis": "The initiative’s disclosed workflow aligns with the proposed segment; classification follows the workflow/asset described in the sourced record, not merely the company’s general identity.", "Segment_classification_confidence": "high", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Company investor disclosure", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "Mixed digital and AI program; end-2025 output counts Company transformation program; project and asset populations undisclosed.", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved.", "Outcome_counted": false, "Outcome_count_note": "Pilot installation, tool names and program outputs do not establish a measured operating or cash result.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "claim_status": "Mixed digital and AI program; end-2025 output counts", "independent_validation": "Not established"}, "Current_source_check": "2026-10-22", "Detail_words": 54, "Research_origin": ["research/watch-2026-10-22-pm/README.md"], "Current_review_note": "End-2025 mixed program outputs verified in indexed original excerpt. Earlier access audit is historical; full PDF remains unavailable."}, {"ID": "AI-200", "Company": "Shaya Ecuador / SLB", "Company_primary": "Shaya Ecuador", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Neuro autonomous geosteering", "Extracted_finding": "AI interprets live downhole measurements and changes the well trajectory.", "Stage_at_disclosure": "One Ecuador well disclosed in December 2024", "Reported_metric": "25 autonomous trajectory changes across a 2,392-ft lateral.", "Interpretation_limit": "SLB-reported; no numeric production uplift, control comparison or fleet-scale outcome disclosed.", "Source_IDs": "FU002", "Source_date": "2024-12-09", "Source_type": "Provider press release", "Source_locator": "Full article / relevant case section", "Source_URL": "https://www.slb.com/newsroom/press-release/2024/slb-adds-ai-geosteering", "Evidence_status": "Provider-authored source reviewed", "PE_relationship": null, "Coverage_bucket": "Supplemental technical / provider source", "Time_scope": "Dated disclosure 2024-12-09; operating dates are specified separately.", "Context": "Operating decisions require timely interpretation of field data.", "Implementation": "AI interprets live downhole measurements and changes the well trajectory.", "Assessment": "Functional deployment evidence; reported outcomes remain subject to the stated comparison and attribution limits.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Opportunity research follow-up"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "followup-ai-200", "Initiative_ID": "AI-200", "Title": "Neuro autonomous geosteering", "Date": "2024-12-09", "Finding": "AI interprets live downhole measurements and changes the well trajectory.", "Metric": "25 autonomous trajectory changes across a 2,392-ft lateral.", "Source_IDs": "FU002", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}], "Latest_disclosure": "2024-12-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["followup-ai-200"], "Research_search": "AI interprets live downhole measurements and changes the well trajectory.", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-28", "decision": "New profile", "note": "Source scope, chronology, metric and duplicate checks completed; no independent performance validation."}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Provider-authored source reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-201", "Company": "Chevron / OPX Ai", "Company_primary": "Chevron", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Kaybob integrated AI operations and lift control", "Extracted_finding": "IOCaaS links SCADA and historian data to surveillance and bounded gas-lift actions.", "Stage_at_disclosure": "Deployment began 2020; live by early 2022; retrospective disclosure April 2026", "Reported_metric": "5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.", "Interpretation_limit": "Vendor retrospective disclosed in April 2026 about a deployment begun in 2020. Facility changes and commodity prices also affected economics. Integration labor, service fees, recurring costs and a net-cash comparison are missing.", "Source_IDs": "FU003", "Source_date": "2026-04-01", "Source_type": "Vendor-authored technical case article", "Source_locator": "Full article / relevant case section", "Source_URL": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "Evidence_status": "Provider-authored source reviewed", "PE_relationship": null, "Coverage_bucket": "Supplemental technical / provider source", "Time_scope": "Dated disclosure 2026-04-01; operating dates are specified separately.", "Context": "Operating decisions require timely interpretation of field data.", "Implementation": "IOCaaS connects SCADA and historian data to surveillance and bounded lift adjustments. The vendor describes roughly 12 months of phased integration, including tag mapping and historical-data cleaning; fully live by early 2022.", "Assessment": "Functional deployment evidence; reported outcomes remain subject to the stated comparison and attribution limits.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Opportunity research follow-up"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "followup-ai-201", "Initiative_ID": "AI-201", "Title": "Kaybob integrated AI operations and lift control", "Date": "2026-04-01", "Finding": "IOCaaS links SCADA and historian data to surveillance and bounded gas-lift actions.", "Metric": "5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.", "Source_IDs": "FU003", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}], "Latest_disclosure": "2026-04-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-followup-ai-201", "followup-ai-201"], "Research_search": "{\"packet_id\": \"oct6-followup-ai-201\", \"review_date\": \"2026-10-06\", \"title\": \"6 October source review: Chevron\", \"workflow\": \"IOCaaS connects SCADA and historian data to surveillance and bounded lift adjustments. The vendor describes roughly 12 months of phased integration, including tag mapping and historical-data cleaning; fully live by early 2022.\", \"deployment_scope\": \"Deployment began 2020; live by early 2022; retrospective disclosure April 2026\", \"economics\": \"Full implementation and recurring costs are not disclosed.\", \"result_label\": \"Reported result and attribution limits\", \"attribution\": \"Vendor-authored retrospective\", \"verification\": \"verified partner\", \"limitations\": [\"Vendor retrospective disclosed in April 2026 about a deployment begun in 2020. Facility changes and commodity prices also affected economics. Integration labor, service fees, recurring costs and a net-cash comparison are missing.\"], \"timeline_event\": \"Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.\", \"source_review\": \"Full vendor-authored JPT article rechecked 6 October 2026\", \"timeline_date\": \"2026-10-06\", \"stage\": \"Deployment began 2020; live by early 2022; retrospective disclosure April 2026\", \"source_locator\": \"Kaybob and Montney deployment sections; integration and measurement-period paragraphs\", \"key\": \"oct6-followup-ai-201\", \"record_ids\": [\"AI-201\"], \"company\": \"Chevron\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"source_ids\": [\"FU003\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"FU003\"], \"workflow\": [\"FU003\"], \"technology_data\": [\"FU003\"], \"deployment_scope\": [\"FU003\"], \"practical_relevance\": [\"FU003\"], \"economics\": [\"FU003\"], \"integration_human_role\": [\"FU003\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and attribution limits\", \"value\": \"5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.\", \"period\": \"2026-04-01\", \"attribution\": \"Vendor-authored retrospective\", \"source_ids\": [\"FU003\"]}], \"timeline\": [{\"date\": \"2026-10-06\", \"event\": \"Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.\", \"source_ids\": [\"FU003\"]}], \"sources\": [{\"id\": \"FU003\", \"title\": \"Field deployments of AI-based IOCaaS\", \"url\": \"https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance\", \"date\": \"2026-04-01\", \"type\": \"Vendor-authored technical case article\", \"publisher\": \"SPE / JPT; OPX Ai authors\", \"locator\": \"Kaybob and Montney deployment sections; integration and measurement-period paragraphs\", \"review_status\": \"Full vendor-authored JPT article rechecked 6 October 2026\"}]}", "Research_status": "Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-28", "decision": "New profile", "note": "Source scope, chronology, metric and duplicate checks completed; no independent performance validation."}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Provider-authored source reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-202", "Company": "ConocoPhillips / OPX Ai", "Company_primary": "ConocoPhillips", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Montney AI surveillance and flow assurance", "Extracted_finding": "IOCaaS monitors lift, hydrate risk and compressor condition.", "Stage_at_disclosure": "Pilot. The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "Reported_metric": "3–4% production above forecast after four months of operation; about 5% lower LOE.", "Interpretation_limit": "Vendor-authored four-month operating comparison. 6% uplift was prospective, not observed. No new hardware does not mean zero cost; service fees, labor and net cash are undisclosed.", "Source_IDs": "FU003", "Source_date": "2026-04-01", "Source_type": "Vendor-authored technical case article", "Source_locator": "Full article / relevant case section", "Source_URL": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "Evidence_status": "Provider-authored source reviewed", "PE_relationship": null, "Coverage_bucket": "Supplemental technical / provider source", "Time_scope": "Dated disclosure 2026-04-01; operating dates are specified separately.", "Context": "Operating decisions require timely interpretation of field data.", "Implementation": "IOCaaS monitors lift, hydrate risk and compressors. The vendor reports four months from kickoff to full deployment, helped by standardized SCADA/historians and reused integration templates; operating results cover a further four months.", "Assessment": "Functional deployment evidence; reported outcomes remain subject to the stated comparison and attribution limits.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Opportunity research follow-up"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "followup-ai-202", "Initiative_ID": "AI-202", "Title": "Montney AI surveillance and flow assurance", "Date": "2026-04-01", "Finding": "IOCaaS monitors lift, hydrate risk and compressor condition.", "Metric": "3–4% production above forecast after four months of operation; about 5% lower LOE.", "Source_IDs": "FU003", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "SW7-D-AI-202", "Initiative_ID": "AI-202", "Title": "Montney AI surveillance, artificial lift and flow assurance", "Date": "2026-04-01", "Finding": "IOCaaS monitors Montney wells, gathering equipment and compressors for gas-lift opportunities, hydrate risk, liquid loading and equipment anomalies. Operators review alerts and act.", "Metric": "Vendor authors report the asset team calculated 3–4% production above forecast and ~5% lower LOE after four months; no hydrate outages in the evaluation. Approximately 6% future uplift is a projection.", "Source_IDs": "FU003", "Evidence": "Original full article opened/read, including the Montney deployment, architecture, timeline, outcomes, and authorship. The source gives operator/team attributions but does not provide underlying data or independent validation.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2026-04-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-followup-ai-202", "swarm7-ai-202", "followup-ai-202"], "Research_search": "{\"packet_id\": \"oct6-followup-ai-202\", \"review_date\": \"2026-10-06\", \"title\": \"6 October source review: ConocoPhillips\", \"workflow\": \"IOCaaS monitors lift, hydrate risk and compressors. The vendor reports four months from kickoff to full deployment, helped by standardized SCADA/historians and reused integration templates; operating results cover a further four months.\", \"deployment_scope\": \"Pilot. The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.\", \"economics\": \"Full implementation and recurring costs are not disclosed.\", \"result_label\": \"Reported result and attribution limits\", \"attribution\": \"Vendor-authored retrospective\", \"verification\": \"verified partner\", \"limitations\": [\"Vendor-authored four-month operating comparison. 6% uplift was prospective, not observed. No new hardware does not mean zero cost; service fees, labor and net cash are undisclosed.\"], \"timeline_event\": \"Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.\", \"source_review\": \"Full vendor-authored JPT article rechecked 6 October 2026\", \"timeline_date\": \"2026-10-06\", \"stage\": \"Pilot. The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.\", \"source_locator\": \"Kaybob and Montney deployment sections; integration and measurement-period paragraphs\", \"key\": \"oct6-followup-ai-202\", \"record_ids\": [\"AI-202\"], \"company\": \"ConocoPhillips\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"source_ids\": [\"FU003\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"FU003\"], \"workflow\": [\"FU003\"], \"technology_data\": [\"FU003\"], \"deployment_scope\": [\"FU003\"], \"practical_relevance\": [\"FU003\"], \"economics\": [\"FU003\"], \"integration_human_role\": [\"FU003\"]}, \"disclosed_results\": [{\"metric\": \"Reported result and attribution limits\", \"value\": \"3–4% production above forecast after four months of operation; about 5% lower LOE.\", \"period\": \"2026-04-01\", \"attribution\": \"Vendor-authored retrospective\", \"source_ids\": [\"FU003\"]}], \"timeline\": [{\"date\": \"2026-10-06\", \"event\": \"Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.\", \"source_ids\": [\"FU003\"]}], \"sources\": [{\"id\": \"FU003\", \"title\": \"Field deployments of AI-based IOCaaS\", \"url\": \"https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance\", \"date\": \"2026-04-01\", \"type\": \"Vendor-authored technical case article\", \"publisher\": \"SPE / JPT; OPX Ai authors\", \"locator\": \"Kaybob and Montney deployment sections; integration and measurement-period paragraphs\", \"review_status\": \"Full vendor-authored JPT article rechecked 6 October 2026\"}]}", "Research_status": "Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-28", "decision": "New profile", "note": "Source scope, chronology, metric and duplicate checks completed; no independent performance validation."}, "Normalized_stage": "Pilot", "Stage_review": {"id": "AI-202", "stage": "Pilot", "stage_basis": "The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "evidence_origin": "Vendor", "limits": "Four-month observation dates are not specified. Production/LOE and no-hydrate-outage claims are source/team-reported and not independently validated; projected uplift is not an achieved outcome.", "source_id": "FU003"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Four-month pilot result from a vendor-contributed JPT technical case.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Provider-authored source reviewed", "independent_validation": "Not established by this collection"}}, {"ID": "AI-203", "Company": "Marathon Pipe Line (MPLX / Marathon Petroleum)", "Company_primary": "Marathon Petroleum", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "AI-assisted pipeline right-of-way surveillance", "Extracted_finding": "Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.", "Stage_at_disclosure": "Deployment stated; expanding across operations", "Reported_metric": "No quantified detection, incident reduction, or cost result disclosed.", "Interpretation_limit": "Company interview from a senior MPL field-services/planning executive; operational claims are company-reported and lack independent performance data or specific rollout coverage.", "Source_IDs": "M29-001", "Source_date": "2025-04-09", "Source_type": "Operator interview / company news", "Source_locator": "Paragraphs under ‘What drives MPL’s success in safety, innovation and sustainability?’, especially lines 43–44 in web text; page dates article Apr 9 2025", "Source_URL": "https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/", "Evidence_status": "Full company page opened; article text directly accessible.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-04-09. Operating dates and deployment status are stated separately.", "Context": "MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.", "Implementation": "MPL says it adopted AI-driven monitoring technology and is deploying it across operations. The disclosure does not identify the vendor, sensor/data architecture, alert handling, or whether imagery comes from aircraft, satellites, or other sources.", "Assessment": "MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-203", "Initiative_ID": "AI-203", "Title": "AI-assisted pipeline right-of-way surveillance", "Date": "2025-04-09", "Finding": "Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.", "Metric": "No quantified detection, incident reduction, or cost result disclosed.", "Source_IDs": "M29-001", "Evidence": "Full company page opened; article text directly accessible.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-04-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-203"], "Research_search": "{\"key\": \"missing-sept29-ai-203\", \"packet_id\": \"missing-sept29-ai-203\", \"record_ids\": [\"AI-203\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Marathon Pipe Line (MPLX / Marathon Petroleum)\", \"title\": \"AI-assisted pipeline right-of-way surveillance\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Deployment stated; expanding across operations\", \"practical_relevance\": \"MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.\", \"business_problem\": \"\", \"workflow\": \"Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.\", \"technology_data\": \"MPL says it adopted AI-driven monitoring technology and is deploying it across operations. The disclosure does not identify the vendor, sensor/data architecture, alert handling, or whether imagery comes from aircraft, satellites, or other sources.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Deployment stated; expanding across operations\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"No quantified detection, incident reduction, or cost result disclosed.\", \"period\": \"2025-04-09\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-001\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-04-09\", \"event\": \"Deployment stated; expanding across operations\", \"source_ids\": [\"M29-001\"]}], \"limitations\": [\"Company interview from a senior MPL field-services/planning executive; operational claims are company-reported and lack independent performance data or specific rollout coverage.\"], \"assessment\": \"MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.\", \"sources\": [{\"id\": \"M29-001\", \"title\": \"Our People: How Paige Chenevey drives success at Marathon Pipe Line\", \"url\": \"https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/\", \"publisher\": \"Marathon Pipe Line (MPLX / Marathon Petroleum)\", \"date\": \"2025-04-09\", \"type\": \"Operator interview / company news\", \"locator\": \"Paragraphs under ‘What drives MPL’s success in safety, innovation and sustainability?’, especially lines 43–44 in web text; page dates article Apr 9 2025\", \"review_status\": \"Full company page opened; article text directly accessible.\", \"supports\": [\"Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.\", \"No quantified detection, incident reduction, or cost result disclosed.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-001\"], \"field_sources\": {\"workflow\": [\"M29-001\"], \"technology_data\": [\"M29-001\"], \"deployment_scope\": [\"M29-001\"], \"practical_relevance\": [\"M29-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "United States; across MPL operations (specific route/site not disclosed)", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full company page opened; article text directly accessible.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-204", "Company": "ORLEN Group", "Company_primary": "ORLEN", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "Group-wide AI rollout with market-monitoring and supply-chain assistants", "Extracted_finding": "Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.", "Stage_at_disclosure": "Rollout / implementation underway; use cases partly prospective", "Reported_metric": "More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.", "Interpretation_limit": "Company announcement states scope and intent; no measured outcomes and no specific production asset use case attributed to the new AI assistants.", "Source_IDs": "M29-002", "Source_date": "2025-07-08", "Source_type": "Operator press release", "Source_locator": "Web lines 312–323: AI deployment objectives; Copilot and Agentic AI; use cases monitoring global oil and gas market data and streamlining supply-chain logistics; internal competence centre to identify, develop and scale projects.", "Source_URL": "https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment", "Evidence_status": "Full English company announcement opened; exact application wording accessible.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-07-08. Operating dates and deployment status are stated separately.", "Context": "ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.", "Implementation": "Microsoft partnership covers consolidated licensing, secure cloud and intelligent assistants. ORLEN states solutions are being implemented and names Microsoft Copilot and Agentic AI; planned AI competence centre is intended to identify and scale higher-impact applications. This is enterprise deployment/rollout, not a single autonomous O&G control loop.", "Assessment": "ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-204", "Initiative_ID": "AI-204", "Title": "Group-wide AI rollout with market-monitoring and supply-chain assistants", "Date": "2025-07-08", "Finding": "Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.", "Metric": "More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.", "Source_IDs": "M29-002", "Evidence": "Full English company announcement opened; exact application wording accessible.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-07-08", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-204"], "Research_search": "{\"key\": \"missing-sept29-ai-204\", \"packet_id\": \"missing-sept29-ai-204\", \"record_ids\": [\"AI-204\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"ORLEN Group\", \"title\": \"Group-wide AI rollout with market-monitoring and supply-chain assistants\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Rollout / implementation underway; use cases partly prospective\", \"practical_relevance\": \"ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.\", \"business_problem\": \"\", \"workflow\": \"Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.\", \"technology_data\": \"Microsoft partnership covers consolidated licensing, secure cloud and intelligent assistants. ORLEN states solutions are being implemented and names Microsoft Copilot and Agentic AI; planned AI competence centre is intended to identify and scale higher-impact applications. This is enterprise deployment/rollout, not a single autonomous O&G control loop.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Rollout / implementation underway; use cases partly prospective\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.\", \"period\": \"2025-07-08\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-002\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-07-08\", \"event\": \"Rollout / implementation underway; use cases partly prospective\", \"source_ids\": [\"M29-002\"]}], \"limitations\": [\"Company announcement states scope and intent; no measured outcomes and no specific production asset use case attributed to the new AI assistants.\"], \"assessment\": \"ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.\", \"sources\": [{\"id\": \"M29-002\", \"title\": \"ORLEN’s transformation to gain momentum with CEE’s largest AI deployment\", \"url\": \"https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment\", \"publisher\": \"ORLEN Group\", \"date\": \"2025-07-08\", \"type\": \"Operator press release\", \"locator\": \"Web lines 312–323: AI deployment objectives; Copilot and Agentic AI; use cases monitoring global oil and gas market data and streamlining supply-chain logistics; internal competence centre to identify, develop and scale projects.\", \"review_status\": \"Full English company announcement opened; exact application wording accessible.\", \"supports\": [\"Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.\", \"More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-002\"], \"field_sources\": {\"workflow\": [\"M29-002\"], \"technology_data\": [\"M29-002\"], \"deployment_scope\": [\"M29-002\"], \"practical_relevance\": [\"M29-002\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "ORLEN Group; Central and Eastern Europe, more than a dozen countries", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full English company announcement opened; exact application wording accessible.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-205", "Company": "Apache Corporation / APA Corporation", "Company_primary": "APA Corporation", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "AI use in emissions reporting, logistics and reservoir-management decisions", "Extracted_finding": "Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.", "Stage_at_disclosure": "Use asserted as ongoing; specific maturity and deployment status unclear", "Reported_metric": "None disclosed.", "Interpretation_limit": "Company interview in Egypt Oil & Gas industry supplement. It names AI functions (emissions reporting, supply-chain logistics, reservoir decisions) but gives no implementation example, product, asset by use case, maturity by use case, or measured result. These are high-level company-reported claims.", "Source_IDs": "M29-003", "Source_date": "2024-11", "Source_type": "Operator executive interview in industry supplement / original report", "Source_locator": "Printed p. 14 (PDF page 14 / extracted web PDF P13), interview with Greg McDaniel, VP Apache Egypt Assets and Country Manager; technology question/answer lines 826–837. Supplement imprint: November 2024 at printed p. 3 / PDF P2; publisher publication listing corroborates.", "Source_URL": "https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence-1.pdf", "Evidence_status": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2024-11. Operating dates and deployment status are stated separately.", "Context": "Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.", "Implementation": "The interview names three functions and says Apache works with service and high-tech partners; no product, model, operating asset, deployment stage by use case, or implementation detail is supplied.", "Assessment": "Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-205", "Initiative_ID": "AI-205", "Title": "AI use in emissions reporting, logistics and reservoir-management decisions", "Date": "2024-11", "Finding": "Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.", "Metric": "None disclosed.", "Source_IDs": "M29-003", "Evidence": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2024-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-205"], "Research_search": "{\"key\": \"missing-sept29-ai-205\", \"packet_id\": \"missing-sept29-ai-205\", \"record_ids\": [\"AI-205\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Apache Corporation / APA Corporation\", \"title\": \"AI use in emissions reporting, logistics and reservoir-management decisions\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Use asserted as ongoing; specific maturity and deployment status unclear\", \"practical_relevance\": \"Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.\", \"business_problem\": \"\", \"workflow\": \"Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.\", \"technology_data\": \"The interview names three functions and says Apache works with service and high-tech partners; no product, model, operating asset, deployment stage by use case, or implementation detail is supplied.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Use asserted as ongoing; specific maturity and deployment status unclear\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"None disclosed.\", \"period\": \"2024-11\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-003\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2024-11\", \"event\": \"Use asserted as ongoing; specific maturity and deployment status unclear\", \"source_ids\": [\"M29-003\"]}], \"limitations\": [\"Company interview in Egypt Oil & Gas industry supplement. It names AI functions (emissions reporting, supply-chain logistics, reservoir decisions) but gives no implementation example, product, asset by use case, maturity by use case, or measured result. These are high-level company-reported claims.\"], \"assessment\": \"Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.\", \"sources\": [{\"id\": \"M29-003\", \"title\": \"Egypt Leading Oil & Gas Excellence Supplement — November 2024 (Apache interview)\", \"url\": \"https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence-1.pdf\", \"publisher\": \"Apache Corporation / APA Corporation\", \"date\": \"2024-11\", \"type\": \"Operator executive interview in industry supplement / original report\", \"locator\": \"Printed p. 14 (PDF page 14 / extracted web PDF P13), interview with Greg McDaniel, VP Apache Egypt Assets and Country Manager; technology question/answer lines 826–837. Supplement imprint: November 2024 at printed p. 3 / PDF P2; publisher publication listing corroborates.\", \"review_status\": \"Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.\", \"supports\": [\"Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.\", \"None disclosed.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-003\"], \"field_sources\": {\"workflow\": [\"M29-003\"], \"technology_data\": [\"M29-003\"], \"deployment_scope\": [\"M29-003\"], \"practical_relevance\": [\"M29-003\"]}, \"verification\": \"verified partner\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Egypt and Apache’s wider operations are referenced; no specific field tied to each use case", "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-206", "Company": "Murphy Oil Corporation", "Company_primary": "Murphy Oil", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "AI/ML-assisted continuous cyber monitoring and response", "Extracted_finding": "Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.", "Stage_at_disclosure": "In use", "Reported_metric": "No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.", "Interpretation_limit": "The AI/ML use case is cybersecurity continuous monitoring/detection and response, not hydrocarbon production optimization. The report does not isolate model role, vendors, OT scope or performance. Separate GenAI Hub/Copilot deployment is reported on p. 79.", "Source_IDs": "M29-004", "Source_date": "2025-08-06", "Source_type": "Operator sustainability report", "Source_locator": "Printed p. 78 (PDF page index P77), “Risk Management Strategy” Technology column: continuous monitoring and analytics built on advanced machine learning and AI; tools include endpoint/email protection and Managed Detection and Response. Printed p. 79 (PDF index P78) describes the Murphy Generative AI Hub and enterprise Microsoft Copilot rollout. Report printed p. 3 (PDF P2) gives publication date August 6, 2025; issuer Q2 announcement/update also dates Aug. 6 and confirms the report was published.", "Source_URL": "https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf", "Evidence_status": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-08-06. Operating dates and deployment status are stated separately.", "Context": "Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.", "Implementation": "Murphy reports internal and third-party AI/ML-backed monitoring and analytics within a layered cyber stack; AI use is explicit but it is a security workflow, not hydrocarbon production optimization.", "Assessment": "Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-206", "Initiative_ID": "AI-206", "Title": "AI/ML-assisted continuous cyber monitoring and response", "Date": "2025-08-06", "Finding": "Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.", "Metric": "No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.", "Source_IDs": "M29-004", "Evidence": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-08-06", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-206"], "Research_search": "{\"key\": \"missing-sept29-ai-206\", \"packet_id\": \"missing-sept29-ai-206\", \"record_ids\": [\"AI-206\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Murphy Oil Corporation\", \"title\": \"AI/ML-assisted continuous cyber monitoring and response\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"In use\", \"practical_relevance\": \"Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.\", \"business_problem\": \"\", \"workflow\": \"Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.\", \"technology_data\": \"Murphy reports internal and third-party AI/ML-backed monitoring and analytics within a layered cyber stack; AI use is explicit but it is a security workflow, not hydrocarbon production optimization.\", \"integration_human_role\": \"\", \"deployment_scope\": \"In use\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.\", \"period\": \"2025-08-06\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-004\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-08-06\", \"event\": \"In use\", \"source_ids\": [\"M29-004\"]}], \"limitations\": [\"The AI/ML use case is cybersecurity continuous monitoring/detection and response, not hydrocarbon production optimization. The report does not isolate model role, vendors, OT scope or performance. Separate GenAI Hub/Copilot deployment is reported on p. 79.\"], \"assessment\": \"Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.\", \"sources\": [{\"id\": \"M29-004\", \"title\": \"2025 Murphy Sustainability Report\", \"url\": \"https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf\", \"publisher\": \"Murphy Oil Corporation\", \"date\": \"2025-08-06\", \"type\": \"Operator sustainability report\", \"locator\": \"Printed p. 78 (PDF page index P77), “Risk Management Strategy” Technology column: continuous monitoring and analytics built on advanced machine learning and AI; tools include endpoint/email protection and Managed Detection and Response. Printed p. 79 (PDF index P78) describes the Murphy Generative AI Hub and enterprise Microsoft Copilot rollout. Report printed p. 3 (PDF P2) gives publication date August 6, 2025; issuer Q2 announcement/update also dates Aug. 6 and confirms the report was published.\", \"review_status\": \"Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.\", \"supports\": [\"Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.\", \"No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-004\"], \"field_sources\": {\"workflow\": [\"M29-004\"], \"technology_data\": [\"M29-004\"], \"deployment_scope\": [\"M29-004\"], \"practical_relevance\": [\"M29-004\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Enterprise-wide; asset geography not specified", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-207", "Company": "TotalEnergies", "Company_primary": "TotalEnergies", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit", "Extracted_finding": "Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.", "Stage_at_disclosure": "Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.", "Reported_metric": "Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.", "Interpretation_limit": "Preliminary, company-reported; sample size is five events, with no false-alarm rate, missed-event rate, production or emissions quantity, counterfactual, deployment cost or payback. The source does not establish autonomous control authority.", "Source_IDs": "M29-005", "Source_date": "2025-11-11", "Source_type": "Operator-hosted joint pilot announcement", "Source_locator": "Lines 11–19; five event predictions and 12-minute lead at line 17", "Source_URL": "https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial", "Evidence_status": "Full primary announcement reviewed.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-11-11. Operating dates and deployment status are stated separately.", "Context": "A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.", "Implementation": "AI assistant built on Honeywell Experion distributed control system; it provides predictive insights to operators. The disclosure describes operator corrective action, not autonomous process control.", "Assessment": "A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-207", "Initiative_ID": "AI-207", "Title": "Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit", "Date": "2025-11-11", "Finding": "Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.", "Metric": "Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.", "Source_IDs": "M29-005", "Evidence": "Full primary announcement reviewed.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-11-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-207"], "Research_search": "{\"key\": \"missing-sept29-ai-207\", \"packet_id\": \"missing-sept29-ai-207\", \"record_ids\": [\"AI-207\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"TotalEnergies\", \"title\": \"Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.\", \"practical_relevance\": \"A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.\", \"business_problem\": \"\", \"workflow\": \"Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.\", \"technology_data\": \"AI assistant built on Honeywell Experion distributed control system; it provides predictive insights to operators. The disclosure describes operator corrective action, not autonomous process control.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.\", \"period\": \"2025-11-11\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-005\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-11-11\", \"event\": \"Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.\", \"source_ids\": [\"M29-005\"]}], \"limitations\": [\"Preliminary, company-reported; sample size is five events, with no false-alarm rate, missed-event rate, production or emissions quantity, counterfactual, deployment cost or payback. The source does not establish autonomous control authority.\"], \"assessment\": \"A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.\", \"sources\": [{\"id\": \"M29-005\", \"title\": \"TotalEnergies and Honeywell Pilot AI-Assisted Control Room to Accelerated Shift to Industrial Autonomy\", \"url\": \"https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial\", \"publisher\": \"TotalEnergies\", \"date\": \"2025-11-11\", \"type\": \"Operator-hosted joint pilot announcement\", \"locator\": \"Lines 11–19; five event predictions and 12-minute lead at line 17\", \"review_status\": \"Full primary announcement reviewed.\", \"supports\": [\"Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.\", \"Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-005\"], \"field_sources\": {\"workflow\": [\"M29-005\"], \"technology_data\": [\"M29-005\"], \"deployment_scope\": [\"M29-005\"], \"practical_relevance\": [\"M29-005\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Port Arthur Refinery, Texas, United States", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full primary announcement reviewed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-208", "Company": "Snam / Engineering Group", "Company_primary": "Snam", "Segment": "Midstream", "Theme": "Enterprise AI", "Initiative": "AI extraction of gas-plant design data and equipment relationships", "Extracted_finding": "Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.", "Stage_at_disclosure": "Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.", "Reported_metric": "Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.", "Interpretation_limit": "The abstract gives different aggregate accuracy figures. Complex diagrams remain a weakness. No independently validated performance, cash saving, measured time saving or fleet rollout is established; the 11,000-plant fleet is contextual.", "Source_IDs": "M29-006", "Source_date": "2025-11-05", "Source_type": "Operator/provider-authored technical preprint", "Source_locator": "Sections 3.1, 4 (Tables 1–3) and 5; abstract uses different summary accuracy figures", "Source_URL": "https://arxiv.org/html/2511.08609v1", "Evidence_status": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-11-05. Operating dates and deployment status are stated separately.", "Context": "Adds a named midstream engineering-document workflow and its technical limitations.", "Implementation": "Combines OCR, language models, diagram recognition and constrained optimization. Snam supplied documents for approximately 1,500 regulating/metering plants; this is a dataset size, not a deployment count.", "Assessment": "Adds a named midstream engineering-document workflow and its technical limitations.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Research & development", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-208", "Initiative_ID": "AI-208", "Title": "AI extraction of gas-plant design data and equipment relationships", "Date": "2025-11-05", "Finding": "Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.", "Metric": "Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.", "Source_IDs": "M29-006", "Evidence": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-11-05", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-208"], "Research_search": "{\"key\": \"missing-sept29-ai-208\", \"packet_id\": \"missing-sept29-ai-208\", \"record_ids\": [\"AI-208\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Snam / Engineering Group\", \"title\": \"AI extraction of gas-plant design data and equipment relationships\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Research & development\", \"stage\": \"Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.\", \"practical_relevance\": \"Adds a named midstream engineering-document workflow and its technical limitations.\", \"business_problem\": \"\", \"workflow\": \"Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.\", \"technology_data\": \"Combines OCR, language models, diagram recognition and constrained optimization. Snam supplied documents for approximately 1,500 regulating/metering plants; this is a dataset size, not a deployment count.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.\", \"period\": \"2025-11-05\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-006\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-11-05\", \"event\": \"Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.\", \"source_ids\": [\"M29-006\"]}], \"limitations\": [\"The abstract gives different aggregate accuracy figures. Complex diagrams remain a weakness. No independently validated performance, cash saving, measured time saving or fleet rollout is established; the 11,000-plant fleet is contextual.\"], \"assessment\": \"Adds a named midstream engineering-document workflow and its technical limitations.\", \"sources\": [{\"id\": \"M29-006\", \"title\": \"Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants\", \"url\": \"https://arxiv.org/html/2511.08609v1\", \"publisher\": \"Snam / Engineering Group\", \"date\": \"2025-11-05\", \"type\": \"Operator/provider-authored technical preprint\", \"locator\": \"Sections 3.1, 4 (Tables 1–3) and 5; abstract uses different summary accuracy figures\", \"review_status\": \"Full original preprint reviewed; not represented as a peer-reviewed deployment study.\", \"supports\": [\"Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.\", \"Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-006\"], \"field_sources\": {\"workflow\": [\"M29-006\"], \"technology_data\": [\"M29-006\"], \"deployment_scope\": [\"M29-006\"], \"practical_relevance\": [\"M29-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Research & development", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Italian gas transportation network", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "independent", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-209", "Company": "PETRONAS", "Company_primary": "PETRONAS", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "PROTEAN predictive rotating-equipment analytics", "Extracted_finding": "Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.", "Stage_at_disclosure": "Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.", "Reported_metric": ">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.", "Interpretation_limit": "The report establishes deployment breadth, not the number of sites, time in service, model performance, false alarms, realized avoided downtime or savings. It gives no implementation cost or payback and does not quantify the claimed outcome in the text.", "Source_IDs": "RS017", "Source_date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "Source_type": "Operator integrated report", "Source_locator": "Printed pp. 54–55 (PDF page 29), Predicting Failures Before They Happen", "Source_URL": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "Evidence_status": "Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2026 (PETRONAS Integrated Report 2025; precise publication date not established). Operating dates and deployment status are stated separately.", "Context": "New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.", "Implementation": "PETRONAS reports deploying PROTEAN across Malaysia and international assets to monitor more than 350 critical rotating equipment items.", "Assessment": "New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-209", "Initiative_ID": "AI-209", "Title": "PROTEAN predictive rotating-equipment analytics", "Date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "Finding": "Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.", "Metric": ">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.", "Source_IDs": "RS017", "Evidence": "Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-209"], "Research_search": "{\"key\": \"missing-sept29-ai-209\", \"packet_id\": \"missing-sept29-ai-209\", \"record_ids\": [\"AI-209\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"PETRONAS\", \"title\": \"PROTEAN predictive rotating-equipment analytics\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.\", \"practical_relevance\": \"New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.\", \"business_problem\": \"\", \"workflow\": \"Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.\", \"technology_data\": \"PETRONAS reports deploying PROTEAN across Malaysia and international assets to monitor more than 350 critical rotating equipment items.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.\", \"period\": \"2026 (PETRONAS Integrated Report 2025; precise publication date not established)\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"RS017\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2026 (PETRONAS Integrated Report 2025; precise publication date not established)\", \"event\": \"Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.\", \"source_ids\": [\"RS017\"]}], \"limitations\": [\"The report establishes deployment breadth, not the number of sites, time in service, model performance, false alarms, realized avoided downtime or savings. It gives no implementation cost or payback and does not quantify the claimed outcome in the text.\"], \"assessment\": \"New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.\", \"sources\": [{\"id\": \"RS017\", \"title\": \"PETRONAS Integrated Report 2025\", \"url\": \"https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf\", \"publisher\": \"PETRONAS\", \"date\": \"2026 (PETRONAS Integrated Report 2025; precise publication date not established)\", \"type\": \"Operator integrated report\", \"locator\": \"Printed pp. 54–55 (PDF page 29), Predicting Failures Before They Happen\", \"review_status\": \"Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.\", \"supports\": [\"Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.\", \">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"RS017\"], \"field_sources\": {\"workflow\": [\"RS017\"], \"technology_data\": [\"RS017\"], \"deployment_scope\": [\"RS017\"], \"practical_relevance\": [\"RS017\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Malaysia and international assets; specific sites not named", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "350+ critical rotating equipment items", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "An earlier audit logged the URL inaccessible; the later source review downloaded the full report and read the relevant spread.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-210", "Company": "Diversified Energy Company (DEC)", "Company_primary": "Diversified Energy", "Segment": "Cross-segment / enterprise", "Theme": "Enterprise AI", "Initiative": "OneDEC producer-statement automation and contract repository for revenue accounting and settlements", "Extracted_finding": "Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.", "Stage_at_disclosure": "Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.", "Reported_metric": "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.", "Interpretation_limit": "Source is a company-authored annual sustainability report rather than a technical deployment postmortem. It reports environmental/operational data for 2025 but publishes in 2026. The workflow list sits under 'AI Efficiency at Scale,' yet no specific model is linked to statement automation or the contract repository. Architecture is company-reported; no independent validation of automation benefits. Financial controls are unusually specific, but no actual JIB or AFE workflow is claimed.", "Source_IDs": "M29-007; SW8-002", "Source_date": "2026-05-28", "Source_type": "Operator-published sustainability report; issuer release confirms report date", "Source_locator": "Printed p. 64 (PDF physical page 63), Artificial Intelligence / AI Efficiency at Scale; issuer release dated 28 May 2026", "Source_URL": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "Evidence_status": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2026-05-28. Operating dates and deployment status are stated separately.", "Context": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "Implementation": "OneDEC is described as a cloud-based data-centric operating environment. Microsoft Cloud provides identity/security/monitoring; operational and financial systems feed Snowflake through governed data contracts; ServiceNow enterprise playbooks provide approval/evidence workflows; APIs support portable model/agent use without data duplication. RAG agents use governed company data. Prompts, actions and outputs are versioned and monitored; logging through ServiceNow GRC supports SOX-aligned change control. The report does not specify document AI/OCR model vendors or whether producer-statement digitization itself uses generative AI.", "Assessment": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-210", "Initiative_ID": "AI-210", "Title": "OneDEC producer-statement automation and contract repository for revenue accounting and settlements", "Date": "2026-05-28", "Finding": "Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.", "Metric": "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.", "Source_IDs": "M29-007", "Evidence": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2026-05-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct4-ai-210", "missing-sept29-ai-210"], "Research_search": "{\"key\": \"missing-sept29-ai-210\", \"packet_id\": \"missing-sept29-ai-210\", \"record_ids\": [\"AI-210\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Diversified Energy Company (DEC)\", \"title\": \"OneDEC producer-statement automation and contract repository for revenue accounting and settlements\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.\", \"practical_relevance\": \"Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.\", \"business_problem\": \"\", \"workflow\": \"Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.\", \"technology_data\": \"OneDEC connects governed data in Snowflake with ServiceNow approvals and RAG agents. Human oversight and monitored/versioned outputs are described at platform level; the AI contribution to each financial workflow is not isolated.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.\", \"period\": \"2026-05-28\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-007\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2026-05-28\", \"event\": \"Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.\", \"source_ids\": [\"M29-007\"]}], \"limitations\": [\"No quantified financial outcome or workflow-specific deployment denominator. Platform architecture does not prove every statement or contract step uses AI.\"], \"assessment\": \"Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.\", \"sources\": [{\"id\": \"M29-007\", \"title\": \"Diversified Energy 2025 Sustainability Report\", \"url\": \"https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf\", \"publisher\": \"Diversified Energy Company (DEC)\", \"date\": \"2026-05-28\", \"type\": \"Operator-published sustainability report; issuer release confirms report date\", \"locator\": \"Printed p. 64 (PDF physical page 63), Artificial Intelligence / AI Efficiency at Scale; issuer release dated 28 May 2026\", \"review_status\": \"Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.\", \"supports\": [\"Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.\", \"No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-007\"], \"field_sources\": {\"workflow\": [\"M29-007\"], \"technology_data\": [\"M29-007\"], \"deployment_scope\": [\"M29-007\"], \"practical_relevance\": [\"M29-007\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"} {\"key\": \"oct4-ai-210\", \"packet_id\": \"oct4-ai-210\", \"record_ids\": [\"AI-210\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Diversified Energy Company (DEC)\", \"title\": \"OneDEC: producer-statement normalization and governed finance/operations AI foundation\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.\", \"practical_relevance\": \"Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.\", \"business_problem\": \"\", \"workflow\": \"New company-authored detail for the existing OneDEC profile: statement automation digitizes and normalizes producer data across all sales points to reduce re-keying for Production Accounting, Marketing and Revenue Distribution; a centralized contract repository is described as the authoritative source for commercial terms used in pricing and settlements. The report also describes a single historian supporting manage-by-exception surveillance. This is an enrichment of the existing work, not a new duplicate record.\", \"technology_data\": \"OneDEC is described as a cloud-based data-centric operating environment. Microsoft Cloud provides identity/security/monitoring; operational and financial systems feed Snowflake through governed data contracts; ServiceNow enterprise playbooks provide approval/evidence workflows; APIs support portable model/agent use without data duplication. RAG agents use governed company data. Prompts, actions and outputs are versioned and monitored; logging through ServiceNow GRC supports SOX-aligned change control. The report does not specify document AI/OCR model vendors or whether producer-statement digitization itself uses generative AI.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.\", \"period\": \"2026-05-28\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-007\"]}], \"economics\": \"No quantified AI/statement-automation savings, costs or payback. ROI is an evaluation gate for each AI use case, not a reported result. Separate company-wide financial performance should not be attributed to this program.\", \"timeline\": [{\"date\": \"2026-05-28\", \"event\": \"Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.\", \"source_ids\": [\"M29-007\"]}], \"limitations\": [\"Source is a company-authored annual sustainability report rather than a technical deployment postmortem. It reports environmental/operational data for 2025 but publishes in 2026. The workflow list sits under 'AI Efficiency at Scale,' yet no specific model is linked to statement automation or the contract repository. Architecture is company-reported; no independent validation of automation benefits. Financial controls are unusually specific, but no actual JIB or AFE workflow is claimed.\"], \"assessment\": \"Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.\", \"sources\": [{\"title\": \"Diversified Energy Company Demonstrates Proven Sustainability Leadership (2025 Sustainability Report release)\", \"url\": \"https://ir.div.energy/news-events/us-press-releases/detail/225/diversified-energy-company-demonstrates-proven-sustainability-leadership\", \"date\": \"2026-05-28 (issuer release announcing 2025 report); IR report index separately lists June 17, 2026\", \"locator\": \"Official issuer release confirms report publication; report content fully opened/read in linked PDF. OneDEC/Snowflake/ServiceNow architecture: report pp. 62-63; data and controls, HITL, audit, AI governance: p. 63; statement automation, historian surveillance and pricing/settlements: p. 63. Report data period is calendar 2025 (p. 67).\", \"review_status\": \"Full issuer-authored report PDF opened/read and cross-checked against official IR release and report index. Metric year (2025) and publication date (2026) kept separate.\", \"id\": \"SW8-002\", \"type\": \"Issuer report\", \"accessed\": \"2026-10-04\"}, {\"title\": \"Diversified 2025 Sustainability Report\", \"url\": \"https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf\", \"date\": \"Announced 2026-05-28; report index lists 2026-06-17\", \"locator\": \"Full 85-page issuer PDF read; especially pp. 62-63 lines 4044-4217 and reporting-period notes p. 67 lines 4256-4274.\", \"review_status\": \"Full primary issuer document reviewed. No dated workflow-specific operational metric disclosed.\", \"id\": \"M29-007\", \"type\": \"Issuer report\", \"accessed\": \"2026-10-04\"}], \"source_ids\": [\"SW8-002\", \"M29-007\"], \"field_sources\": {\"workflow\": [\"SW8-002\", \"M29-007\"], \"technology_data\": [\"SW8-002\", \"M29-007\"], \"stage\": [\"SW8-002\", \"M29-007\"], \"economics\": [\"SW8-002\", \"M29-007\"]}, \"verification\": \"verified primary\", \"cohort\": \"October 4 evidence review\", \"report_date\": \"2026-10-04\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "United States; individual workflow rollout locations not specified", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-211", "Company": "Trafigura", "Company_primary": "Trafigura", "Segment": "Trading & shipping", "Theme": "Enterprise AI", "Initiative": "AI-enabled maritime voyage optimization across controlled fleet", "Extracted_finding": "AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.", "Stage_at_disclosure": "Planned fleet rollout in July 2025; completion and active vessel count unverified.", "Reported_metric": "No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.", "Interpretation_limit": "Future-tense announcement. Fleet size is not a verified AI usage count, and expected savings are not operating results.", "Source_IDs": "M29-008", "Source_date": "2025-07-09", "Source_type": "Operator press release with named partner and attributed Global Head of Shipping", "Source_locator": "Trafigura press release, 9 July 2025, paragraphs 2-5", "Source_URL": "https://www.trafigura.com/news-and-insights/press-releases/2025/trafigura-announces-strategic-alliance-with-maritime-technology-provider-zeronorth/", "Evidence_status": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-07-09. Operating dates and deployment status are stated separately.", "Context": "No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.", "Implementation": "Trafigura announced a ZeroNorth alliance and equity investment, with rollout intended across more than 350 controlled vessels. Human approvals and control authority are not described.", "Assessment": "No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.", "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-211", "Initiative_ID": "AI-211", "Title": "AI-enabled maritime voyage optimization across controlled fleet", "Date": "2025-07-09", "Finding": "AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.", "Metric": "No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.", "Source_IDs": "M29-008", "Evidence": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-07-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-211"], "Research_search": "{\"key\": \"missing-sept29-ai-211\", \"packet_id\": \"missing-sept29-ai-211\", \"record_ids\": [\"AI-211\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Trafigura\", \"title\": \"AI-enabled maritime voyage optimization across controlled fleet\", \"department\": \"Supply chain & logistics\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Planned fleet rollout in July 2025; completion and active vessel count unverified.\", \"practical_relevance\": \"No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.\", \"business_problem\": \"\", \"workflow\": \"AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.\", \"technology_data\": \"Trafigura announced a ZeroNorth alliance and equity investment, with rollout intended across more than 350 controlled vessels. Human approvals and control authority are not described.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Planned fleet rollout in July 2025; completion and active vessel count unverified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.\", \"period\": \"2025-07-09\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-008\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-07-09\", \"event\": \"Planned fleet rollout in July 2025; completion and active vessel count unverified.\", \"source_ids\": [\"M29-008\"]}], \"limitations\": [\"Future-tense announcement. Fleet size is not a verified AI usage count, and expected savings are not operating results.\"], \"assessment\": \"No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.\", \"sources\": [{\"id\": \"M29-008\", \"title\": \"Trafigura announces strategic alliance with maritime technology provider ZeroNorth\", \"url\": \"https://www.trafigura.com/news-and-insights/press-releases/2025/trafigura-announces-strategic-alliance-with-maritime-technology-provider-zeronorth/\", \"publisher\": \"Trafigura\", \"date\": \"2025-07-09\", \"type\": \"Operator press release with named partner and attributed Global Head of Shipping\", \"locator\": \"Trafigura press release, 9 July 2025, paragraphs 2-5\", \"review_status\": \"Full original Trafigura press release opened and read; date and rollout phrasing verified in source.\", \"supports\": [\"AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.\", \"No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-008\"], \"field_sources\": {\"workflow\": [\"M29-008\"], \"technology_data\": [\"M29-008\"], \"deployment_scope\": [\"M29-008\"], \"practical_relevance\": [\"M29-008\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Global fleet; no specific operating geography disclosed", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Cross-segment", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-212", "Company": "SM Energy", "Company_primary": "SM Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Machine-learning gas-lift injection-rate optimization in Midland Basin production operations", "Extracted_finding": "A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.", "Stage_at_disclosure": "Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.", "Reported_metric": "Management reports incremental production without a numerical uplift, baseline or measurement period.", "Interpretation_limit": "Company-authored conference transcript; no independently measured result. Recommendations do not establish autonomous valve control. Civitas extension was prospective.", "Source_IDs": "M29-009", "Source_date": "2025-11-21", "Source_type": "Company conference transcript filed with SEC under Rule 425", "Source_locator": "Transcript p. 7, COO Beth McDonald; conference 20 November, filing 21 November 2025", "Source_URL": "https://www.sec.gov/Archives/edgar/data/893538/000089353825000160/stephensannualinvestorco.htm", "Evidence_status": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "PE_relationship": null, "Coverage_bucket": "Targeted missing-company and workflow search", "Time_scope": "Disclosure: 2025-11-21. Operating dates and deployment status are stated separately.", "Context": "SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.", "Implementation": "Management describes a compression-linked analytics platform and engineers applying learned practices. Vendor, well count, execution controls and validation method are undisclosed.", "Assessment": "SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Gap review · 29 September 2026"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "M29-D-AI-212", "Initiative_ID": "AI-212", "Title": "Machine-learning gas-lift injection-rate optimization in Midland Basin production operations", "Date": "2025-11-21", "Finding": "A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.", "Metric": "Management reports incremental production without a numerical uplift, baseline or measurement period.", "Source_IDs": "M29-009", "Evidence": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "Origin": "29 September 2026 gap review"}], "Latest_disclosure": "2025-11-21", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["missing-sept29-ai-212"], "Research_search": "{\"key\": \"missing-sept29-ai-212\", \"packet_id\": \"missing-sept29-ai-212\", \"record_ids\": [\"AI-212\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"SM Energy\", \"title\": \"Machine-learning gas-lift injection-rate optimization in Midland Basin production operations\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.\", \"practical_relevance\": \"SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.\", \"business_problem\": \"\", \"workflow\": \"A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.\", \"technology_data\": \"Management describes a compression-linked analytics platform and engineers applying learned practices. Vendor, well count, execution controls and validation method are undisclosed.\", \"integration_human_role\": \"\", \"deployment_scope\": \"Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Reported evidence\", \"value\": \"Management reports incremental production without a numerical uplift, baseline or measurement period.\", \"period\": \"2025-11-21\", \"attribution\": \"Source-reported; see evidence limits\", \"source_ids\": [\"M29-009\"]}], \"economics\": \"Deployment cost and cash payback are not established by this source.\", \"timeline\": [{\"date\": \"2025-11-21\", \"event\": \"Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.\", \"source_ids\": [\"M29-009\"]}], \"limitations\": [\"Company-authored conference transcript; no independently measured result. Recommendations do not establish autonomous valve control. Civitas extension was prospective.\"], \"assessment\": \"SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.\", \"sources\": [{\"id\": \"M29-009\", \"title\": \"SM Energy Company — Fireside Chat Stephens Annual Investment Conference (internally generated transcript; Rule 425 filing)\", \"url\": \"https://www.sec.gov/Archives/edgar/data/893538/000089353825000160/stephensannualinvestorco.htm\", \"publisher\": \"SM Energy\", \"date\": \"2025-11-21\", \"type\": \"Company conference transcript filed with SEC under Rule 425\", \"locator\": \"Transcript p. 7, COO Beth McDonald; conference 20 November, filing 21 November 2025\", \"review_status\": \"Full SEC EDGAR transcript and issuer-hosted copy independently checked.\", \"supports\": [\"A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.\", \"Management reports incremental production without a numerical uplift, baseline or measurement period.\"], \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"M29-009\"], \"field_sources\": {\"workflow\": [\"M29-009\"], \"technology_data\": [\"M29-009\"], \"deployment_scope\": [\"M29-009\"], \"practical_relevance\": [\"M29-009\"]}, \"verification\": \"verified primary\", \"cohort\": \"Gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Source review establishes the disclosure, not independently validated performance.\"}", "Research_status": "Expanded dossier", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Source, date, stage, metric and duplicate review; source attribution retained."}, "Asset_location": "Midland Basin, Texas", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-213", "Company": "Koch Energy Services", "Company_primary": "Koch Energy Services", "Segment": "Trading & shipping", "Theme": "Operational AI", "Initiative": "LLM-assisted physical gas trade capture from chat into ETRM", "Extracted_finding": "An LLM converts abbreviated trader chat into structured physical-gas trade fields for booking in Koch’s ETRM system.", "Stage_at_disclosure": "Named Koch executive reports operating use; announcement body dated 18 September 2025, webpage metadata dated 26 November.", "Reported_metric": "Koch reports 25 traders, more than 35,000 trades through the platform, and approximately 75% less trade-input time.", "Interpretation_limit": "Platform trade count does not establish that every trade was AI-parsed. No baseline minutes or measurement method. Later undated 90,000-trade/85% figures are not treated as dated progress.", "Source_IDs": "SW2-001; SW2-002", "Source_date": "2025-09-18", "Source_type": "Vendor announcement with named customer quote", "Source_locator": "Opening dateline; Darnell Bortz, Director of Natural Gas Trading quote", "Source_URL": "https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat", "Evidence_status": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "Release body: 18 September 2025. Webpage metadata: 26 November 2025. Companion implementation page is undated.", "Context": "Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.", "Implementation": "Source chat stays linked to the trade for checking and downstream reconciliation. An undated vendor case study describes 12–13 extracted fields and trader validation before submission.", "Assessment": "Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.", "Primary_department": "Trading & marketing", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-213", "Initiative_ID": "AI-213", "Title": "LLM-assisted physical gas trade capture from chat into ETRM", "Date": "2025-09-18", "Finding": "An LLM converts abbreviated trader chat into structured physical-gas trade fields for booking in Koch’s ETRM system.", "Metric": "Koch reports 25 traders, more than 35,000 trades through the platform, and approximately 75% less trade-input time.", "Source_IDs": "SW2-001; SW2-002", "Evidence": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2025-09-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-213"], "Research_search": "{\"key\": \"swarm2-ai-213\", \"packet_id\": \"swarm2-ai-213\", \"record_ids\": [\"AI-213\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Koch Energy Services\", \"title\": \"LLM-assisted physical gas trade capture from chat into ETRM\", \"department\": \"Trading & marketing\", \"secondary_departments\": [\"Finance, investment & strategy\"], \"category\": \"Internal adoption\", \"stage\": \"Named Koch executive reports operating use; announcement body dated 18 September 2025, webpage metadata dated 26 November.\", \"practical_relevance\": \"Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.\", \"business_problem\": \"\", \"workflow\": \"An LLM converts abbreviated trader chat into structured physical-gas trade fields for booking in Koch’s ETRM system.\", \"technology_data\": \"Source chat stays linked to the trade for checking and downstream reconciliation. An undated vendor case study describes 12–13 extracted fields and trader validation before submission.\", \"integration_human_role\": \"Traders validate and submit generated order tickets, according to the undated case study.\", \"deployment_scope\": \"Named Koch executive reports operating use; announcement body dated 18 September 2025, webpage metadata dated 26 November.\", \"partners\": [\"ipushpull\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Koch reports 25 traders, more than 35,000 trades through the platform, and approximately 75% less trade-input time.\", \"period\": \"At the September 2025 announcement; measurement period unstated\", \"attribution\": \"Customer executive statement published by vendor; not independently validated.\", \"source_ids\": [\"SW2-001\", \"SW2-002\"]}], \"economics\": \"Process-time improvement; no labor-cost realization, implementation cost or payback disclosed.\", \"timeline\": [{\"date\": \"2025-09-18\", \"event\": \"Release body: 18 September 2025. Webpage metadata: 26 November 2025. Companion implementation page is undated.\", \"source_ids\": [\"SW2-001\", \"SW2-002\"]}], \"limitations\": [\"Platform trade count does not establish that every trade was AI-parsed. No baseline minutes or measurement method. Later undated 90,000-trade/85% figures are not treated as dated progress.\"], \"assessment\": \"Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.\", \"sources\": [{\"title\": \"ipushpull launches LLM-enabled agent to automate physical gas trade booking via chat\", \"url\": \"https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat\", \"date\": \"2025-09-18\", \"type\": \"Vendor announcement with named customer quote\", \"locator\": \"Opening dateline; Darnell Bortz, Director of Natural Gas Trading quote\", \"review_status\": \"Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.\", \"publisher\": \"ipushpull\", \"id\": \"SW2-001\", \"accessed\": \"2026-09-29\"}, {\"title\": \"Koch Energy Services and ipushpull automate physical gas trade booking through chat\", \"url\": \"https://ipushpull.com/case-studies/koch-energy-services-and-ipushpull-automate-physical-gas-trade-booking-through-chat\", \"date\": \"Undated\", \"type\": \"Vendor customer case study\", \"locator\": \"Reimagining trade capture; Human-in-the-loop\", \"review_status\": \"Indexed full source reviewed; used for workflow controls, not dated outcome update.\", \"publisher\": \"ipushpull\", \"id\": \"SW2-002\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-001\", \"SW2-002\"], \"field_sources\": {\"workflow\": [\"SW2-001\", \"SW2-002\"], \"technology_data\": [\"SW2-001\", \"SW2-002\"], \"deployment_scope\": [\"SW2-001\", \"SW2-002\"], \"practical_relevance\": [\"SW2-001\", \"SW2-002\"]}, \"verification\": \"verified partner\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Customer executive statement published by vendor; not independently validated.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "North American natural-gas trading", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Cross-segment", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-214", "Company": "Sentinel Peak Resources", "Company_primary": "Sentinel Peak Resources", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Intelligent Oilfield steam-injection optimization with Quantum", "Extracted_finding": "An AI-based thermal-management tool supports lower steam-injection rates in California heavy-oil operations.", "Stage_at_disclosure": "Reported operational use before the disclosure window; included for an in-window sponsor disclosure of earlier results.", "Reported_metric": "For 2023, the case study reports more than 3 million barrels less steam injection and over 120,000 metric tons less Scope 1 GHG.", "Interpretation_limit": "Results predate the 24-month disclosure window and include stewardship and analysis beyond AI. AI-only causality and baseline method are unproven. May 2025 upload is established by source path; exact May 2 date is file-indicated.", "Source_IDs": "RS076; SW5-016", "Source_date": "2025-05-02", "Source_type": "Sponsor report", "Source_locator": "Printed/PDF page 91, Sentinel Peak Resources", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "Evidence_status": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.", "PE_relationship": "Quantum Energy Partners portfolio company; sponsor report identifies investment in 2016 and joint tool development.", "Coverage_bucket": "Second targeted gap review", "Time_scope": "2024-titled report uploaded May 2025; filename indicates 2 May. Reported operating results relate to 2023.", "Context": "A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.", "Implementation": "Sentinel Peak developed the Intelligent Oilfield tool with Quantum’s digital team and applied steam analysis to acquired Seneca assets.", "Assessment": "A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.", "Primary_department": "Production & field operations", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-214", "Initiative_ID": "AI-214", "Title": "Intelligent Oilfield steam-injection optimization with Quantum", "Date": "2025-05-02", "Finding": "An AI-based thermal-management tool supports lower steam-injection rates in California heavy-oil operations.", "Metric": "For 2023, the case study reports more than 3 million barrels less steam injection and over 120,000 metric tons less Scope 1 GHG.", "Source_IDs": "RS076", "Evidence": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.", "Origin": "Second September research swarm"}, {"ID": "SW5-D-AI-214", "Initiative_ID": "AI-214", "Title": "Undated operator corroboration of Intelligent Oilfield steam optimization", "Date": "Undated", "Finding": "Sentinel Peak identifies machine-learning and AI analysis of reservoir data for real-time steam management.", "Metric": "The operator estimates 200,000 metric tons CO2e per year enabled through intelligent-oilfield technologies. Period, baseline and AI-only contribution are unspecified.", "Source_IDs": "SW5-016", "Evidence": "Full original HTML page opened and read. It names Sentinel Peak's Intelligent Oilfield as an ML/AI-powered data system for real-time steam-management optimization and estimates 200,000 MT CO2e/year enabled through intelligent oilfield technologies. The page gives no publication date, calculation, baseline, or AI-only attribution split.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2025-05-02", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-214", "swarm2-ai-214"], "Research_search": "{\"key\": \"swarm2-ai-214\", \"packet_id\": \"swarm2-ai-214\", \"record_ids\": [\"AI-214\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Sentinel Peak Resources\", \"title\": \"Intelligent Oilfield steam-injection optimization with Quantum\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"HSE & emissions\"], \"category\": \"Internal adoption\", \"stage\": \"Reported operational use before the disclosure window; included for an in-window sponsor disclosure of earlier results.\", \"practical_relevance\": \"A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.\", \"business_problem\": \"\", \"workflow\": \"An AI-based thermal-management tool supports lower steam-injection rates in California heavy-oil operations.\", \"technology_data\": \"Sentinel Peak developed the Intelligent Oilfield tool with Quantum’s digital team and applied steam analysis to acquired Seneca assets.\", \"integration_human_role\": \"Review and actuation controls are not disclosed.\", \"deployment_scope\": \"Reported operational use before the disclosure window; included for an in-window sponsor disclosure of earlier results.\", \"partners\": [\"Quantum Energy Partners\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"For 2023, the case study reports more than 3 million barrels less steam injection and over 120,000 metric tons less Scope 1 GHG.\", \"period\": \"2023 operating results reported in 2024-titled report uploaded in May 2025\", \"attribution\": \"Portfolio-company data; report explicitly says Quantum and third parties have not verified it.\", \"source_ids\": [\"RS076\"]}], \"economics\": \"No quantified cash saving, deployment cost or payback disclosed.\", \"timeline\": [{\"date\": \"2025-05-02\", \"event\": \"2024-titled report uploaded May 2025; filename indicates 2 May. Reported operating results relate to 2023.\", \"source_ids\": [\"RS076\"]}], \"limitations\": [\"Results predate the 24-month disclosure window and include stewardship and analysis beyond AI. AI-only causality and baseline method are unproven. May 2025 upload is established by source path; exact May 2 date is file-indicated.\"], \"assessment\": \"A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.\", \"sources\": [{\"title\": \"Quantum Capital Group 2024 Stakeholder Report — Sentinel Peak case study\", \"url\": \"https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf\", \"date\": \"2025-05-02 (file-indicated)\", \"type\": \"Sponsor report\", \"locator\": \"Printed/PDF page 91, Sentinel Peak Resources\", \"review_status\": \"Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.\", \"publisher\": \"Quantum Capital Group\", \"id\": \"RS076\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"RS076\"], \"field_sources\": {\"workflow\": [\"RS076\"], \"technology_data\": [\"RS076\"], \"deployment_scope\": [\"RS076\"], \"practical_relevance\": [\"RS076\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Portfolio-company data; report explicitly says Quantum and third parties have not verified it.\"} {\"key\": \"swarm5-ai-214\", \"packet_id\": \"swarm5-ai-214\", \"record_ids\": [\"AI-214\"], \"existing_ids\": [\"AI-214\"], \"related_existing_ids\": [], \"company\": \"Sentinel Peak Resources / Quantum Energy Partners\", \"title\": \"Undated operator corroboration of Intelligent Oilfield steam optimization\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"HSE & emissions\"], \"category\": \"Internal adoption\", \"stage\": \"Operator describes use; undated webpage does not establish a new rollout.\", \"practical_relevance\": \"Adds operator corroboration while preserving the earlier dated sponsor disclosure and its historical outcome limits.\", \"business_problem\": \"\", \"workflow\": \"Sentinel Peak identifies machine-learning and AI analysis of reservoir data for real-time steam management.\", \"technology_data\": \"Operator corroboration of the existing Quantum/Sentinel Peak case; no new control architecture described.\", \"integration_human_role\": \"Approval, override and fallback controls undisclosed.\", \"deployment_scope\": \"Operator describes use; undated webpage does not establish a new rollout.\", \"partners\": [\"Quantum Energy Partners (sponsor and digital-team co-developer)\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"The operator estimates 200,000 metric tons CO2e per year enabled through intelligent-oilfield technologies. Period, baseline and AI-only contribution are unspecified.\", \"period\": \"Undated operator estimate; not a new measured 2026 result.\", \"attribution\": \"Operator estimate, not independently validated. Sponsor co-development comes from the existing dated dossier, not this page.\", \"source_ids\": [\"SW5-016\"]}], \"economics\": \"No project spend, monetized gas savings or payback disclosed.\", \"timeline\": [{\"date\": \"Undated\", \"event\": \"Current page accessed 1 October 2026; publication and measurement dates undisclosed.\", \"source_ids\": [\"SW5-016\"]}], \"limitations\": [\"Do not add this estimate to the baseline’s historical 2023 figures or treat either as AI-only net savings.\"], \"assessment\": \"Adds operator corroboration while preserving the earlier dated sponsor disclosure and its historical outcome limits.\", \"sources\": [{\"title\": \"Environment — Sentinel Peak Resources\", \"url\": \"https://sentinelpeakresources.com/sustainability/environment/\", \"date\": \"undated; accessed 2026-10-01\", \"type\": \"Operator environment disclosure\", \"locator\": \"Step 1: Reduced Steam, lines 15–26\", \"review_status\": \"Full original HTML page opened and read. It names Sentinel Peak's Intelligent Oilfield as an ML/AI-powered data system for real-time steam-management optimization and estimates 200,000 MT CO2e/year enabled through intelligent oilfield technologies. The page gives no publication date, calculation, baseline, or AI-only attribution split.\", \"publisher\": \"Sentinel Peak Resources\", \"id\": \"SW5-016\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-016\"], \"field_sources\": {\"workflow\": [\"SW5-016\"], \"technology_data\": [\"SW5-016\"], \"deployment_scope\": [\"SW5-016\"], \"practical_relevance\": [\"SW5-016\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator estimate, not independently validated. Sponsor co-development comes from the existing dated dossier, not this page.\"}", "Research_status": "Supplemental source added · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "California heavy-oil steamflood assets", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "An earlier audit logged the URL inaccessible; the later source review downloaded the full PDF and read the relevant page.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-215", "Company": "Occidental", "Company_primary": "Occidental", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI-assisted prioritization of lease-operator visits in routeless operations", "Extracted_finding": "Automation, field sensors and AI prioritize lease-operator routes; management reports routeless operations across part of onshore production.", "Stage_at_disclosure": "Management reports approximately 40% of onshore production had transitioned to routeless operations.", "Reported_metric": "Approximately 40% of onshore production covered; this is a production-share denominator, not a share of wells or personnel.", "Interpretation_limit": "No route-level time, cost, failure-rate or production-uplift result. The contribution of AI cannot be separated from sensors and automation.", "Source_IDs": "SW2-003", "Source_date": "2025-08-07", "Source_type": "Issuer-hosted earnings-call transcript", "Source_locator": "Page 3, Vicki Hollub prepared remarks", "Source_URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy2q25transcript.pdf", "Evidence_status": "Full 17-page company-hosted transcript reviewed.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "7 August 2025 earnings call reports existing operating coverage.", "Context": "A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.", "Implementation": "CEO described the integrated operating model and production coverage; model, data pipeline, dispatch logic and human override arrangements are not specified.", "Assessment": "A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-153"], "Disclosures": [{"ID": "SW2-D-AI-215", "Initiative_ID": "AI-215", "Title": "AI-assisted prioritization of lease-operator visits in routeless operations", "Date": "2025-08-07", "Finding": "Automation, field sensors and AI prioritize lease-operator routes; management reports routeless operations across part of onshore production.", "Metric": "Approximately 40% of onshore production covered; this is a production-share denominator, not a share of wells or personnel.", "Source_IDs": "SW2-003", "Evidence": "Full 17-page company-hosted transcript reviewed.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2025-08-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-215"], "Research_search": "{\"key\": \"swarm2-ai-215\", \"packet_id\": \"swarm2-ai-215\", \"record_ids\": [\"AI-215\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-153\"], \"company\": \"Occidental\", \"title\": \"AI-assisted prioritization of lease-operator visits in routeless operations\", \"department\": \"Production & field operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Management reports approximately 40% of onshore production had transitioned to routeless operations.\", \"practical_relevance\": \"A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.\", \"business_problem\": \"\", \"workflow\": \"Automation, field sensors and AI prioritize lease-operator routes; management reports routeless operations across part of onshore production.\", \"technology_data\": \"CEO described the integrated operating model and production coverage; model, data pipeline, dispatch logic and human override arrangements are not specified.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Management reports approximately 40% of onshore production had transitioned to routeless operations.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Approximately 40% of onshore production covered; this is a production-share denominator, not a share of wells or personnel.\", \"period\": \"Q2 2025 earnings call\", \"attribution\": \"CEO-reported operating scope; no independently measured outcome.\", \"source_ids\": [\"SW2-003\"]}], \"economics\": \"Adjacent $150 million expected 2025 onshore cost savings cover broader initiatives and cannot be assigned to this AI workflow.\", \"timeline\": [{\"date\": \"2025-08-07\", \"event\": \"7 August 2025 earnings call reports existing operating coverage.\", \"source_ids\": [\"SW2-003\"]}], \"limitations\": [\"No route-level time, cost, failure-rate or production-uplift result. The contribution of AI cannot be separated from sensors and automation.\"], \"assessment\": \"A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.\", \"sources\": [{\"title\": \"Occidental Q2 2025 Earnings Call Transcript\", \"url\": \"https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy2q25transcript.pdf\", \"date\": \"2025-08-07\", \"type\": \"Issuer-hosted earnings-call transcript\", \"locator\": \"Page 3, Vicki Hollub prepared remarks\", \"review_status\": \"Full 17-page company-hosted transcript reviewed.\", \"publisher\": \"Occidental\", \"id\": \"SW2-003\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-003\"], \"field_sources\": {\"workflow\": [\"SW2-003\"], \"technology_data\": [\"SW2-003\"], \"deployment_scope\": [\"SW2-003\"], \"practical_relevance\": [\"SW2-003\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"CEO-reported operating scope; no independently measured outcome.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "US onshore operations", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "~40% of onshore production", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full 17-page company-hosted transcript reviewed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-216", "Company": "ADNOC Gas", "Company_primary": "ADNOC Gas", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Real-time AI process optimization across gas plants", "Extracted_finding": "ADNOC Gas reports AI-driven process optimization across multiple plants and >$20m delivered to date; its FY2025 results slide separately places $50m realized savings in the AI-driven production-optimization pillar alongside closed-loop RTO deployed in Ruwais. The slide does not isolate RTO value or reconcile the claims.", "Stage_at_disclosure": "Multiple plant optimization described as deployed; closed-loop RTO at Ruwais listed among FY2025 achievements.", "Reported_metric": "Two separate issuer disclosures: >$20m delivered to date across multiple plants (period/value components unstated); $50m savings realized listed under FY2025 AI-driven production-optimization achievements (period, baseline and RTO-only allocation unstated). Do not add.", "Interpretation_limit": "$50m is pillar-level, not explicitly RTO-only, and gross/net, period, baseline and measurement method are unspecified. Existing >$20m claim has a different/unclear cumulative perimeter. The results deck disclaimer says information is background-only and unverified.", "Source_IDs": "SW2-005; OCT5-S047", "Source_date": "2026-02-09", "Source_type": "Operator results presentation", "Source_locator": "Slide 7 / PDF p. 6 (AI-driven production optimization row); PDF p. 2 disclaimer", "Source_URL": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "Evidence_status": "Relevant official presentation text/layout reviewed on 2026-10-06.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "Context": "A gas-processing workflow with a monetary result and a stated operating deployment, separate from ADNOC’s upstream ENERGYai program.", "Implementation": "The report confirms multi-plant deployment but gives no model architecture, plant list, control interface or operator-approval process.", "Assessment": "A gas-processing workflow with a monetary result and a stated operating deployment, separate from ADNOC’s upstream ENERGYai program.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-216", "Initiative_ID": "AI-216", "Title": "Real-time AI process optimization across gas plants", "Date": "2026", "Finding": "AI-driven optimizers allocate feed gas, reduce energy consumption and improve product recovery across multiple plants.", "Metric": "More than US$20 million in value delivered to date; value composition and cumulative period unstated.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "OCT5-AI-216-1", "Initiative_ID": "AI-216", "Title": "ADNOC Gas Q4/FY2025 Results Presentation", "Date": "2026-02-09", "Finding": "The FY2025 achievements table places “Closed Loop RTO deployed in Ruwais” and “$50m savings realized” within the AI-driven production-optimization row (separate 2026–30 opportunity: up to $300m). Existing AI-216 cites the 2025 Integrated Report claim of more than $20m value delivered to date across multiple plants. The presentation does not isolate RTO’s contribution or reconcile the two scopes.", "Metric": "Company-reported row-level $50m savings realization; period, baseline and calculation perimeter unstated. Separate existing claim: >$20m value to date across multiple plants; do not add or replace without reconciliation.", "Source_IDs": "OCT5-S047", "Source_URL": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Latest_disclosure": "2026-02-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai-216", "swarm2-ai-216"], "Research_search": "{\"packet_id\": \"oct6-executive-ai-216\", \"review_date\": \"2026-10-06\", \"title\": \"ADNOC Gas Q4/FY2025 Results Presentation\", \"publisher\": \"ADNOC Gas\", \"source_title\": \"ADNOC Gas Q4/FY2025 Results Presentation\", \"source_date\": \"2026-02-09\", \"source_type\": \"Operator results presentation\", \"source_locator\": \"Slide 7 / PDF p. 6; disclaimer PDF p. 2\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"ADNOC Gas describes AI process optimization across plants and closed-loop real-time optimization at Ruwais.\", \"technology_data\": \"The FY2025 slide names closed-loop RTO within the production-optimization pillar. It does not specify the model architecture, control limits or how much of the pillar’s savings came from RTO.\", \"deployment_scope\": \"Ruwais closed-loop RTO deployment listed; $50m is pillar-level, not expressly RTO-only.\", \"practical_relevance\": \"Production optimization has a reported financial result to investigate. Ask ADNOC Gas to reconcile the plant-program figure with the wider pillar before comparing savings.\", \"business_problem\": \"Improving plant operating decisions and production performance.\", \"economics\": \"No costs, baseline or net/gross value bridge. Presentation states information is unverified.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Do not sum separate savings claims or allocate the full $50m to RTO; row-level scope and period are unclear.\"], \"timeline_event\": \"FY2025 achievements slide published 2026-02-09; exact savings period not stated.\", \"timeline_date\": \"2026-02-09\", \"source_ids\": [\"OCT5-S047\"], \"stage\": \"Multiple plant optimization described as deployed; closed-loop RTO at Ruwais listed among FY2025 achievements.\", \"key\": \"oct6-executive-ai-216\", \"record_ids\": [\"AI-216\"], \"company\": \"ADNOC Gas\", \"category\": \"Internal adoption\", \"department\": \"Refining, LNG & process operations\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S047\"], \"workflow\": [\"OCT5-S047\"], \"technology_data\": [\"OCT5-S047\"], \"deployment_scope\": [\"OCT5-S047\"], \"practical_relevance\": [\"OCT5-S047\"], \"economics\": [\"OCT5-S047\"], \"integration_human_role\": [\"OCT5-S047\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \"Issuer reports $50m savings realized on row; period/baseline/method unknown. Preserve separate >$20m disclosure.\", \"period\": \"2026-02-09\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S047\"]}], \"timeline\": [{\"date\": \"2026-02-09\", \"event\": \"FY2025 achievements slide published 2026-02-09; exact savings period not stated.\", \"source_ids\": [\"OCT5-S047\"]}], \"sources\": [{\"id\": \"OCT5-S047\", \"title\": \"ADNOC Gas Q4/FY2025 Results Presentation\", \"url\": \"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\", \"date\": \"2026-02-09\", \"type\": \"Operator results presentation\", \"publisher\": \"ADNOC Gas\", \"locator\": \"Slide 7 / PDF p. 6; disclaimer PDF p. 2\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}]}", "Research_status": "Updated · second September review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Gas-processing facilities; individual sites not named", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Preserve both operator disclosures separately: FY2025 slide places $50m under production-optimization pillar; existing >$20m-to-date across plants not reconciled. Neither sum nor ascribe all $50m to RTO.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Relevant official presentation text/layout reviewed on 2026-10-06.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-217", "Company": "ADNOC Gas", "Company_primary": "ADNOC Gas", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Neuron 5 asset-integrity and process-optimization platform", "Extracted_finding": "A centralized AI platform combines predictive analytics and first-principles models to detect anomalies and support gas-plant reliability and process optimization.", "Stage_at_disclosure": "Company describes an operating platform; deployment coverage and go-live dates undisclosed.", "Reported_metric": "No Neuron 5-specific numerical result disclosed in the reviewed report section.", "Interpretation_limit": "No quantified reduction in shutdowns or project economics. Neuron 5 appeared in an earlier unreviewed ADNOC portfolio disclosure; this is a dedicated, operator-attributed profile. Historical CPAD lineage added to existing Neuron 5 dossier, not a new counted initiative. CPAD amounts are not Neuron 5 results, annual savings, net ROI or additive benefit.", "Source_IDs": "SW2-005; OCT5-S042", "Source_date": "2026", "Source_type": "Issuer integrated report", "Source_locator": "Printed pages 10–11, AI Highlights; PDF physical page 7", "Source_URL": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "Evidence_status": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "Context": "Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.", "Implementation": "The report names Neuron 5 and the modeling approach; equipment coverage, interfaces and response controls are not described. The earlier CPAD system, launched in 2019, monitored rotating-equipment health. Its 2024 report frames Neuron 5 as an evolution; predecessor benefit claims are kept in source history and are not Neuron 5 results.", "Assessment": "Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Refining, LNG & process operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-040"], "Disclosures": [{"ID": "SW2-D-AI-217", "Initiative_ID": "AI-217", "Title": "Neuron 5 asset-integrity and process-optimization platform", "Date": "2026", "Finding": "A centralized AI platform combines predictive analytics and first-principles models to detect anomalies and support gas-plant reliability and process optimization.", "Metric": "No Neuron 5-specific numerical result disclosed in the reviewed report section.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "OCT5-AI-217-1", "Initiative_ID": "AI-217", "Title": "CPAD predecessor: historical deployment and benefit boundaries", "Date": "2025 (2024 report; exact publication day not stated)", "Finding": "The issuer reports US$29m saved through performance/reliability improvement and more than US$90m avoided costs through early prediction. GenAI integration and expansion to static equipment are roadmap work.", "Metric": "US$29m reported savings and >US$90m avoided costs. These have no reconciled overlap or common measurement period.", "Source_IDs": "OCT5-S042", "Source_URL": "https://www.adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas-integrated-report-2024-english.ashx", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-217", "oct6-three-rounds-ai-217"], "Research_search": "{\"key\": \"swarm2-ai-217\", \"packet_id\": \"swarm2-ai-217\", \"record_ids\": [\"AI-217\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-040\"], \"company\": \"ADNOC Gas\", \"title\": \"Neuron 5 asset-integrity and process-optimization platform\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"Refining, LNG & process operations\"], \"category\": \"Internal adoption\", \"stage\": \"Company describes an operating platform; deployment coverage and go-live dates undisclosed.\", \"practical_relevance\": \"Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.\", \"business_problem\": \"\", \"workflow\": \"A centralized AI platform combines predictive analytics and first-principles models to detect anomalies and support gas-plant reliability and process optimization.\", \"technology_data\": \"The report names Neuron 5 and the modeling approach; equipment coverage, interfaces and response controls are not described.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Company describes an operating platform; deployment coverage and go-live dates undisclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No Neuron 5-specific numerical result disclosed in the reviewed report section.\", \"period\": \"FY2025 report; underlying operating period not specified\", \"attribution\": \"Issuer-reported disclosure, not independently validated operating performance.\", \"source_ids\": [\"SW2-005\"]}], \"economics\": \"The adjacent RTO US$20 million claim is not assigned to Neuron 5.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.\", \"source_ids\": [\"SW2-005\"]}], \"limitations\": [\"No quantified reduction in shutdowns or project economics. Neuron 5 appeared in an earlier unreviewed ADNOC portfolio disclosure; this is a dedicated, operator-attributed profile.\"], \"assessment\": \"Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.\", \"sources\": [{\"title\": \"ADNOC Gas Integrated Report 2025\", \"url\": \"https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx\", \"date\": \"2026\", \"type\": \"Issuer integrated report\", \"locator\": \"Printed pages 10–11, AI Highlights; PDF physical page 7\", \"review_status\": \"Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.\", \"publisher\": \"ADNOC Gas\", \"id\": \"SW2-005\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-005\"], \"field_sources\": {\"workflow\": [\"SW2-005\"], \"technology_data\": [\"SW2-005\"], \"deployment_scope\": [\"SW2-005\"], \"practical_relevance\": [\"SW2-005\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Issuer-reported disclosure, not independently validated operating performance.\"}", "Research_status": "Updated · second September review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Gas-processing facilities; individual sites not named", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-218", "Company": "ADNOC Gas", "Company_primary": "ADNOC Gas", "Segment": "Refining, LNG & processing", "Theme": "Operational AI", "Initiative": "Cantilever AI-enabled inspection management", "Extracted_finding": "ADNOC Gas announced a three-year AIQ/Gecko Robotics agreement using Cantilever to digitize inspection-to-decision. Year 1 is joint deployment/data modeling, followed by scaled access/continuous optimization in years 2–3. Issuer forecasts >$300m maintenance and inspection cost savings over five years. ADNOC Gas says it used Gecko robots in 2023; a later deck says Cantilever/Gecko in all assets but gives no denominator/completion definition.", "Stage_at_disclosure": "Cantilever deployment described in current operator reporting; the three-year AIQ/Gecko collaboration is phased, with five-year savings forecast.", "Reported_metric": "Issuer forecast: >$300m maintenance and inspection cost savings over five years under a three-year phased agreement; no realized savings reported.", "Interpretation_limit": "Forward-looking issuer claim. No baseline, contract/implementation/ongoing cost, site denominator, realized result or measurement method. Separate from the AI-driven production-optimization pillar’s up-to-$300m 2026–30 opportunity. Do not call forecast net ROI.", "Source_IDs": "SW2-005; OCT5-S048; OCT5-S047", "Source_date": "2025-11-02", "Source_type": "Operator press release", "Source_locator": "Paragraphs 107–125; phased plan at paragraph 119; forward-looking caveat paragraphs 123–125", "Source_URL": "https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq", "Evidence_status": "Relevant original operator release independently reviewed on 2026-10-06.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "Context": "Adds a named inspection workflow with explicit AI attribution and a company deployment statement.", "Implementation": "ADNOC Gas names Cantilever as a deployed inspection-management platform. The report does not distinguish AI-model functions from drone hardware and other inspection methods.", "Assessment": "Adds a named inspection workflow with explicit AI attribution and a company deployment statement.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-218", "Initiative_ID": "AI-218", "Title": "Cantilever AI-enabled inspection management", "Date": "2026", "Finding": "An AI-enabled inspection platform supports non-intrusive inspections, forecasting and drone-enabled monitoring.", "Metric": "Safety, accuracy and efficiency improvements described without numerical results.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "OCT5-AI-218-1", "Initiative_ID": "AI-218", "Title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "Date": "2025-11-02", "Finding": "ADNOC Gas announced a three-year AI/robotics collaboration with AIQ and Gecko Robotics using Cantilever to digitize inspection-to-decision. Year 1 is joint deployment/data modeling, years 2–3 scaled access and continuous optimization. The issuer forecasts over $300m maintenance and inspection cost savings over five years. ADNOC Gas says it had already used Gecko robots in 2023; the new element is the AI/Cantilever program. A 2026 results slide later says “Cantiliver GECKO - in all assets,” without asset denominator or completion definition.", "Metric": "Issuer forecast: >$300m maintenance/inspection cost savings over five years under a three-year phased program. No realized savings reported.", "Source_IDs": "OCT5-S048", "Source_URL": "https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai-218", "swarm2-ai-218"], "Research_search": "{\"packet_id\": \"oct6-executive-ai-218\", \"review_date\": \"2026-10-06\", \"title\": \"ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance\", \"publisher\": \"ADNOC Gas\", \"source_title\": \"ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance\", \"source_date\": \"2025-11-02\", \"source_type\": \"Operator press release\", \"source_locator\": \"Paragraphs 107–125\", \"source_review\": \"Relevant original primary source reviewed on 2026-10-06.\", \"workflow\": \"Gecko inspection data and Cantilever are intended to connect inspection, analysis and maintenance decisions, with AIQ supplying predictive models.\", \"technology_data\": \"The three-year program phases deployment and data modeling before wider access and continuous optimization. The company’s later results describe Cantilever across assets without a completion denominator.\", \"deployment_scope\": \"Phased agreement; later slide says Cantilever/Gecko “in all assets,” denominator/completion undefined.\", \"practical_relevance\": \"This is a practical example of connecting inspection data to maintenance planning. Ask which assets and decisions are live, and how responsibilities are divided between the operator and providers.\", \"business_problem\": \"Turning inspection findings into maintenance decisions.\", \"economics\": \"No baseline, contract/system costs, denominator, realized results or net return disclosed.\", \"result_label\": \"Issuer-reported disclosure and evidence boundary\", \"attribution\": \"Operator account\", \"verification\": \"verified primary\", \"limitations\": [\"Forward-looking cost-savings claim, separate from production-optimization pillar; not net ROI.\"], \"timeline_event\": \"Agreement announced 2025-11-02; year-one deployment/modeling, then years 2–3 scaling.\", \"timeline_date\": \"2025-11-02\", \"source_ids\": [\"OCT5-S048\", \"OCT5-S047\"], \"supplementary_source_urls\": [\"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\"], \"field_source_urls\": {\"deployment_scope\": [\"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\"], \"technology_data\": [\"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\"], \"workflow\": [\"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\"]}, \"stage\": \"Cantilever deployment described in current operator reporting; the three-year AIQ/Gecko collaboration is phased, with five-year savings forecast.\", \"key\": \"oct6-executive-ai-218\", \"record_ids\": [\"AI-218\"], \"company\": \"ADNOC Gas\", \"category\": \"Internal adoption\", \"department\": \"Maintenance & reliability\", \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S048\"], \"workflow\": [\"OCT5-S048\", \"OCT5-S047\"], \"technology_data\": [\"OCT5-S048\", \"OCT5-S047\"], \"deployment_scope\": [\"OCT5-S048\", \"OCT5-S047\"], \"practical_relevance\": [\"OCT5-S048\"], \"economics\": [\"OCT5-S048\"], \"integration_human_role\": [\"OCT5-S048\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and evidence boundary\", \"value\": \"Issuer forecasts >$300m maintenance/inspection cost savings over five years; no realized value reported.\", \"period\": \"2025-11-02\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S048\"]}], \"timeline\": [{\"date\": \"2025-11-02\", \"event\": \"Agreement announced 2025-11-02; year-one deployment/modeling, then years 2–3 scaling.\", \"source_ids\": [\"OCT5-S048\"]}], \"sources\": [{\"id\": \"OCT5-S048\", \"title\": \"ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance\", \"url\": \"https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq\", \"date\": \"2025-11-02\", \"type\": \"Operator press release\", \"publisher\": \"ADNOC Gas\", \"locator\": \"Paragraphs 107–125\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}, {\"id\": \"OCT5-S047\", \"title\": \"ADNOC Gas Q4/FY2025 Results Presentation\", \"url\": \"https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx\", \"date\": \"2026-02-09\", \"type\": \"Operator results presentation\", \"publisher\": \"ADNOC Gas\", \"locator\": \"Slide 7 / PDF p. 6; disclaimer PDF p. 2\", \"review_status\": \"Relevant original primary source reviewed on 2026-10-06.\"}]}", "Research_status": "Updated · second September review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Gas-processing facilities; individual sites not named", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Forecast >$300m over five years from three-year phased AI/robotics maintenance program; not achieved savings/net ROI. Keep distinct from production-optimization pillar.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Relevant original operator release independently reviewed on 2026-10-06.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-219", "Company": "Petrobras", "Company_primary": "Petrobras", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Machine learning for well configurations integrated with simulation", "Extracted_finding": "Machine-learning well-configuration analysis is integrated with simulation systems.", "Stage_at_disclosure": "Presentation reports value already captured; deployment extent unspecified.", "Reported_metric": "R$350 million captured; period and definition of captured value unstated.", "Interpretation_limit": "No baseline, project duration, well count or value-calculation method. Separate from Petrobras flare control and monitoring.", "Source_IDs": "SW2-006", "Source_date": "2025-11", "Source_type": "Issuer investor presentation", "Source_locator": "Slide 158, Accelerate value generation through digital technologies", "Source_URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Evidence_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "Context": "Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.", "Implementation": "Slide identifies the decision workflow and its connection to simulation; model, well population and decision-approval controls are not disclosed.", "Assessment": "Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-219", "Initiative_ID": "AI-219", "Title": "Machine learning for well configurations integrated with simulation", "Date": "2025-11", "Finding": "Machine-learning well-configuration analysis is integrated with simulation systems.", "Metric": "R$350 million captured; period and definition of captured value unstated.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2025-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-219"], "Research_search": "{\"key\": \"swarm2-ai-219\", \"packet_id\": \"swarm2-ai-219\", \"record_ids\": [\"AI-219\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Petrobras\", \"title\": \"Machine learning for well configurations integrated with simulation\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Presentation reports value already captured; deployment extent unspecified.\", \"practical_relevance\": \"Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.\", \"business_problem\": \"\", \"workflow\": \"Machine-learning well-configuration analysis is integrated with simulation systems.\", \"technology_data\": \"Slide identifies the decision workflow and its connection to simulation; model, well population and decision-approval controls are not disclosed.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Presentation reports value already captured; deployment extent unspecified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"R$350 million captured; period and definition of captured value unstated.\", \"period\": \"November 2025 disclosure; result periods not specified except where annualized\", \"attribution\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\", \"source_ids\": [\"SW2-006\"]}], \"economics\": \"Reported captured value; no cash reconciliation, implementation cost or net return. Not an annual figure.\", \"timeline\": [{\"date\": \"2025-11\", \"event\": \"November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.\", \"source_ids\": [\"SW2-006\"]}], \"limitations\": [\"No baseline, project duration, well count or value-calculation method. Separate from Petrobras flare control and monitoring.\"], \"assessment\": \"Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.\", \"sources\": [{\"title\": \"Petrobras Business Plan 2026–2030 investor presentation\", \"url\": \"https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0\", \"date\": \"2025-11\", \"type\": \"Issuer investor presentation\", \"locator\": \"Slide 158, Accelerate value generation through digital technologies\", \"review_status\": \"Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.\", \"publisher\": \"Petrobras\", \"id\": \"SW2-006\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-006\"], \"field_sources\": {\"workflow\": [\"SW2-006\"], \"technology_data\": [\"SW2-006\"], \"deployment_scope\": [\"SW2-006\"], \"practical_relevance\": [\"SW2-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Assets not specified on slide", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-220", "Company": "Petrobras", "Company_primary": "Petrobras", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI predictive monitoring and integrated operations control rooms", "Extracted_finding": "AI enhances predictive monitoring and integrated operations control rooms (COI).", "Stage_at_disclosure": "Annualized avoided-loss result reported; asset coverage and persistence unspecified.", "Reported_metric": "8.3 million BOE/year of losses avoided, described as approximately three days of production per year.", "Interpretation_limit": "No counterfactual, event count, baseline or implementation cost. Do not monetize without entitlement, pricing and realization evidence.", "Source_IDs": "SW2-006", "Source_date": "2025-11", "Source_type": "Issuer investor presentation", "Source_locator": "Slide 158, Accelerate value generation through digital technologies", "Source_URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Evidence_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "Context": "Adds an operating-availability case while keeping physical production measures separate from financial returns.", "Implementation": "Presentation connects AI-supported monitoring to avoided production losses; assets, event detection methods and intervention process are not described.", "Assessment": "Adds an operating-availability case while keeping physical production measures separate from financial returns.", "Primary_department": "Production & field operations", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-220", "Initiative_ID": "AI-220", "Title": "AI predictive monitoring and integrated operations control rooms", "Date": "2025-11", "Finding": "AI enhances predictive monitoring and integrated operations control rooms (COI).", "Metric": "8.3 million BOE/year of losses avoided, described as approximately three days of production per year.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2025-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-220"], "Research_search": "{\"key\": \"swarm2-ai-220\", \"packet_id\": \"swarm2-ai-220\", \"record_ids\": [\"AI-220\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Petrobras\", \"title\": \"AI predictive monitoring and integrated operations control rooms\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"Maintenance & reliability\"], \"category\": \"Internal adoption\", \"stage\": \"Annualized avoided-loss result reported; asset coverage and persistence unspecified.\", \"practical_relevance\": \"Adds an operating-availability case while keeping physical production measures separate from financial returns.\", \"business_problem\": \"\", \"workflow\": \"AI enhances predictive monitoring and integrated operations control rooms (COI).\", \"technology_data\": \"Presentation connects AI-supported monitoring to avoided production losses; assets, event detection methods and intervention process are not described.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Annualized avoided-loss result reported; asset coverage and persistence unspecified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"8.3 million BOE/year of losses avoided, described as approximately three days of production per year.\", \"period\": \"November 2025 disclosure; result periods not specified except where annualized\", \"attribution\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\", \"source_ids\": [\"SW2-006\"]}], \"economics\": \"Avoided production loss is an availability measure; it is not disclosed revenue, cash savings or incremental sales.\", \"timeline\": [{\"date\": \"2025-11\", \"event\": \"November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.\", \"source_ids\": [\"SW2-006\"]}], \"limitations\": [\"No counterfactual, event count, baseline or implementation cost. Do not monetize without entitlement, pricing and realization evidence.\"], \"assessment\": \"Adds an operating-availability case while keeping physical production measures separate from financial returns.\", \"sources\": [{\"title\": \"Petrobras Business Plan 2026–2030 investor presentation\", \"url\": \"https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0\", \"date\": \"2025-11\", \"type\": \"Issuer investor presentation\", \"locator\": \"Slide 158, Accelerate value generation through digital technologies\", \"review_status\": \"Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.\", \"publisher\": \"Petrobras\", \"id\": \"SW2-006\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-006\"], \"field_sources\": {\"workflow\": [\"SW2-006\"], \"technology_data\": [\"SW2-006\"], \"deployment_scope\": [\"SW2-006\"], \"practical_relevance\": [\"SW2-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Assets not specified on slide", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-221", "Company": "Petrobras", "Company_primary": "Petrobras", "Segment": "Trading & shipping", "Theme": "Operational AI", "Initiative": "Machine-learning revenue forecasting for the domestic market", "Extracted_finding": "Advanced machine-learning algorithms forecast domestic-market revenue.", "Stage_at_disclosure": "Improved model performance reported; operating scope undisclosed.", "Reported_metric": "51% greater forecast accuracy; accuracy measure and comparison baseline unspecified.", "Interpretation_limit": "The 51% is not an absolute accuracy rate or a percentage-point change. No test population, horizon, baseline or validation method.", "Source_IDs": "SW2-006", "Source_date": "2025-11", "Source_type": "Issuer investor presentation", "Source_locator": "Slide 158, Accelerate value generation through digital technologies", "Source_URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Evidence_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "PE_relationship": null, "Coverage_bucket": "Second targeted gap review", "Time_scope": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "Context": "Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.", "Implementation": "The presentation identifies the forecasting objective; features, model type, horizon and relationship to commercial decisions are unspecified.", "Assessment": "Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.", "Primary_department": "Trading & marketing", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Second September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW2-D-AI-221", "Initiative_ID": "AI-221", "Title": "Machine-learning revenue forecasting for the domestic market", "Date": "2025-11", "Finding": "Advanced machine-learning algorithms forecast domestic-market revenue.", "Metric": "51% greater forecast accuracy; accuracy measure and comparison baseline unspecified.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}], "Latest_disclosure": "2025-11", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm2-ai-221"], "Research_search": "{\"key\": \"swarm2-ai-221\", \"packet_id\": \"swarm2-ai-221\", \"record_ids\": [\"AI-221\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Petrobras\", \"title\": \"Machine-learning revenue forecasting for the domestic market\", \"department\": \"Trading & marketing\", \"secondary_departments\": [\"Finance, investment & strategy\"], \"category\": \"Internal adoption\", \"stage\": \"Improved model performance reported; operating scope undisclosed.\", \"practical_relevance\": \"Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.\", \"business_problem\": \"\", \"workflow\": \"Advanced machine-learning algorithms forecast domestic-market revenue.\", \"technology_data\": \"The presentation identifies the forecasting objective; features, model type, horizon and relationship to commercial decisions are unspecified.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Improved model performance reported; operating scope undisclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"51% greater forecast accuracy; accuracy measure and comparison baseline unspecified.\", \"period\": \"November 2025 disclosure; result periods not specified except where annualized\", \"attribution\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\", \"source_ids\": [\"SW2-006\"]}], \"economics\": \"Forecast accuracy is a model-performance measure, not measured revenue growth or savings.\", \"timeline\": [{\"date\": \"2025-11\", \"event\": \"November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.\", \"source_ids\": [\"SW2-006\"]}], \"limitations\": [\"The 51% is not an absolute accuracy rate or a percentage-point change. No test population, horizon, baseline or validation method.\"], \"assessment\": \"Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.\", \"sources\": [{\"title\": \"Petrobras Business Plan 2026–2030 investor presentation\", \"url\": \"https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0\", \"date\": \"2025-11\", \"type\": \"Issuer investor presentation\", \"locator\": \"Slide 158, Accelerate value generation through digital technologies\", \"review_status\": \"Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.\", \"publisher\": \"Petrobras\", \"id\": \"SW2-006\", \"accessed\": \"2026-09-29\"}], \"source_ids\": [\"SW2-006\"], \"field_sources\": {\"workflow\": [\"SW2-006\"], \"technology_data\": [\"SW2-006\"], \"deployment_scope\": [\"SW2-006\"], \"practical_relevance\": [\"SW2-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Second gap review · September 2026\", \"report_date\": \"2026-09-29\", \"evidence_note\": \"Issuer-reported presentation metric; calculation method and independent assurance undisclosed.\"}", "Research_status": "Updated · second September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-29", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Assets not specified on slide", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-222", "Company": "Enbridge", "Company_primary": "Enbridge", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "Energy Optimizer for liquids-pipeline power decisions", "Extracted_finding": "AI supplies operational insights to control-center staff choosing how to transport liquids with less power.", "Stage_at_disclosure": "Company describes an operating tool; deployment footprint undisclosed.", "Reported_metric": "No quantified energy, cost or emissions result.", "Interpretation_limit": "No asset count, baseline or measured outcome. Enterprise cloud-migration statistics do not measure this tool’s deployment.", "Source_IDs": "SW3-001", "Source_date": "2024-10-07", "Source_type": "Operator press release", "Source_locator": "AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "Source_URL": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "Evidence_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.", "Context": "Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.", "Implementation": "The release identifies Azure Machine Learning and the Energy Optimizer tool; control-center operators make the decisions.", "Assessment": "Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.", "Primary_department": "Production & field operations", "Secondary_departments": ["HSE & emissions"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW3-D-AI-222", "Initiative_ID": "AI-222", "Title": "Energy Optimizer for liquids-pipeline power decisions", "Date": "2024-10-07", "Finding": "AI supplies operational insights to control-center staff choosing how to transport liquids with less power.", "Metric": "No quantified energy, cost or emissions result.", "Source_IDs": "SW3-001", "Evidence": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "Origin": "Third September research swarm"}], "Latest_disclosure": "2024-10-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-222"], "Research_search": "{\"key\": \"swarm3-ai-222\", \"packet_id\": \"swarm3-ai-222\", \"record_ids\": [\"AI-222\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Enbridge\", \"title\": \"Energy Optimizer for liquids-pipeline power decisions\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"HSE & emissions\"], \"category\": \"Internal adoption\", \"stage\": \"Company describes an operating tool; deployment footprint undisclosed.\", \"practical_relevance\": \"Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.\", \"business_problem\": \"\", \"workflow\": \"AI supplies operational insights to control-center staff choosing how to transport liquids with less power.\", \"technology_data\": \"The release identifies Azure Machine Learning and the Energy Optimizer tool; control-center operators make the decisions.\", \"integration_human_role\": \"Decision support for control-center operators; autonomous dispatch is not established.\", \"deployment_scope\": \"Company describes an operating tool; deployment footprint undisclosed.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No quantified energy, cost or emissions result.\", \"period\": \"October 2024 disclosure; operating measurement period unstated\", \"attribution\": \"Operator disclosure; performance not independently validated.\", \"source_ids\": [\"SW3-001\"]}], \"economics\": \"Potential cost and emissions benefits described; project cost and payback absent.\", \"timeline\": [{\"date\": \"2024-10-07\", \"event\": \"7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.\", \"source_ids\": [\"SW3-001\"]}], \"limitations\": [\"No asset count, baseline or measured outcome. Enterprise cloud-migration statistics do not measure this tool’s deployment.\"], \"assessment\": \"Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.\", \"sources\": [{\"title\": \"Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization\", \"url\": \"https://www.enbridge.com/media-center/news/details?id=123831&lang=en\", \"date\": \"2024-10-07\", \"type\": \"Operator press release\", \"locator\": \"AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring\", \"review_status\": \"Original company release read in full; workflow-specific claims retained without assigning general platform statistics.\", \"publisher\": \"Enbridge\", \"id\": \"SW3-001\", \"accessed\": \"2026-09-30\"}], \"source_ids\": [\"SW3-001\"], \"field_sources\": {\"workflow\": [\"SW3-001\"], \"technology_data\": [\"SW3-001\"], \"deployment_scope\": [\"SW3-001\"], \"practical_relevance\": [\"SW3-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Third gap review · September 2026\", \"report_date\": \"2026-09-30\", \"evidence_note\": \"Operator disclosure; performance not independently validated.\"}", "Research_status": "Updated · third September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Pipeline operations; individual assets not identified", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-223", "Company": "Enbridge", "Company_primary": "Enbridge", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "AI-assisted aerial right-of-way threat monitoring", "Extracted_finding": "AI helps review aerial pipeline-surveillance data for potentially damaging activity, hazardous conditions and loss of containment.", "Stage_at_disclosure": "Company says it uses the capability; monitored mileage and rollout dates undisclosed.", "Reported_metric": "Qualitative detection and response improvements only.", "Interpretation_limit": "No accuracy, false-positive rate or response-time baseline. Separate from in-line inspection and deformation-feature assessment.", "Source_IDs": "SW3-001", "Source_date": "2024-10-07", "Source_type": "Operator press release", "Source_locator": "AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "Source_URL": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "Evidence_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.", "Context": "Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.", "Implementation": "The operator describes faster issue review and threat response; model, sensor mix and escalation controls are not specified.", "Assessment": "Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.", "Primary_department": "HSE & emissions", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW3-D-AI-223", "Initiative_ID": "AI-223", "Title": "AI-assisted aerial right-of-way threat monitoring", "Date": "2024-10-07", "Finding": "AI helps review aerial pipeline-surveillance data for potentially damaging activity, hazardous conditions and loss of containment.", "Metric": "Qualitative detection and response improvements only.", "Source_IDs": "SW3-001", "Evidence": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "Origin": "Third September research swarm"}], "Latest_disclosure": "2024-10-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-223"], "Research_search": "{\"key\": \"swarm3-ai-223\", \"packet_id\": \"swarm3-ai-223\", \"record_ids\": [\"AI-223\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Enbridge\", \"title\": \"AI-assisted aerial right-of-way threat monitoring\", \"department\": \"HSE & emissions\", \"secondary_departments\": [\"Maintenance & reliability\"], \"category\": \"Internal adoption\", \"stage\": \"Company says it uses the capability; monitored mileage and rollout dates undisclosed.\", \"practical_relevance\": \"Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.\", \"business_problem\": \"\", \"workflow\": \"AI helps review aerial pipeline-surveillance data for potentially damaging activity, hazardous conditions and loss of containment.\", \"technology_data\": \"The operator describes faster issue review and threat response; model, sensor mix and escalation controls are not specified.\", \"integration_human_role\": \"Alert triage and response responsibilities are not described.\", \"deployment_scope\": \"Company says it uses the capability; monitored mileage and rollout dates undisclosed.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Qualitative detection and response improvements only.\", \"period\": \"October 2024 disclosure; operating measurement period unstated\", \"attribution\": \"Operator disclosure; performance not independently validated.\", \"source_ids\": [\"SW3-001\"]}], \"economics\": \"No quantified damage avoidance, inspection saving or payback.\", \"timeline\": [{\"date\": \"2024-10-07\", \"event\": \"7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.\", \"source_ids\": [\"SW3-001\"]}], \"limitations\": [\"No accuracy, false-positive rate or response-time baseline. Separate from in-line inspection and deformation-feature assessment.\"], \"assessment\": \"Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.\", \"sources\": [{\"title\": \"Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization\", \"url\": \"https://www.enbridge.com/media-center/news/details?id=123831&lang=en\", \"date\": \"2024-10-07\", \"type\": \"Operator press release\", \"locator\": \"AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring\", \"review_status\": \"Original company release read in full; workflow-specific claims retained without assigning general platform statistics.\", \"publisher\": \"Enbridge\", \"id\": \"SW3-001\", \"accessed\": \"2026-09-30\"}], \"source_ids\": [\"SW3-001\"], \"field_sources\": {\"workflow\": [\"SW3-001\"], \"technology_data\": [\"SW3-001\"], \"deployment_scope\": [\"SW3-001\"], \"practical_relevance\": [\"SW3-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Third gap review · September 2026\", \"report_date\": \"2026-09-30\", \"evidence_note\": \"Operator disclosure; performance not independently validated.\"}", "Research_status": "Updated · third September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Pipeline operations; individual assets not identified", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-224", "Company": "Williams", "Company_primary": "Williams", "Segment": "Cross-segment / enterprise", "Theme": "Operational AI", "Initiative": "Companywide use of Copilot and approved AI tools", "Extracted_finding": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.", "Stage_at_disclosure": "Employee-built alert workflow described in use", "Reported_metric": "More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.", "Interpretation_limit": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.", "Source_IDs": "SW3-002; WATCH14PM-WILLIAMS", "Source_date": "2026-09-09", "Source_type": "Vendor customer story", "Source_locator": "Bringing AI to the frontline first; Accelerating operations with Copilot", "Source_URL": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "Published 9 September 2026; the existing profile retains Williams’s June company-wide adoption disclosure. The detector’s start date and asset count are unspecified.", "Context": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.", "Implementation": "The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "Assessment": "Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Workforce & learning"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW3-D-AI-224", "Initiative_ID": "AI-224", "Title": "Companywide use of Copilot and approved AI tools", "Date": "2026-06-01", "Finding": "Employees use Copilot and other approved AI tools to reduce routine work and support decisions.", "Metric": "More than 90% employee adoption; active-use threshold and observation period unspecified.", "Source_IDs": "SW3-002", "Evidence": "Original operator article read in full; date from article URL; adoption definition limitation retained.", "Origin": "Third September research swarm"}, {"ID": "WATCH14PM-AI-224", "Initiative_ID": "AI-224", "Title": "A field technician turns equipment reports into turbine alerts", "Date": "2026-09-09", "Finding": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration. The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "Metric": "More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.", "Source_IDs": "WATCH14PM-WILLIAMS", "Source_URL": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share."}], "Latest_disclosure": "2026-09-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-224", "watch14pm-ai-224"], "Research_search": "{\"packet_id\": \"watch14pm-ai-224\", \"key\": \"watch14pm-ai-224\", \"record_ids\": [\"AI-224\"], \"company\": \"Williams\", \"title\": \"A field technician turns equipment reports into turbine alerts\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Employee-built alert workflow described in use\", \"business_problem\": \"A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.\", \"workflow\": \"The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.\", \"technology_data\": \"The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.\", \"deployment_scope\": \"Published 9 September 2026; the existing profile retains Williams’s June company-wide adoption disclosure. The detector’s start date and asset count are unspecified.\", \"integration_human_role\": \"Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.\", \"economics\": \"More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.\", \"practical_relevance\": \"Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?\", \"assessment\": \"Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?\", \"limitations\": [\"Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.\"], \"source_ids\": [\"WATCH14PM-WILLIAMS\"], \"sources\": [{\"id\": \"WATCH14PM-WILLIAMS\", \"title\": \"Williams transforms frontline work with Copilot, gaining 90% adoption\", \"url\": \"https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot\", \"date\": \"2026-09-09\", \"type\": \"Vendor customer story\", \"publisher\": \"Microsoft / named Williams speakers\", \"locator\": \"Bringing AI to the frontline first; Accelerating operations with Copilot\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-09-09\", \"event\": \"A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.\", \"source_ids\": [\"WATCH14PM-WILLIAMS\"]}], \"field_sources\": {\"business_problem\": [\"WATCH14PM-WILLIAMS\"], \"workflow\": [\"WATCH14PM-WILLIAMS\"], \"technology_data\": [\"WATCH14PM-WILLIAMS\"], \"deployment_scope\": [\"WATCH14PM-WILLIAMS\"], \"economics\": [\"WATCH14PM-WILLIAMS\"]}}", "Research_status": "Reviewed 14 October 2026 · evening watch", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Companywide", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "90%+ of employees", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original operator article read in full; date from article URL; adoption definition limitation retained.", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-14", "Research_origin": ["research/watch-2026-10-14-pm/README.md"]}, {"ID": "AI-225", "Company": "Oil and Natural Gas Corporation Limited (ONGC)", "Company_primary": "ONGC", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI/ML data-physics water-injection optimization across Heera assets", "Extracted_finding": "Use AI/ML data-physics study recommendations to optimize water-injection rates: pilot in Heera, implement recommendations in South Heera and later in North and Mid Heera wells, adjusting rates for field performance and regularly monitoring production-injection performance.", "Stage_at_disclosure": "Scaled. ONGC’s 2025 annual report says AI/ML-plus-physics water-injection recommendations started implementation at South Heera in March 2023 and were subsequently implemented in North and Mid Heera wells; the 2025 filing says field-performance monitoring continues. Multi-area implementation supports scaled within Heera, not whole-company scale.", "Reported_metric": "No quantified production gain, recovery uplift, injection reduction, accuracy, cost saving, or ROI disclosed.", "Interpretation_limit": "Implementation is described, but no quantified result or validation is reported. The pilot and dated South Heera start precede the window; only the August 2025 report disclosure is in-window. The source does not state well count, input data, decision rights, or model design.", "Source_IDs": "SW3-003", "Source_date": "2025-08-07", "Source_type": "integrated annual report / statutory filing", "Source_locator": "Annexures to the Board's Report, printed p. 131, item 30; PDF p. 157", "Source_URL": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "Evidence_status": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "The report dates the Heera pilot to June 2022 and South Heera implementation to March 2023; later implementation in North and Mid Heera is undated. The first located in-window disclosure covering this implementation is ONGC's FY2024-25 annual report filed August 7, 2025.", "Context": "Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.", "Implementation": "ONGC says Tachyus's AI/ML data-physics technique was adopted for water-injection optimization. Its FY2024-25 report describes recommendations implemented across the Heera assets, with ongoing production-injection monitoring. The source does not identify model architecture, data fields, or an automated control connection.", "Assessment": "Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["Production & field operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-152", "AI-106"], "Disclosures": [{"ID": "SW3-D-AI-225", "Initiative_ID": "AI-225", "Title": "AI/ML data-physics water-injection optimization across Heera assets", "Date": "2025-08-07", "Finding": "Use AI/ML data-physics study recommendations to optimize water-injection rates: pilot in Heera, implement recommendations in South Heera and later in North and Mid Heera wells, adjusting rates for field performance and regularly monitoring production-injection performance.", "Metric": "No quantified production gain, recovery uplift, injection reduction, accuracy, cost saving, or ROI disclosed.", "Source_IDs": "SW3-003", "Evidence": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "Origin": "Third September research swarm"}], "Latest_disclosure": "2025-08-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-225"], "Research_search": "{\"key\": \"swarm3-ai-225\", \"packet_id\": \"swarm3-ai-225\", \"record_ids\": [\"AI-225\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-152\", \"AI-106\"], \"company\": \"Oil and Natural Gas Corporation Limited (ONGC)\", \"title\": \"AI/ML data-physics water-injection optimization across Heera assets\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [\"Production & field operations\"], \"category\": \"Internal adoption\", \"stage\": \"Pilot began in June 2022; recommendations implemented in South Heera from March 2023 and subsequently in North and Mid Heera wells. ONGC reports continued field-performance monitoring in August 2025.\", \"practical_relevance\": \"Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.\", \"business_problem\": \"\", \"workflow\": \"Use AI/ML data-physics study recommendations to optimize water-injection rates: pilot in Heera, implement recommendations in South Heera and later in North and Mid Heera wells, adjusting rates for field performance and regularly monitoring production-injection performance.\", \"technology_data\": \"ONGC says Tachyus's AI/ML data-physics technique was adopted for water-injection optimization. Its FY2024-25 report describes recommendations implemented across the Heera assets, with ongoing production-injection monitoring. The source does not identify model architecture, data fields, or an automated control connection.\", \"integration_human_role\": \"The report describes changes for field performance and regular monitoring; it does not identify human approvers or automated execution authority.\", \"deployment_scope\": \"Pilot began in June 2022; recommendations implemented in South Heera from March 2023 and subsequently in North and Mid Heera wells. ONGC reports continued field-performance monitoring in August 2025.\", \"partners\": [\"Tachyus\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No quantified production gain, recovery uplift, injection reduction, accuracy, cost saving, or ROI disclosed.\", \"period\": \"August 7, 2025 report disclosure; pilot and implementation dates precede the research window.\", \"attribution\": \"Operator-reported implementation; no quantified AI outcome or independent validation disclosed.\", \"source_ids\": [\"SW3-003\"]}], \"economics\": \"No AI-specific economic result disclosed.\", \"timeline\": [{\"date\": \"2025-08-07\", \"event\": \"The report dates the Heera pilot to June 2022 and South Heera implementation to March 2023; later implementation in North and Mid Heera is undated. The first located in-window disclosure covering this implementation is ONGC's FY2024-25 annual report filed August 7, 2025.\", \"source_ids\": [\"SW3-003\"]}], \"limitations\": [\"Implementation is described, but no quantified result or validation is reported. The pilot and dated South Heera start precede the window; only the August 2025 report disclosure is in-window. The source does not state well count, input data, decision rights, or model design.\"], \"assessment\": \"Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.\", \"sources\": [{\"title\": \"ONGC Integrated Annual Report 2024-25\", \"url\": \"https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf\", \"date\": \"2025-08-07\", \"type\": \"integrated annual report / statutory filing\", \"locator\": \"Annexures to the Board's Report, printed p. 131, item 30; PDF p. 157\", \"review_status\": \"Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.\", \"publisher\": \"Oil and Natural Gas Corporation Limited; filed with NSE\", \"id\": \"SW3-003\", \"accessed\": \"2026-09-30\"}], \"source_ids\": [\"SW3-003\"], \"field_sources\": {\"workflow\": [\"SW3-003\"], \"technology_data\": [\"SW3-003\"], \"deployment_scope\": [\"SW3-003\"], \"practical_relevance\": [\"SW3-003\"]}, \"verification\": \"verified primary\", \"cohort\": \"Third gap review · September 2026\", \"report_date\": \"2026-09-30\", \"evidence_note\": \"Operator-reported implementation; no quantified AI outcome or independent validation disclosed.\"}", "Research_status": "Updated · third September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Offshore Heera, South Heera, North Heera, and Mid Heera fields, India", "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-225", "stage": "Scaled", "stage_basis": "ONGC’s 2025 annual report says AI/ML-plus-physics water-injection recommendations started implementation at South Heera in March 2023 and were subsequently implemented in North and Mid Heera wells; the 2025 filing says field-performance monitoring continues. Multi-area implementation supports scaled within Heera, not whole-company scale.", "source_url": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "source_date": "2025-08-07 (disclosed; implementation 2023 onward)", "evidence_origin": "Operator", "limits": "Pilot and first implementation predate the review period; in-window evidence is continued monitoring. No quantified results or company-wide scope disclosed.", "source_id": "SW3-003"}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-226", "Company": "ADNOC", "Company_primary": "ADNOC", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Real-Time Operations Center AI drilling oversight across ADNOC's 120+ rig fleet", "Extracted_finding": "Live rig data is consolidated into dashboards and AI insights for drilling teams to identify risks and coordinate decisions.", "Stage_at_disclosure": "Operator reports deployed across more than 120 rigs as of August 4, 2026.", "Reported_metric": "ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.", "Interpretation_limit": "All performance figures are company-reported and lack measurement definitions, baselines, or independent validation. More than 120 rigs indicates platform deployment scope, not autonomous control. No model architecture, sensor list, or specific intervention examples are disclosed. 23 October scope limit: Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet.", "Source_IDs": "SW3-004; WATCH23AM-SLB-Q3", "Source_date": "2026-08-04", "Source_type": "Operator press release", "Source_locator": "Release body: paragraphs on RTOC deployment, workflow, operating scope and reported metrics", "Source_URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/", "Evidence_status": "Full primary release opened and read; release body reviewed in full.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "ADNOC announced on August 4, 2026 that the AI-enabled Real-Time Operations Center platform had been deployed across more than 120 rigs. The release provides no initial go-live date or rollout milestones.", "Context": "Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.", "Implementation": "SLB DrillOps enables the RTOC within ADNOC’s sovereign cloud. Engineers retain oversight; the release describes progression toward autonomy, not autonomous rig control.\n\n23 October supplier update: SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs.", "Assessment": "Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.", "Primary_department": "Drilling & completions", "Secondary_departments": ["IT, data & enterprise transformation", "Production & field operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW3-D-AI-226", "Initiative_ID": "AI-226", "Title": "Real-Time Operations Center AI drilling oversight across ADNOC's 120+ rig fleet", "Date": "2026-08-04", "Finding": "Live rig data is consolidated into dashboards and AI insights for drilling teams to identify risks and coordinate decisions.", "Metric": "ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.", "Source_IDs": "SW3-004", "Evidence": "Full primary release opened and read; release body reviewed in full.", "Origin": "Third September research swarm"}, {"ID": "WATCH23AM-AI-226", "Initiative_ID": "AI-226", "Title": "SLB repeats ADNOC’s rig-center scope, without a new rollout count", "Date": "2026-10-23", "Finding": "SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs.", "Metric": "No additional rollout count or newly measured benefit. Earlier ADNOC claims retain their own source, period and limitations.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet."}], "Latest_disclosure": "2026-10-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-226"], "Research_search": "{\"key\": \"swarm3-ai-226\", \"packet_id\": \"swarm3-ai-226\", \"record_ids\": [\"AI-226\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"ADNOC\", \"title\": \"Real-Time Operations Center AI drilling oversight across ADNOC's 120+ rig fleet\", \"department\": \"Drilling & completions\", \"secondary_departments\": [\"IT, data & enterprise transformation\", \"Production & field operations\"], \"category\": \"Internal adoption\", \"stage\": \"Operator reports deployed across more than 120 rigs as of August 4, 2026.\", \"practical_relevance\": \"Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.\", \"business_problem\": \"\", \"workflow\": \"Live rig data is consolidated into dashboards and AI insights for drilling teams to identify risks and coordinate decisions.\", \"technology_data\": \"SLB DrillOps enables the RTOC within ADNOC’s sovereign cloud. Engineers retain oversight; the release describes progression toward autonomy, not autonomous rig control.\", \"integration_human_role\": \"Engineers retain oversight and make operational decisions.\", \"deployment_scope\": \"Operator reports deployed across more than 120 rigs as of August 4, 2026.\", \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.\", \"period\": \"August 4, 2026 deployment announcement; outcome measurement period and baselines unspecified.\", \"attribution\": \"ADNOC operator press release, with SLB as technology partner; reported performance claims are not independently validated and their measurement basis is not given.\", \"source_ids\": [\"SW3-004\"]}], \"economics\": \"ADNOC describes reduced engineering effort and avoided rig downtime but reports no monetary savings, implementation cost, or payback.\", \"timeline\": [{\"date\": \"2026-08-04\", \"event\": \"ADNOC announced on August 4, 2026 that the AI-enabled Real-Time Operations Center platform had been deployed across more than 120 rigs. The release provides no initial go-live date or rollout milestones.\", \"source_ids\": [\"SW3-004\"]}], \"limitations\": [\"All performance figures are company-reported and lack measurement definitions, baselines, or independent validation. More than 120 rigs indicates platform deployment scope, not autonomous control. No model architecture, sensor list, or specific intervention examples are disclosed.\"], \"assessment\": \"Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.\", \"sources\": [{\"title\": \"ADNOC and SLB Deploy AI Platform Across Over 120 Drilling Rigs to Strengthen Upstream Performance\", \"url\": \"https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/\", \"date\": \"2026-08-04\", \"type\": \"Operator press release\", \"locator\": \"Release body: paragraphs on RTOC deployment, workflow, operating scope and reported metrics\", \"review_status\": \"Full primary release opened and read; release body reviewed in full.\", \"publisher\": \"ADNOC\", \"id\": \"SW3-004\", \"accessed\": \"2026-09-30\"}], \"source_ids\": [\"SW3-004\"], \"field_sources\": {\"workflow\": [\"SW3-004\"], \"technology_data\": [\"SW3-004\"], \"deployment_scope\": [\"SW3-004\"], \"practical_relevance\": [\"SW3-004\"]}, \"verification\": \"verified primary\", \"cohort\": \"Third gap review · September 2026\", \"report_date\": \"2026-09-30\", \"evidence_note\": \"ADNOC operator press release, with SLB as technology partner; reported performance claims are not independently validated and their measurement basis is not given.\"}", "Research_status": "Updated · third September review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "ADNOC Drilling rig fleet across ADNOC Onshore and ADNOC Offshore assets, UAE", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "120+ rigs", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "All performance figures are company-reported and lack measurement definitions, baselines, or independent validation. More than 120 rigs indicates platform deployment scope, not autonomous control. No model architecture, sensor list, or specific intervention examples are disclosed. 23 October scope limit: Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full primary release opened and read; release body reviewed in full.", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-23"}, {"ID": "AI-227", "Company": "Devon Energy", "Company_primary": "Devon Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Harvey legal AI selection for land, transactional, litigation and compliance work", "Extracted_finding": "Devon announced adoption of Harvey's legal AI platform to support its in-house team's litigation, land, transactional and compliance functions. The announcement does not specify individual tasks or workflows within those functions.", "Stage_at_disclosure": "Harvey announced Devon’s selection/adoption on 14 August 2025. Go-live, production use, rollout completion and adoption breadth are not established.", "Reported_metric": "No workflow performance, usage or adoption metric reported.", "Interpretation_limit": "Vendor announcement with no task-level examples, rollout date, production evidence, controls or quantified results.", "Source_IDs": "SW3-005", "Source_date": "2025-08-14", "Source_type": "Vendor announcement with named customer executive quote", "Source_locator": "Byline and announcement text; Devon Senior Counsel Matt Tilly quote", "Source_URL": "https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform", "Evidence_status": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "PE_relationship": null, "Coverage_bucket": "Third targeted gap review", "Time_scope": "Harvey byline explicitly dates the announcement 14 August 2025. The announcement says Devon 'is adopting' the platform but gives no go-live date.", "Context": "Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.", "Implementation": "Harvey is named as the selected platform. No implementation architecture, integration, model configuration, access controls, rollout schedule or deployment scope is disclosed.", "Assessment": "Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.", "Primary_department": "Land & legal", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Third September research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-006", "AI-007", "AI-086"], "Disclosures": [{"ID": "SW3-D-AI-227", "Initiative_ID": "AI-227", "Title": "Harvey legal AI selection for land, transactional, litigation and compliance work", "Date": "2025-08-14", "Finding": "Devon announced adoption of Harvey's legal AI platform to support its in-house team's litigation, land, transactional and compliance functions. The announcement does not specify individual tasks or workflows within those functions.", "Metric": "No workflow performance, usage or adoption metric reported.", "Source_IDs": "SW3-005", "Evidence": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "Origin": "Third September research swarm"}], "Latest_disclosure": "2025-08-14", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm3-ai-227"], "Research_search": "{\"key\": \"swarm3-ai-227\", \"packet_id\": \"swarm3-ai-227\", \"record_ids\": [\"AI-227\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-006\", \"AI-007\", \"AI-086\"], \"company\": \"Devon Energy\", \"title\": \"Harvey legal AI selection for land, transactional, litigation and compliance work\", \"department\": \"Land & legal\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Harvey announced Devon’s selection/adoption on 14 August 2025. Go-live, production use, rollout completion and adoption breadth are not established.\", \"practical_relevance\": \"Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.\", \"business_problem\": \"\", \"workflow\": \"Devon announced adoption of Harvey's legal AI platform to support its in-house team's litigation, land, transactional and compliance functions. The announcement does not specify individual tasks or workflows within those functions.\", \"technology_data\": \"Harvey is named as the selected platform. No implementation architecture, integration, model configuration, access controls, rollout schedule or deployment scope is disclosed.\", \"integration_human_role\": \"Devon's Senior Counsel describes an aim to devote more time to higher-value work. Review steps, approval authority and controls around AI outputs are not described.\", \"deployment_scope\": \"Harvey announced Devon’s selection/adoption on 14 August 2025. Go-live, production use, rollout completion and adoption breadth are not established.\", \"partners\": [\"Harvey\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No workflow performance, usage or adoption metric reported.\", \"period\": \"14 August 2025 adoption announcement; no production measurement period disclosed\", \"attribution\": \"Vendor announcement with a named Devon Senior Counsel quote; not independently validated.\", \"source_ids\": [\"SW3-005\"]}], \"economics\": \"No realized savings, labor-cost reduction, implementation cost or payback disclosed. Efficiency and scale are stated objectives, not measured outcomes.\", \"timeline\": [{\"date\": \"2025-08-14\", \"event\": \"Harvey byline explicitly dates the announcement 14 August 2025. The announcement says Devon 'is adopting' the platform but gives no go-live date.\", \"source_ids\": [\"SW3-005\"]}], \"limitations\": [\"Vendor announcement with no task-level examples, rollout date, production evidence, controls or quantified results.\"], \"assessment\": \"Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.\", \"sources\": [{\"title\": \"Devon Energy Selects Harvey's Legal AI Platform\", \"url\": \"https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform\", \"date\": \"2025-08-14\", \"type\": \"Vendor announcement with named customer executive quote\", \"locator\": \"Byline and announcement text; Devon Senior Counsel Matt Tilly quote\", \"review_status\": \"Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.\", \"publisher\": \"Harvey\", \"id\": \"SW3-005\", \"accessed\": \"2026-09-30\"}], \"source_ids\": [\"SW3-005\"], \"field_sources\": {\"workflow\": [\"SW3-005\"], \"technology_data\": [\"SW3-005\"], \"deployment_scope\": [\"SW3-005\"], \"practical_relevance\": [\"SW3-005\"]}, \"verification\": \"verified partner\", \"cohort\": \"Third gap review · September 2026\", \"report_date\": \"2026-09-30\", \"evidence_note\": \"Vendor announcement with a named Devon Senior Counsel quote; not independently validated.\"}", "Research_status": "Updated · third September review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-09-30", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Devon in-house legal team; geographic rollout scope not specified", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-228", "Company": "Moeve (formerly Cepsa)", "Company_primary": "Moeve", "Segment": "Downstream", "Theme": "Operational AI", "Initiative": "Visbreaker stability and sulphur prediction at La Rábida", "Extracted_finding": "AI models predict outlet stability and sulphur content in the La Rábida Visbreaker unit.", "Stage_at_disclosure": "Model development reported for named process units; production rollout extent unspecified.", "Reported_metric": "No accuracy, yield or financial result quantified.", "Interpretation_limit": "No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.", "Source_IDs": "SW4-001", "Source_date": "2026-02-26", "Source_type": "Operator annual management report", "Source_locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "Source_URL": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "Evidence_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "PE_relationship": "FY2025 report identifies Carlyle-controlled Matador Bidco holding 38.41%, and Mubadala-controlled Cepsa Holding 61.36%. Ownership does not establish sponsor involvement in the AI project.", "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "FY2025 report adopted by the board 26 February 2026; online publication day unverified.", "Context": "Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.", "Implementation": "Process-property prediction; model design, input data and system integration are unspecified.", "Assessment": "Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-228", "Initiative_ID": "AI-228", "Title": "Visbreaker stability and sulphur prediction at La Rábida", "Date": "2026-02-26", "Finding": "AI models predict outlet stability and sulphur content in the La Rábida Visbreaker unit.", "Metric": "No accuracy, yield or financial result quantified.", "Source_IDs": "SW4-001", "Evidence": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2026-02-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-228"], "Research_search": "{\"key\": \"swarm4-ai-228\", \"packet_id\": \"swarm4-ai-228\", \"record_ids\": [\"AI-228\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Moeve (formerly Cepsa)\", \"title\": \"Visbreaker stability and sulphur prediction at La Rábida\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Model development reported for named process units; production rollout extent unspecified.\", \"practical_relevance\": \"Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.\", \"business_problem\": \"\", \"workflow\": \"AI models predict outlet stability and sulphur content in the La Rábida Visbreaker unit.\", \"technology_data\": \"Process-property prediction; model design, input data and system integration are unspecified.\", \"integration_human_role\": \"Approval roles and control-system write authority are not disclosed.\", \"deployment_scope\": \"Model development reported for named process units; production rollout extent unspecified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No accuracy, yield or financial result quantified.\", \"period\": \"FY2025; precise deployment and measurement dates undisclosed\", \"attribution\": \"Operator disclosure, not independent performance validation.\", \"source_ids\": [\"SW4-001\"]}], \"economics\": \"No project cost or payback disclosed.\", \"timeline\": [{\"date\": \"2026-02-26\", \"event\": \"FY2025 report adopted by the board 26 February 2026; online publication day unverified.\", \"source_ids\": [\"SW4-001\"]}], \"limitations\": [\"No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.\"], \"assessment\": \"Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.\", \"sources\": [{\"title\": \"Moeve Consolidated Management Report 2025\", \"url\": \"https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf\", \"date\": \"2026-02-26\", \"type\": \"Operator annual management report\", \"locator\": \"p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174\", \"review_status\": \"Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.\", \"publisher\": \"Moeve\", \"id\": \"SW4-001\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-001\"], \"field_sources\": {\"workflow\": [\"SW4-001\"], \"technology_data\": [\"SW4-001\"], \"deployment_scope\": [\"SW4-001\"], \"practical_relevance\": [\"SW4-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator disclosure, not independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "La Rábida Energy Park, Huelva, Spain", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-229", "Company": "Moeve (formerly Cepsa)", "Company_primary": "Moeve", "Segment": "Downstream", "Theme": "Operational AI", "Initiative": "Real-time diesel and kerosene blending-property estimation at San Roque", "Extracted_finding": "AI estimates diesel and kerosene blending properties in real time at San Roque.", "Stage_at_disclosure": "Model development reported for named process units; production rollout extent unspecified.", "Reported_metric": "No accuracy, yield or financial result quantified.", "Interpretation_limit": "No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.", "Source_IDs": "SW4-001", "Source_date": "2026-02-26", "Source_type": "Operator annual management report", "Source_locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "Source_URL": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "Evidence_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "PE_relationship": "FY2025 report identifies Carlyle-controlled Matador Bidco holding 38.41%, and Mubadala-controlled Cepsa Holding 61.36%. Ownership does not establish sponsor involvement in the AI project.", "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "FY2025 report adopted by the board 26 February 2026; online publication day unverified.", "Context": "Distinct from Visbreaker prediction: separate site, process and operational decision.", "Implementation": "The report identifies the blending process but does not specify estimator inputs, property list or optimization constraints.", "Assessment": "Distinct from Visbreaker prediction: separate site, process and operational decision.", "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-229", "Initiative_ID": "AI-229", "Title": "Real-time diesel and kerosene blending-property estimation at San Roque", "Date": "2026-02-26", "Finding": "AI estimates diesel and kerosene blending properties in real time at San Roque.", "Metric": "No accuracy, yield or financial result quantified.", "Source_IDs": "SW4-001", "Evidence": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2026-02-26", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-229"], "Research_search": "{\"key\": \"swarm4-ai-229\", \"packet_id\": \"swarm4-ai-229\", \"record_ids\": [\"AI-229\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Moeve (formerly Cepsa)\", \"title\": \"Real-time diesel and kerosene blending-property estimation at San Roque\", \"department\": \"Refining, LNG & process operations\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Model development reported for named process units; production rollout extent unspecified.\", \"practical_relevance\": \"Distinct from Visbreaker prediction: separate site, process and operational decision.\", \"business_problem\": \"\", \"workflow\": \"AI estimates diesel and kerosene blending properties in real time at San Roque.\", \"technology_data\": \"The report identifies the blending process but does not specify estimator inputs, property list or optimization constraints.\", \"integration_human_role\": \"Approval roles and control-system write authority are not disclosed.\", \"deployment_scope\": \"Model development reported for named process units; production rollout extent unspecified.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No accuracy, yield or financial result quantified.\", \"period\": \"FY2025; precise deployment and measurement dates undisclosed\", \"attribution\": \"Operator disclosure, not independent performance validation.\", \"source_ids\": [\"SW4-001\"]}], \"economics\": \"No project cost or payback disclosed.\", \"timeline\": [{\"date\": \"2026-02-26\", \"event\": \"FY2025 report adopted by the board 26 February 2026; online publication day unverified.\", \"source_ids\": [\"SW4-001\"]}], \"limitations\": [\"No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.\"], \"assessment\": \"Distinct from Visbreaker prediction: separate site, process and operational decision.\", \"sources\": [{\"title\": \"Moeve Consolidated Management Report 2025\", \"url\": \"https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf\", \"date\": \"2026-02-26\", \"type\": \"Operator annual management report\", \"locator\": \"p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174\", \"review_status\": \"Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.\", \"publisher\": \"Moeve\", \"id\": \"SW4-001\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-001\"], \"field_sources\": {\"workflow\": [\"SW4-001\"], \"technology_data\": [\"SW4-001\"], \"deployment_scope\": [\"SW4-001\"], \"practical_relevance\": [\"SW4-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator disclosure, not independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "San Roque Energy Park, Cádiz, Spain", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-230", "Company": "Moeve (formerly Cepsa)", "Company_primary": "Moeve", "Segment": "Downstream", "Theme": "Operational AI", "Initiative": "Mistral document-understanding assistant pilot for industrial centers", "Extracted_finding": "Planned industrial digital assistant to help staff understand documents and carry out business tasks.", "Stage_at_disclosure": "Partnership and planned pilot announced; operating deployment not established.", "Reported_metric": "No measured adoption or productivity result.", "Interpretation_limit": "Announcement is prospective. Neither rollout across all sites nor energy/cost savings is evidenced.", "Source_IDs": "SW4-002; SW4-001", "Source_date": "2025-12-04", "Source_type": "Operator press release", "Source_locator": "Pilot digital assistant paragraphs; press index spells out 4 December 2025", "Source_URL": "https://www.moeveglobal.com/en/press/moeve-mistral-ai-alliance-ai-digital-assistant-industrial-efficiency", "Evidence_status": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "PE_relationship": "FY2025 report identifies Carlyle-controlled Matador Bidco holding 38.41%, and Mubadala-controlled Cepsa Holding 61.36%. Ownership does not establish sponsor involvement in the AI project.", "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "4 December 2025 agreement announces a first pilot to be developed; go-live unreported.", "Context": "Separates generative document assistance from Moeve’s process-prediction models.", "Implementation": "Mistral Applied AI and Moeve teams plan a tailored language-model assistant; document corpus, integrations and user count are not stated.", "Assessment": "Separates generative document assistance from Moeve’s process-prediction models.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-230", "Initiative_ID": "AI-230", "Title": "Mistral document-understanding assistant pilot for industrial centers", "Date": "2025-12-04", "Finding": "Planned industrial digital assistant to help staff understand documents and carry out business tasks.", "Metric": "No measured adoption or productivity result.", "Source_IDs": "SW4-002; SW4-001", "Evidence": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2025-12-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-230"], "Research_search": "{\"key\": \"swarm4-ai-230\", \"packet_id\": \"swarm4-ai-230\", \"record_ids\": [\"AI-230\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Moeve (formerly Cepsa)\", \"title\": \"Mistral document-understanding assistant pilot for industrial centers\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Partnership and planned pilot announced; operating deployment not established.\", \"practical_relevance\": \"Separates generative document assistance from Moeve’s process-prediction models.\", \"business_problem\": \"\", \"workflow\": \"Planned industrial digital assistant to help staff understand documents and carry out business tasks.\", \"technology_data\": \"Mistral Applied AI and Moeve teams plan a tailored language-model assistant; document corpus, integrations and user count are not stated.\", \"integration_human_role\": \"Intended staff assistance; review and escalation controls not described.\", \"deployment_scope\": \"Partnership and planned pilot announced; operating deployment not established.\", \"partners\": [\"Mistral AI\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No measured adoption or productivity result.\", \"period\": \"December 2025 announcement; no operating results\", \"attribution\": \"Operator disclosure, not independent performance validation.\", \"source_ids\": [\"SW4-002\", \"SW4-001\"]}], \"economics\": \"No project cost or payback disclosed.\", \"timeline\": [{\"date\": \"2025-12-04\", \"event\": \"4 December 2025 agreement announces a first pilot to be developed; go-live unreported.\", \"source_ids\": [\"SW4-002\", \"SW4-001\"]}], \"limitations\": [\"Announcement is prospective. Neither rollout across all sites nor energy/cost savings is evidenced.\"], \"assessment\": \"Separates generative document assistance from Moeve’s process-prediction models.\", \"sources\": [{\"title\": \"Moeve and Mistral AI join forces to spearhead generative AI solutions in the energy sector\", \"url\": \"https://www.moeveglobal.com/en/press/moeve-mistral-ai-alliance-ai-digital-assistant-industrial-efficiency\", \"date\": \"2025-12-04\", \"type\": \"Operator press release\", \"locator\": \"Pilot digital assistant paragraphs; press index spells out 4 December 2025\", \"review_status\": \"Full operator release and dated issuer press index read; planned pilot kept separate from deployment.\", \"publisher\": \"Moeve\", \"id\": \"SW4-002\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Moeve Consolidated Management Report 2025\", \"url\": \"https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf\", \"date\": \"2026-02-26\", \"type\": \"Operator annual management report\", \"locator\": \"p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174\", \"review_status\": \"Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.\", \"publisher\": \"Moeve\", \"id\": \"SW4-001\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-002\", \"SW4-001\"], \"field_sources\": {\"workflow\": [\"SW4-002\", \"SW4-001\"], \"technology_data\": [\"SW4-002\", \"SW4-001\"], \"deployment_scope\": [\"SW4-002\", \"SW4-001\"], \"practical_relevance\": [\"SW4-002\", \"SW4-001\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator disclosure, not independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Moeve industrial centers; pilot sites unspecified", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-231", "Company": "Presidio Production Company (NYSE: FTW)", "Company_primary": "Presidio Production", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "FTW Technologies live AI well surveillance and production intelligence at Presidio", "Extracted_finding": "Operating surveillance and production intelligence, with an achieved-result claim kept separate from wider agent plans.", "Stage_at_disclosure": "Operating: approximately 2,000 wells on intelligence platform; field engineers receive DOUG recommendations. This is not fleet-wide autonomous control.", "Reported_metric": "Operator reports 2.3% AI-linked production uplift through Q2: 1.4 points from DOUG and nearly 1 point from adjacent initiatives. DOUG averaged ~400 boe/d over the prior three months. Controllers on 24 wells: 2,680 to 2,946 Mcf/day, roughly 10% on that population. Weekend production reportedly rose 2.5%, with comparison baseline undisclosed. These overlap the program total; do not add them. $4.5m annualized revenue covers the initiatives, not net cash or DOUG alone. August 17 deck repeats 2.3% YTD, targets 3–5%, and cites ~2,300 platform wells on its GOAL slide. $2.5m per uplift point is illustrative revenue sensitivity.", "Interpretation_limit": "Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs. Controller pre/post periods, matched wells, decline adjustments, weekend baseline and controller safeguards remain unspecified. The 24 controller wells are a separate population from platform coverage.", "Source_IDs": "SW4-003; OCT5-S024; OCT10-PRESIDIO-0817", "Source_date": "2026-08-17", "Source_type": "SEC-filed investor presentation", "Source_locator": "Slides 20–23; current surveillance versus 2026 deployment column on slide 21; GOAL slide 23", "Source_URL": "https://ir.bypresidio.com/sec-filings/all-sec-filings/content/0001213900-26-090532/ea030229001ex99-1.htm", "Evidence_status": "Relevant original issuer presentation text reviewed 10 October 2026; results remain issuer-reported.", "PE_relationship": "Historically backed by Morgan Stanley Energy Partners; publicly listed by March 2026. No sponsor-directed AI attribution established.", "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "Latest disclosure: August 12, 2026, reporting Q2 results. Earlier June evidence retained in disclosure history.", "Context": "Surveillance and field-work priorities on mature producing assets.", "Implementation": "Live 24/7 well surveillance and natural-language production queries support field operators and engineers. DOUG sends recommendations to field engineers; weekend routing provides pumpers a prioritized map. Twenty-four plunger boxes continuously adjust cycle timing; safeguards and overrides are undisclosed. Budget, acquisition-screening and fleet-orchestration agents remain 2026 deployment work, not established full-fleet autonomy.", "Assessment": "The issuer connects field surveillance, staff knowledge and operating recommendations. The wider agent program and production target require follow-through; comparable net returns remain undisclosed.", "Primary_department": "Production & field operations", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-231", "Initiative_ID": "AI-231", "Title": "FTW Technologies live AI well surveillance and production intelligence at Presidio", "Date": "2026-06", "Finding": "Presidio says a 24/7 AI surveillance tool is live in well surveillance and production intelligence and available to all operators and engineers. Users can query live production data in natural language. The underlying well-intelligence layer combines field pumper voice/text observations with well files, drilling and completion reports, production volumes, and geology. Operations staff remain central; an asset-intelligence group uses analytics to drive field efficiencies. The deck also describes machine-learning pump-by-exception routing: high-cash-flow wells receive daily visits and other wells are visited on exceptions; field crews remain responsible for action.", "Metric": "Company reports approximately 1% YTD production uplift toward a 3–5% 2026 target. No measurement cutoff, baseline, attribution method or independent validation is supplied. Separately, its pump-by-exception operating case reports 50% fewer well visits; scope is selected acquisition case studies, not necessarily the entire fleet.", "Source_IDs": "SW4-003", "Evidence": "Full SEC-hosted HTML presentation opened and reviewed, including legal disclaimer and AI slides.", "Origin": "Fourth research swarm"}, {"ID": "OCT5-AI-231-1", "Initiative_ID": "AI-231", "Title": "DOUG surveillance, weekend routing and plunger optimization: Q2 update", "Date": "2026-08-12", "Finding": "Presidio’s Q2 call distinguishes DOUG recommendations, weekend visit routing and automatic plunger-cycle adjustments. DOUG alerts field engineers; pumpers receive a prioritized map. Controllers on 24 wells continuously adjust cycle timing. Management says field judgment remains with staff; controller safeguards are unspecified.", "Metric": "Reported gas rates on the controller population: 2,680 to 2,946 Mcf/day, roughly 10%. Weekend production reportedly rose 2.5%. These overlap the existing 2.3% program-uplift claim and must not be added to it. Existing program revenue is not net cash.", "Source_IDs": "OCT5-S024", "Source_URL": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review", "Assessment": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage."}, {"ID": "OCT10-AI-231-0817", "Initiative_ID": "AI-231", "Title": "August presentation: achieved uplift, platform scope and 2026 roadmap", "Date": "2026-08-17", "Finding": "The August deck repeats 2.3% YTD AI-linked production uplift and a 3–5% 2026 target. It describes live surveillance and natural-language production queries. Its goal slide cites roughly 2,300 platform wells; fleet-wide agentic orchestration, budget agents and automatic acquisition screening appear in the 2026 deployment column, not the NOW column.", "Metric": "August deck: 2.3% YTD achieved, company-reported; 3–5% 2026 production-uplift target. Roughly 2,300 wells appears on the intelligence-platform GOAL slide; coverage date and active intervention population are not established. $2.5m annual revenue per 1% uplift at $30/boe is illustrative, not a realized return.", "Source_IDs": "OCT10-PRESIDIO-0817", "Source_URL": "https://ir.bypresidio.com/sec-filings/all-sec-filings/content/0001213900-26-090532/ea030229001ex99-1.htm", "Evidence": "Relevant original issuer presentation text reviewed 10 October 2026; results remain issuer-reported.", "Origin": "10 October integration review", "Assessment": "Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs."}], "Latest_disclosure": "2026-08-17", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct5-followup-ai-231", "oct10-presidio-august-deck", "swarm4-ai-231"], "Research_search": "{\"title\": \"Presidio separates field recommendations from controller adjustments\", \"stage\": \"Existing operating profile; Q2 workflow detail clarified\", \"workflow\": \"Live 24/7 well surveillance and natural-language production queries support field operators and engineers. DOUG sends recommendations to field engineers; weekend routing provides pumpers a prioritized map. Twenty-four plunger boxes continuously adjust cycle timing; safeguards and overrides are undisclosed. Budget, acquisition-screening and fleet-orchestration agents remain 2026 deployment work, not established full-fleet autonomy.\", \"technology_data\": \"DOUG alerts field engineers; pumpers receive a prioritized map. Controllers on 24 wells continuously adjust cycle timing. Management says field judgment remains with staff; controller safeguards are unspecified.\", \"integration_human_role\": \"DOUG advises engineers; pumpers use a map; controller timing adjustments are automatic, with safeguards undisclosed.\", \"deployment_scope\": \"24 controller wells within Presidio’s operated base. Call: 12 August 2026; controller pre/post periods and weekend baseline unknown. Platform scope is a separate population.\", \"practical_relevance\": \"What happened on comparable wells, and what net benefit remained after controller and field-work costs?\", \"business_problem\": \"Surveillance and field-work priorities on mature producing assets.\", \"economics\": \"Operator reports 2.3% AI-linked production uplift through Q2: 1.4 points from DOUG and nearly 1 point from adjacent initiatives. DOUG averaged ~400 boe/d over the prior three months. Controllers on 24 wells: 2,680 to 2,946 Mcf/day, roughly 10% on that population. Weekend production reportedly rose 2.5%, with comparison baseline undisclosed. These overlap the program total; do not add them. $4.5m annualized revenue covers the initiatives, not net cash or DOUG alone. August 17 deck repeats 2.3% YTD, targets 3–5%, and cites ~2,300 platform wells on its GOAL slide. $2.5m per uplift point is illustrative revenue sensitivity.\", \"limitations\": [\"Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs. Controller pre/post periods, matched wells, decline adjustments, weekend baseline and controller safeguards remain unspecified. The 24 controller wells are a separate population from platform coverage.\"], \"assessment\": \"The issuer connects field surveillance, staff knowledge and operating recommendations. The wider agent program and production target require follow-through; comparable net returns remain undisclosed.\", \"packet_id\": \"oct5-followup-ai-231\", \"key\": \"oct5-followup-ai-231\", \"record_ids\": [\"AI-231\"], \"company\": \"Presidio Production\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"source_ids\": [\"OCT5-S024\"], \"report_date\": \"2026-10-18\", \"verification\": \"verified primary\", \"field_sources\": {\"business_problem\": [\"OCT5-S024\"], \"workflow\": [\"OCT5-S024\"], \"technology_data\": [\"OCT5-S024\"], \"deployment_scope\": [\"OCT5-S024\"], \"economics\": [\"OCT5-S024\"], \"integration_human_role\": [\"OCT5-S024\"], \"practical_relevance\": [\"OCT5-S024\"]}, \"disclosed_results\": [{\"metric\": \"Q2 2026 company report\", \"value\": \"Operator reports 2.3% AI-linked production uplift through Q2: 1.4 points from DOUG and nearly 1 point from adjacent initiatives. DOUG averaged ~400 boe/d incremental production over the prior three months. $4.5m annualized revenue spans the initiatives; it is not net cash or a DOUG-only result.\", \"period\": \"2026-08-12\", \"attribution\": \"Operator-reported, not independently evaluated\", \"source_ids\": [\"OCT5-S024\"]}], \"timeline\": [{\"date\": \"2026-08-12\", \"event\": \"Q2 update replaces the earlier progress figure; prior evidence remains in the following dossier and disclosure history.\", \"source_ids\": [\"OCT5-S024\"]}, {\"date\": \"2026-08-12\", \"review_on\": \"2026-10-18\", \"event\": \"Presidio’s Q2 call distinguishes DOUG recommendations, weekend visit routing and automatic plunger-cycle adjustments. Reported gas rates on the controller population: 2,680 to 2,946 Mcf/day, roughly 10%. Weekend production reportedly rose 2.5%. These overlap the existing 2.3% program-uplift claim and must not be added to it.\", \"source_ids\": [\"OCT5-S024\"]}], \"sources\": [{\"id\": \"OCT5-S024\", \"title\": \"Presidio Production Company Q2 2026 Earnings Call Transcript\", \"url\": \"https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm\", \"date\": \"2026-08-12\", \"publisher\": \"Presidio Production\", \"locator\": \"Chris Hammack remarks, printed page 3: weekend coverage, 24 plunger boxes, human judgment; Will Ulrich page 2 for aggregate uplift/revenue\", \"excerpt\": \"None of these replace field judgment\", \"access\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\", \"review_status\": \"Original publisher text read by editor 18 October 2026; claims remain source-reported.\"}], \"earlier_review\": {\"report_date\": \"2026-10-05\", \"assessment\": \"A specific path from data to field action is disclosed. Revenue attribution is useful, but missing costs prevent a project return calculation.\", \"limitations\": [\"Company-reported attribution; no independent counterfactual, measurement protocol or project costs. The $4.5m annualized revenue spans AI initiatives and is not realized net cash.\"], \"note\": \"Superseded by targeted 18 October 2026 source check.\"}, \"last_source_check\": \"2026-10-18\"}", "Research_status": "Targeted source detail checked 18 October 2026 evening", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Presidio-operated Anadarko Basin wells, Texas, Oklahoma, and Kansas; its June 2026 deck lists approximately 2,300 operated wells.", "Normalized_stage": "Operating", "Stage_review": {"date": "2026-10-05", "stage": "Operating", "stage_basis": "August 12 Q2 transcript names current DOUG workflow and human handoff", "source_url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "limits": "2,000 wells describes the intelligence platform; not all wells have autonomous control."}, "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Current_source_check": "2026-10-18", "Evidence_review": {"source_read": "verified primary", "claim_status": "Existing operating profile; Q2 workflow detail clarified", "independent_validation": "Not established by this collection"}, "Earlier_source_review": {"as_of": "Before 18 October evening source check", "finding": "Operating surveillance and production intelligence, with an achieved-result claim kept separate from wider agent plans.", "limit": "Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs.", "source_date": "2026-08-17", "note": "Earlier review retained for provenance; current fields below supersede it."}, "Research_origin": ["research/watch-2026-10-18-pm/README.md"]}, {"ID": "AI-232", "Company": "SLB", "Company_primary": "SLB", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "AI Builder flight-invoice extraction for airline GST processing", "Extracted_finding": "SLB's Flight Airline Automation Bot retrieves flight invoices from email, uses AI Builder to extract 22 key fields and tax details, validates the data, and compiles an Airline GST report for the tax team.", "Stage_at_disclosure": "SLB describes the bot as implemented and processing invoices; the source gives no go-live date or rollout scope beyond the stated volume.", "Reported_metric": "SLB reports processing more than 1,000 emails and 200 vendor invoices per month and effort equivalent to one full-time employee saved. The case separately says 22 invoice fields are extracted.", "Interpretation_limit": "The August 8, 2025 date is the Microsoft Learn page's explicit last-updated date, not the deployment date. Vendor-published case; customer executive quote and detailed workflow are disclosed, but the source does not give accuracy, exception rate, baseline minutes, governance details for ambiguous records, or audited savings. Invoice scope is airline travel invoices and tax processing, not supplier AP/JIB/revenue accounting.", "Source_IDs": "SW4-004", "Source_date": "2025-08-08", "Source_type": "Microsoft Learn customer case study with named SLB architect", "Source_locator": "Flight Airline Automation Bot and tax process: page sections 148–174; page last-updated stamp at line 287; date is page last update, not go-live", "Source_URL": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "Evidence_status": "Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.", "PE_relationship": null, "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "Microsoft Learn displays 'Last updated on 2025-08-08'. The page confirms the workflow is implemented but does not date the implementation.", "Context": "A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.", "Implementation": "Power Automate retrieves and filters vendor emails, AI Builder extracts invoice fields, business validations check extracted details, and the result is compiled for tax processing. The page identifies Dataverse and Azure Key Vault in the back end. It does not describe a posting to the general ledger or payment execution.", "Assessment": "A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-232", "Initiative_ID": "AI-232", "Title": "AI Builder flight-invoice extraction for airline GST processing", "Date": "2025-08-08", "Finding": "SLB's Flight Airline Automation Bot retrieves flight invoices from email, uses AI Builder to extract 22 key fields and tax details, validates the data, and compiles an Airline GST report for the tax team.", "Metric": "SLB reports processing more than 1,000 emails and 200 vendor invoices per month and effort equivalent to one full-time employee saved. The case separately says 22 invoice fields are extracted.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2025-08-08", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-232"], "Research_search": "{\"key\": \"swarm4-ai-232\", \"packet_id\": \"swarm4-ai-232\", \"record_ids\": [\"AI-232\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"AI Builder flight-invoice extraction for airline GST processing\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [\"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"SLB describes the bot as implemented and processing invoices; the source gives no go-live date or rollout scope beyond the stated volume.\", \"practical_relevance\": \"A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.\", \"business_problem\": \"\", \"workflow\": \"SLB's Flight Airline Automation Bot retrieves flight invoices from email, uses AI Builder to extract 22 key fields and tax details, validates the data, and compiles an Airline GST report for the tax team.\", \"technology_data\": \"Power Automate retrieves and filters vendor emails, AI Builder extracts invoice fields, business validations check extracted details, and the result is compiled for tax processing. The page identifies Dataverse and Azure Key Vault in the back end. It does not describe a posting to the general ledger or payment execution.\", \"integration_human_role\": \"The tax team receives the compiled GST report and uses it for tax processing. Validation is described, but the page does not specify whether checks are automated, who resolves exceptions, or an approval gate before posting.\", \"deployment_scope\": \"SLB describes the bot as implemented and processing invoices; the source gives no go-live date or rollout scope beyond the stated volume.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"SLB reports processing more than 1,000 emails and 200 vendor invoices per month and effort equivalent to one full-time employee saved. The case separately says 22 invoice fields are extracted.\", \"period\": \"Microsoft Learn page last updated August 8, 2025; underlying bot go-live and metric measurement window not disclosed\", \"attribution\": \"Microsoft customer case study quoting named SLB Lead Architect Rushi Patel; results are customer-reported through a technology-provider case study, not independently validated.\", \"source_ids\": [\"SW4-004\"]}], \"economics\": \"One-FTE equivalent is reported workload capacity, not a disclosed cash payroll saving. No labor cost realization, software/deployment cost, or payback is reported.\", \"timeline\": [{\"date\": \"2025-08-08\", \"event\": \"Microsoft Learn displays 'Last updated on 2025-08-08'. The page confirms the workflow is implemented but does not date the implementation.\", \"source_ids\": [\"SW4-004\"]}], \"limitations\": [\"The August 8, 2025 date is the Microsoft Learn page's explicit last-updated date, not the deployment date. Vendor-published case; customer executive quote and detailed workflow are disclosed, but the source does not give accuracy, exception rate, baseline minutes, governance details for ambiguous records, or audited savings. Invoice scope is airline travel invoices and tax processing, not supplier AP/JIB/revenue accounting.\"], \"assessment\": \"A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.\", \"sources\": [{\"title\": \"SLB enhances productivity with Power Platform and AI\", \"url\": \"https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity\", \"date\": \"2025-08-08\", \"type\": \"Microsoft Learn customer case study with named SLB architect\", \"locator\": \"Flight Airline Automation Bot and tax process: page sections 148–174; page last-updated stamp at line 287; date is page last update, not go-live\", \"review_status\": \"Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.\", \"publisher\": \"Microsoft\", \"id\": \"SW4-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-004\"], \"field_sources\": {\"workflow\": [\"SW4-004\"], \"technology_data\": [\"SW4-004\"], \"deployment_scope\": [\"SW4-004\"], \"practical_relevance\": [\"SW4-004\"]}, \"verification\": \"verified partner\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Microsoft customer case study quoting named SLB Lead Architect Rushi Patel; results are customer-reported through a technology-provider case study, not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "SLB tax team; location/rollout geography not specified", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-233", "Company": "SLB", "Company_primary": "SLB", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "AI-powered project discovery and recommendations across SLB facilities", "Extracted_finding": "SLB employees use an AI-powered project recommendation tool to find similar initiatives across global facilities, identify opportunities to consolidate work, and discover projects to replicate or join. Financial controllers validate project records; management reviews the resulting projects.", "Stage_at_disclosure": "Implemented and in continued use; Microsoft page says SLB continues to benefit from the application. Launch date and global rollout coverage are not stated.", "Reported_metric": "Within first 28 days of launch, SLB reports 800+ AI-powered matches and 150 similar projects inspiring facility initiatives. These are activity/claimed follow-on counts, not quantified savings or verified project consolidation.", "Interpretation_limit": "Source is a Microsoft-hosted customer story; no publication date beyond last-updated 2025-08-08, no deployment launch date, precision/recall, project-value attribution, costs, or independent verification. This is an internal project-discovery workflow, not a field production deployment.", "Source_IDs": "SW4-004", "Source_date": "2025-08-08", "Source_type": "Microsoft Learn customer case study", "Source_locator": "AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live", "Source_URL": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "Evidence_status": "Full official Microsoft Learn page opened and reviewed.", "PE_relationship": null, "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "Page states the initiative is part of SLB’s 2024 internal strategy. It reports outcomes within the first 28 days after launch but does not state the launch calendar date. The page itself was last updated 2025-08-08.", "Context": "Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.", "Implementation": "Power Apps supports project submission/review; Dataverse records synchronize to Azure Blob Storage; a Power Automate flow calls Azure Functions/Python to retrieve similar projects using text similarity. The system excludes the current user’s own projects from recommendations. Financial controllers validate records and management reviews projects.", "Assessment": "Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-233", "Initiative_ID": "AI-233", "Title": "AI-powered project discovery and recommendations across SLB facilities", "Date": "2025-08-08", "Finding": "SLB employees use an AI-powered project recommendation tool to find similar initiatives across global facilities, identify opportunities to consolidate work, and discover projects to replicate or join. Financial controllers validate project records; management reviews the resulting projects.", "Metric": "Within first 28 days of launch, SLB reports 800+ AI-powered matches and 150 similar projects inspiring facility initiatives. These are activity/claimed follow-on counts, not quantified savings or verified project consolidation.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2025-08-08", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-233"], "Research_search": "{\"key\": \"swarm4-ai-233\", \"packet_id\": \"swarm4-ai-233\", \"record_ids\": [\"AI-233\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"AI-powered project discovery and recommendations across SLB facilities\", \"department\": \"IT, data & enterprise transformation\", \"secondary_departments\": [\"Finance, investment & strategy\"], \"category\": \"Internal adoption\", \"stage\": \"Implemented and in continued use; Microsoft page says SLB continues to benefit from the application. Launch date and global rollout coverage are not stated.\", \"practical_relevance\": \"Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.\", \"business_problem\": \"\", \"workflow\": \"SLB employees use an AI-powered project recommendation tool to find similar initiatives across global facilities, identify opportunities to consolidate work, and discover projects to replicate or join. Financial controllers validate project records; management reviews the resulting projects.\", \"technology_data\": \"Power Apps supports project submission/review; Dataverse records synchronize to Azure Blob Storage; a Power Automate flow calls Azure Functions/Python to retrieve similar projects using text similarity. The system excludes the current user’s own projects from recommendations. Financial controllers validate records and management reviews projects.\", \"integration_human_role\": \"Employees review and act on recommendations; controllers validate project records and management reviews projects. Source does not detail model governance or human adjudication of similarity matches.\", \"deployment_scope\": \"Implemented and in continued use; Microsoft page says SLB continues to benefit from the application. Launch date and global rollout coverage are not stated.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Within first 28 days of launch, SLB reports 800+ AI-powered matches and 150 similar projects inspiring facility initiatives. These are activity/claimed follow-on counts, not quantified savings or verified project consolidation.\", \"period\": \"Microsoft Learn page last updated 2025-08-08; workflow chronology/launch dates not disclosed\", \"attribution\": \"Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.\", \"source_ids\": [\"SW4-004\"]}], \"economics\": \"No financial savings, implementation costs, or payback disclosed. Benefits are described as collaboration and innovation.\", \"timeline\": [{\"date\": \"2025-08-08\", \"event\": \"Page states the initiative is part of SLB’s 2024 internal strategy. It reports outcomes within the first 28 days after launch but does not state the launch calendar date. The page itself was last updated 2025-08-08.\", \"source_ids\": [\"SW4-004\"]}], \"limitations\": [\"Source is a Microsoft-hosted customer story; no publication date beyond last-updated 2025-08-08, no deployment launch date, precision/recall, project-value attribution, costs, or independent verification. This is an internal project-discovery workflow, not a field production deployment.\"], \"assessment\": \"Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.\", \"sources\": [{\"title\": \"SLB enhances productivity with Power Platform and AI\", \"url\": \"https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity\", \"date\": \"2025-08-08\", \"type\": \"Microsoft Learn customer case study\", \"locator\": \"AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live\", \"review_status\": \"Full official Microsoft Learn page opened and reviewed.\", \"publisher\": \"Microsoft\", \"id\": \"SW4-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-004\"], \"field_sources\": {\"workflow\": [\"SW4-004\"], \"technology_data\": [\"SW4-004\"], \"deployment_scope\": [\"SW4-004\"], \"practical_relevance\": [\"SW4-004\"]}, \"verification\": \"verified partner\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Global SLB facilities / global workforce; countries or rollout coverage not enumerated", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full official Microsoft Learn page opened and reviewed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-234", "Company": "SLB", "Company_primary": "SLB", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "AI extraction of shipping documents for SLB export-duty recovery", "Extracted_finding": "SLB processes partner shipping documents for tools and parts across multiple countries to associate shipment details with SAP orders and entities and support export-duty recovery.", "Stage_at_disclosure": "Automated document workflow is described as implemented and processing received documents. Exact go-live date, total volume, and country coverage are not disclosed.", "Reported_metric": "Microsoft page reports $10 million recovered in export duties; workflow description says it processes received documents. The source does not disclose the measurement period, baseline, or direct attribution method.", "Interpretation_limit": "Microsoft-hosted customer story, not independently validated. The $10 million figure has no stated period, reconciliation, or counterfactual. The exact publication date is not explicit beyond page last-updated 2025-08-08; no launch date or throughput/accuracy is disclosed. This is a distinct shipping-document/export-duty workflow, not the airline invoice GST bot also described on the same page.", "Source_IDs": "SW4-004", "Source_date": "2025-08-08", "Source_type": "Microsoft Learn customer case study", "Source_locator": "AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live", "Source_URL": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "Evidence_status": "Full official Microsoft Learn page opened and reviewed.", "PE_relationship": null, "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "Page last updated 2025-08-08. Deployment/go-live and $10 million measurement dates are not disclosed.", "Context": "Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.", "Implementation": "Users upload shipment documents or invoices to a SharePoint library. A weekly Power Automate flow invokes AI Builder template models to extract shipment details, updates a SharePoint list and logs, and displays extracted data/documents in Power BI. Azure App Insights monitors completion. Failed/unrecognized documents go to the process-owner team for manual processing or assignment to a new template; once enough examples are accumulated, templates are trained for future recognition. Extracted shipment data are linked to ERP orders for tax claims; source names SAP orders/entities.", "Assessment": "Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": ["Supply chain & logistics", "IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-234", "Initiative_ID": "AI-234", "Title": "AI extraction of shipping documents for SLB export-duty recovery", "Date": "2025-08-08", "Finding": "SLB processes partner shipping documents for tools and parts across multiple countries to associate shipment details with SAP orders and entities and support export-duty recovery.", "Metric": "Microsoft page reports $10 million recovered in export duties; workflow description says it processes received documents. The source does not disclose the measurement period, baseline, or direct attribution method.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2025-08-08", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-234"], "Research_search": "{\"key\": \"swarm4-ai-234\", \"packet_id\": \"swarm4-ai-234\", \"record_ids\": [\"AI-234\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"SLB\", \"title\": \"AI extraction of shipping documents for SLB export-duty recovery\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [\"Supply chain & logistics\", \"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"Automated document workflow is described as implemented and processing received documents. Exact go-live date, total volume, and country coverage are not disclosed.\", \"practical_relevance\": \"Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.\", \"business_problem\": \"\", \"workflow\": \"SLB processes partner shipping documents for tools and parts across multiple countries to associate shipment details with SAP orders and entities and support export-duty recovery.\", \"technology_data\": \"Users upload shipment documents or invoices to a SharePoint library. A weekly Power Automate flow invokes AI Builder template models to extract shipment details, updates a SharePoint list and logs, and displays extracted data/documents in Power BI. Azure App Insights monitors completion. Failed/unrecognized documents go to the process-owner team for manual processing or assignment to a new template; once enough examples are accumulated, templates are trained for future recognition. Extracted shipment data are linked to ERP orders for tax claims; source names SAP orders/entities.\", \"integration_human_role\": \"Process owners receive failure reports and manually process failures or assign a new document template; the team retrains template models when sufficient examples are accumulated.\", \"deployment_scope\": \"Automated document workflow is described as implemented and processing received documents. Exact go-live date, total volume, and country coverage are not disclosed.\", \"partners\": [\"Microsoft\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Microsoft page reports $10 million recovered in export duties; workflow description says it processes received documents. The source does not disclose the measurement period, baseline, or direct attribution method.\", \"period\": \"Microsoft Learn page last updated 2025-08-08; workflow chronology/launch dates not disclosed\", \"attribution\": \"Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.\", \"source_ids\": [\"SW4-004\"]}], \"economics\": \"Reported recovery is $10 million of export duties, not necessarily recurring annual savings. No implementation/software cost or payback disclosed.\", \"timeline\": [{\"date\": \"2025-08-08\", \"event\": \"Page last updated 2025-08-08. Deployment/go-live and $10 million measurement dates are not disclosed.\", \"source_ids\": [\"SW4-004\"]}], \"limitations\": [\"Microsoft-hosted customer story, not independently validated. The $10 million figure has no stated period, reconciliation, or counterfactual. The exact publication date is not explicit beyond page last-updated 2025-08-08; no launch date or throughput/accuracy is disclosed. This is a distinct shipping-document/export-duty workflow, not the airline invoice GST bot also described on the same page.\"], \"assessment\": \"Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.\", \"sources\": [{\"title\": \"SLB enhances productivity with Power Platform and AI\", \"url\": \"https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity\", \"date\": \"2025-08-08\", \"type\": \"Microsoft Learn customer case study\", \"locator\": \"AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live\", \"review_status\": \"Full official Microsoft Learn page opened and reviewed.\", \"publisher\": \"Microsoft\", \"id\": \"SW4-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-004\"], \"field_sources\": {\"workflow\": [\"SW4-004\"], \"technology_data\": [\"SW4-004\"], \"deployment_scope\": [\"SW4-004\"], \"practical_relevance\": [\"SW4-004\"]}, \"verification\": \"verified partner\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Global SLB facilities / global workforce; countries or rollout coverage not enumerated", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full official Microsoft Learn page opened and reviewed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-235", "Company": "ExxonMobil / XTO Energy", "Company_primary": "ExxonMobil", "Segment": "E&P", "Theme": "Operational AI", "Initiative": "Closed-loop machine-learning gas-lift optimization across 1,300+ Permian wells", "Extracted_finding": "The well-by-well workflow remotely changes gas-lift injection-rate setpoints, performs multirate tests, automatically acquires well data, and iteratively uses the test response to select a production-optimizing setting. The abstract says a machine-learning model supplies recommendations on wells without downhole pressure gauges, while production data are quality-checked and uplift is model-tracked using productivity and steady-state tests.", "Stage_at_disclosure": "Scaled. JPT/SPE highlight dated 2025-10-01 reports closed-loop ML gas-lift optimization deployed across more than 1,300 ExxonMobil/XTO Permian wells.", "Reported_metric": "Operator-authored SPE abstract reports about 2% average oil uplift. JPT’s October 2024 interview reports 1.87% in a three-month human-approval stage and about 2.2% in the next three-month, mostly automatic stage. These summaries are not additive; periods and cohorts are incompletely disclosed.", "Interpretation_limit": "The primary paper's accessible abstract states the full paper is unavailable, so full technical validation is not possible. The operator-reported uplift lacks a disclosed counterfactual or confidence interval. Gas-lift performance can drift when water cut or productivity changes; JPT says an existing setting may remain suboptimal for years without frequent monitoring. The control-boundary details and automatic-versus-human exception rates are undisclosed.", "Source_IDs": "SW4-005; SW4-006; SW3-006", "Source_date": "2025-10-01", "Source_type": "SPE Journal of Petroleum Technology paper highlight", "Source_locator": "Accessible abstract and first paragraph; lines 592–620", "Source_URL": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "Evidence_status": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "PE_relationship": null, "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "SPE 219553 was presented in August 2024. JPT reported the staged trial in October 2024 and published a deployment highlight in October 2025. Historical results are not newly measured 2025 outcomes.", "Context": "This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.", "Implementation": "SPE/JPT describes an operating closed loop deployed on more than 1,300 Permian wells. JPT says this required no surface or downhole equipment changes and no additional personnel work for the repeated monthly optimization cycle. No model vendor or project-level software/integration cost is identified. The SPE abstract notes minimal disturbance to surface facilities.", "Assessment": "This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.", "Primary_department": "Production & field operations", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-022"], "Disclosures": [{"ID": "SW4-D-AI-235", "Initiative_ID": "AI-235", "Title": "Closed-loop machine-learning gas-lift optimization across 1,300+ Permian wells", "Date": "2025-10-01", "Finding": "The well-by-well workflow remotely changes gas-lift injection-rate setpoints, performs multirate tests, automatically acquires well data, and iteratively uses the test response to select a production-optimizing setting. The abstract says a machine-learning model supplies recommendations on wells without downhole pressure gauges, while production data are quality-checked and uplift is model-tracked using productivity and steady-state tests.", "Metric": "Operator-authored SPE abstract reports about 2% average oil uplift. JPT’s October 2024 interview reports 1.87% in a three-month human-approval stage and about 2.2% in the next three-month, mostly automatic stage. These summaries are not additive; periods and cohorts are incompletely disclosed.", "Source_IDs": "SW4-005; SW4-006; SW3-006", "Evidence": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2025-10-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-235"], "Research_search": "{\"key\": \"swarm4-ai-235\", \"packet_id\": \"swarm4-ai-235\", \"record_ids\": [\"AI-235\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-022\"], \"company\": \"ExxonMobil / XTO Energy\", \"title\": \"Closed-loop machine-learning gas-lift optimization across 1,300+ Permian wells\", \"department\": \"Production & field operations\", \"secondary_departments\": [\"Reservoir engineering & field development\"], \"category\": \"Internal adoption\", \"stage\": \"Reported deployed across >1,300 operator wells by the October 2025 JPT highlight. The specific test-stage dates, eligibility criteria, and fraction of active wells are not disclosed.\", \"practical_relevance\": \"This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.\", \"business_problem\": \"\", \"workflow\": \"The well-by-well workflow remotely changes gas-lift injection-rate setpoints, performs multirate tests, automatically acquires well data, and iteratively uses the test response to select a production-optimizing setting. The abstract says a machine-learning model supplies recommendations on wells without downhole pressure gauges, while production data are quality-checked and uplift is model-tracked using productivity and steady-state tests.\", \"technology_data\": \"SPE/JPT describes an operating closed loop deployed on more than 1,300 Permian wells. JPT says this required no surface or downhole equipment changes and no additional personnel work for the repeated monthly optimization cycle. No model vendor or project-level software/integration cost is identified. The SPE abstract notes minimal disturbance to surface facilities.\", \"integration_human_role\": \"The first three-month stage retained field-person approval or rejection of recommendations. In the following three-month stage, most recommendations were automatically approved and implemented. The sources do not specify who retains override authority, safety interlocks, exception handling, or the fraction still requiring manual review.\", \"deployment_scope\": \"Reported deployed across >1,300 operator wells by the October 2025 JPT highlight. The specific test-stage dates, eligibility criteria, and fraction of active wells are not disclosed.\", \"partners\": [\"XTO Energy\", \"ExxonMobil Global Services Company\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Operator-authored SPE abstract reports about 2% average oil uplift. JPT’s October 2024 interview reports 1.87% in a three-month human-approval stage and about 2.2% in the next three-month, mostly automatic stage. These summaries are not additive; periods and cohorts are incompletely disclosed.\", \"period\": \"The SPE case was presented in August 2024 and JPT reported two three-month stages on 2024-10-01; exact operating dates and the production-result measurement window are not specified. Treat the stage results as historical by the 2024 conference disclosure. JPT reaffirmed deployment to >1,300 Permian wells on 2025-10-01; that later publication does not establish a new measurement period.\", \"attribution\": \"ExxonMobil/XTO coauthored SPE 219553; the operator's production optimization engineer described the staged test to JPT. All performance claims are operator-reported and not independently validated.\", \"source_ids\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"]}], \"economics\": \"No license, implementation, integration, operating cost, cash benefit, or payback is disclosed. JPT describes incremental production without major expenditure and no equipment changes or added personnel work, but gives no amount; this qualitative statement is not a project cost or ROI estimate.\", \"timeline\": [{\"date\": \"2025-10-01\", \"event\": \"SPE 219553 was presented in August 2024. JPT reported the staged trial in October 2024 and published a deployment highlight in October 2025. Historical results are not newly measured 2025 outcomes.\", \"source_ids\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"]}], \"limitations\": [\"The primary paper's accessible abstract states the full paper is unavailable, so full technical validation is not possible. The operator-reported uplift lacks a disclosed counterfactual or confidence interval. Gas-lift performance can drift when water cut or productivity changes; JPT says an existing setting may remain suboptimal for years without frequent monitoring. The control-boundary details and automatic-versus-human exception rates are undisclosed.\"], \"assessment\": \"This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.\", \"sources\": [{\"title\": \"Gas Lift Optimization in the Permian Uses Machine Learning, Artificial Intelligence\", \"url\": \"https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted\", \"date\": \"2025-10-01\", \"type\": \"SPE Journal of Petroleum Technology paper highlight\", \"locator\": \"Accessible abstract and first paragraph; lines 592–620\", \"review_status\": \"Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.\", \"publisher\": \"Society of Petroleum Engineers\", \"id\": \"SW4-005\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Gas Lift Optimization in the Permian Using Machine Learning and Artificial Intelligence (SPE 219553-MS)\", \"url\": \"https://www.researchgate.net/publication/383230634_Gas_Lift_Optimization_in_the_Permian_Using_Machine_Learning_and_Artificial_Intelligence\", \"date\": \"2024-08\", \"type\": \"Operator-authored SPE conference paper abstract\", \"locator\": \"Complete abstract, paragraphs 1–4\", \"review_status\": \"Full accessible abstract read. Page explicitly says the full text is unavailable and can be requested from the authors; do not characterize the complete paper as reviewed.\", \"publisher\": \"ResearchGate-hosted abstract of an SPE paper\", \"id\": \"SW4-006\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Unsung Hero: Artificial Lift Boosts Production Without Breaking the Bank\", \"url\": \"https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted\", \"date\": \"2024-10-01\", \"type\": \"SPE Journal of Petroleum Technology feature with operator-engineer interview\", \"locator\": \"Article date at lines 592–599; ExxonMobil SPE 219553 workflow and claims at lines 625–646\", \"review_status\": \"Full relevant article section opened and read, including the two 3-month control stages, operator quotations, workflow description, and caveats. Reporting summarizes a paper and interview; it is not independent validation.\", \"publisher\": \"Society of Petroleum Engineers\", \"id\": \"SW3-006\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"], \"field_sources\": {\"workflow\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"], \"technology_data\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"], \"deployment_scope\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"], \"practical_relevance\": [\"SW4-005\", \"SW4-006\", \"SW3-006\"]}, \"verification\": \"verified primary abstract\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"ExxonMobil/XTO coauthored SPE 219553; the operator's production optimization engineer described the staged test to JPT. All performance claims are operator-reported and not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary abstract", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "ExxonMobil/XTO-operated Permian Basin wells", "Normalized_stage": "Scaled", "Stage_review": {"id": "AI-235", "stage": "Scaled", "stage_basis": "JPT/SPE highlight dated 2025-10-01 reports closed-loop ML gas-lift optimization deployed across more than 1,300 ExxonMobil/XTO Permian wells.", "source_url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "source_date": "2025-10-01", "evidence_origin": "Technical paper", "limits": "JPT highlight is member-exclusive after the opening summary. This supports reported workflow and scale, not test history, eligibility or active-well denominator; no operational outcome metric is available in the accessible text. The October 2025 highlight refers to earlier technical work; its underlying August 2024 paper predates the core window.", "source_id": "SW4-005"}, "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "1,300+ wells", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Operator-authored SPE paper summarized by JPT; ~2% average oil uplift.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary abstract", "claim_status": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-236", "Company": "Dagang Oilfield (PetroChina/CNPC)", "Company_primary": "PetroChina", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Machine-learning production forecasting across Dagang Oilfield blocks", "Extracted_finding": "A machine-learning production-forecast model was trialed in multiple Dagang Oilfield blocks to predict oilfield production.", "Stage_at_disclosure": "Trialed in multiple blocks in 2025; production-wide deployment is not established.", "Reported_metric": "CNPC article reports that forecast accuracy “improved by 15%” (预测准确率提升15%). The comparison baseline, calculation, validation set, forecast horizon, and whether this is a relative improvement or percentage-point change are unspecified; it is not an absolute 15% accuracy score.", "Interpretation_limit": "The article says Dagang created 11 AI applications in 2025; this is a portfolio count, not 11 separately evidenced profiles. The 15% accuracy statement is an operator-group-reported model-performance claim without a disclosed denominator or validation method. Adjacent figures in the article—response speed up more than 60% and plan-preparation time down an average 30%—describe broader digital transformation/platform work and are not attributed to this ML forecast model.", "Source_IDs": "SW4-007", "Source_date": "2026-02-04", "Source_type": "CNPC News Center / China Petroleum News Center operator-group news article", "Source_locator": "Article section “让科技生产力‘看得见’” (“Make technological productivity visible”), paragraph beginning “人工智能的应用体现出务实的导向” (“AI applications reflect a practical orientation”); published 2026-02-04 10:17.", "Source_URL": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml", "Evidence_status": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "PE_relationship": null, "Coverage_bucket": "Fourth targeted gap review", "Time_scope": "The article’s 2025 retrospective says the machine-learning production-forecast model was trialed in multiple blocks. It does not provide individual trial dates or say the model was rolled out beyond the trials.", "Context": "Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.", "Implementation": "Article identifies machine learning and production forecasting; input data, forecast horizon, model design, integration, and how forecasts informed operating decisions are undisclosed.", "Assessment": "Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["Production & field operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fourth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW4-D-AI-236", "Initiative_ID": "AI-236", "Title": "Machine-learning production forecasting across Dagang Oilfield blocks", "Date": "2026-02-04", "Finding": "A machine-learning production-forecast model was trialed in multiple Dagang Oilfield blocks to predict oilfield production.", "Metric": "CNPC article reports that forecast accuracy “improved by 15%” (预测准确率提升15%). The comparison baseline, calculation, validation set, forecast horizon, and whether this is a relative improvement or percentage-point change are unspecified; it is not an absolute 15% accuracy score.", "Source_IDs": "SW4-007", "Evidence": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "Origin": "Fourth research swarm"}], "Latest_disclosure": "2026-02-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm4-ai-236"], "Research_search": "{\"key\": \"swarm4-ai-236\", \"packet_id\": \"swarm4-ai-236\", \"record_ids\": [\"AI-236\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Dagang Oilfield (PetroChina/CNPC)\", \"title\": \"Machine-learning production forecasting across Dagang Oilfield blocks\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [\"Production & field operations\"], \"category\": \"Internal adoption\", \"stage\": \"Trialed in multiple blocks in 2025; production-wide deployment is not established.\", \"practical_relevance\": \"Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.\", \"business_problem\": \"\", \"workflow\": \"A machine-learning production-forecast model was trialed in multiple Dagang Oilfield blocks to predict oilfield production.\", \"technology_data\": \"Article identifies machine learning and production forecasting; input data, forecast horizon, model design, integration, and how forecasts informed operating decisions are undisclosed.\", \"integration_human_role\": \"User, review, and approval roles are not disclosed; no automated control or production-setting authority is stated.\", \"deployment_scope\": \"Trialed in multiple blocks in 2025; production-wide deployment is not established.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"CNPC article reports that forecast accuracy “improved by 15%” (预测准确率提升15%). The comparison baseline, calculation, validation set, forecast horizon, and whether this is a relative improvement or percentage-point change are unspecified; it is not an absolute 15% accuracy score.\", \"period\": \"Article published 4 February 2026; the article states the model was trialed in multiple blocks during 2025.\", \"attribution\": \"CNPC news-center report of Dagang Oilfield activity; operator-group reported, not independently validated.\", \"source_ids\": [\"SW4-007\"]}], \"economics\": \"No production uplift, avoided loss, cost, or payback is attributed to this forecast model.\", \"timeline\": [{\"date\": \"2026-02-04\", \"event\": \"The article’s 2025 retrospective says the machine-learning production-forecast model was trialed in multiple blocks. It does not provide individual trial dates or say the model was rolled out beyond the trials.\", \"source_ids\": [\"SW4-007\"]}], \"limitations\": [\"The article says Dagang created 11 AI applications in 2025; this is a portfolio count, not 11 separately evidenced profiles. The 15% accuracy statement is an operator-group-reported model-performance claim without a disclosed denominator or validation method. Adjacent figures in the article—response speed up more than 60% and plan-preparation time down an average 30%—describe broader digital transformation/platform work and are not attributed to this ML forecast model.\"], \"assessment\": \"Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.\", \"sources\": [{\"title\": \"Machine-learning production forecasting across Dagang Oilfield blocks\", \"url\": \"https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml\", \"date\": \"2026-02-04\", \"type\": \"CNPC News Center / China Petroleum News Center operator-group news article\", \"locator\": \"Article section “让科技生产力‘看得见’” (“Make technological productivity visible”), paragraph beginning “人工智能的应用体现出务实的导向” (“AI applications reflect a practical orientation”); published 2026-02-04 10:17.\", \"review_status\": \"Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.\", \"publisher\": \"CNPC News Center / China Petroleum News Center\", \"id\": \"SW4-007\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW4-007\"], \"field_sources\": {\"workflow\": [\"SW4-007\"], \"technology_data\": [\"SW4-007\"], \"deployment_scope\": [\"SW4-007\"], \"practical_relevance\": [\"SW4-007\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fourth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"CNPC news-center report of Dagang Oilfield activity; operator-group reported, not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Dagang Oilfield, China; specific blocks undisclosed.", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "CNPC (PetroChina parent) group news centre; operator-group reporting.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-237", "Company": "Repsol", "Company_primary": "Repsol", "Segment": "Integrated energy", "Theme": "Operational AI", "Initiative": "Repsol legal team Harvey rollout and adoption update", "Extracted_finding": "Repsol’s in-house legal team uses Harvey to assist legal work; the original Repsol announcement listed legal-document analysis, contract/regulation comparison, clause suggestions, translation, legal queries and document summaries. The March 2026 source establishes rollout scope and adoption-support practices, but does not add new task-level workflows.", "Stage_at_disclosure": "A dated 2026 vendor interview confirms that Repsol’s legal team evaluated, piloted and rolled out Harvey across a stated 150-person team. The sources do not establish exact go-live date, percentage adoption in March 2026, or task-level production volume.", "Reported_metric": "Harvey interview: Repsol tracks about 30 power users out of 150; approximately 6–7 meet both usage and qualitative-output definitions. These are adoption-support measures, not productivity outcomes. Repsol’s April 2024 issuer announcement estimated up to three hours per week of repetitive low-legal-content work saved per lawyer during implementation; this was an estimate, not an audited realized saving and predates the window.", "Interpretation_limit": "The new dated source is a Harvey vendor interview, not a Repsol-issued performance report. It supports rollout across 150 people but not a completed adoption percentage or validated performance result. An undated Harvey Repsol customer page makes stronger claims (“96%” adoption; “4–6 hours” saved per lawyer/week, “two years later”) but has no verifiable publication date; treat those claims as an undated supplemental lead, not a dated in-window outcome.", "Source_IDs": "SW5-001; SW5-002", "Source_date": "2026-03-30", "Source_type": "Harvey customer interview with Repsol Legal Counsel", "Source_locator": "Byline/date and rollout scope lines 62–77; legal-team size, rollout lead, training and power users.", "Source_URL": "https://www.harvey.ai/blog/harvey-power-users-people-call-me-harvey", "Evidence_status": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "Repsol announced its Harvey agreement on 2024-04-04, before the search window. A Harvey interview dated 2026-03-30 says Repsol’s 150-person legal team had completed evaluation, piloting and rollout; Javier Díez Tomillo says the team tracks about 30 power users, of whom about six or seven meet both quantitative and qualitative criteria. This is evidence of ongoing rollout/adoption by March 2026, not the date Harvey first entered production.", "Context": "Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.", "Implementation": "Harvey was evaluated against Microsoft Copilot and Repsol’s internal prototype before a pilot/rollout. Repsol’s dated interview describes a 150-person team, a Harvey Taskforce, champions and team-level use-case sharing/training. Technical integrations, model configuration and access controls are not disclosed.", "Assessment": "Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.", "Primary_department": "Land & legal", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-237", "Initiative_ID": "AI-237", "Title": "Repsol legal team Harvey rollout and adoption update", "Date": "2026-03-30", "Finding": "Repsol’s in-house legal team uses Harvey to assist legal work; the original Repsol announcement listed legal-document analysis, contract/regulation comparison, clause suggestions, translation, legal queries and document summaries. The March 2026 source establishes rollout scope and adoption-support practices, but does not add new task-level workflows.", "Metric": "Harvey interview: Repsol tracks about 30 power users out of 150; approximately 6–7 meet both usage and qualitative-output definitions. These are adoption-support measures, not productivity outcomes. Repsol’s April 2024 issuer announcement estimated up to three hours per week of repetitive low-legal-content work saved per lawyer during implementation; this was an estimate, not an audited realized saving and predates the window.", "Source_IDs": "SW5-001; SW5-002", "Evidence": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2026-03-30", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-237"], "Research_search": "{\"key\": \"swarm5-ai-237\", \"packet_id\": \"swarm5-ai-237\", \"record_ids\": [\"AI-237\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Repsol\", \"title\": \"Repsol legal team Harvey rollout and adoption update\", \"department\": \"Land & legal\", \"secondary_departments\": [\"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"A dated 2026 vendor interview confirms that Repsol’s legal team evaluated, piloted and rolled out Harvey across a stated 150-person team. The sources do not establish exact go-live date, percentage adoption in March 2026, or task-level production volume.\", \"practical_relevance\": \"Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.\", \"business_problem\": \"\", \"workflow\": \"Repsol’s in-house legal team uses Harvey to assist legal work; the original Repsol announcement listed legal-document analysis, contract/regulation comparison, clause suggestions, translation, legal queries and document summaries. The March 2026 source establishes rollout scope and adoption-support practices, but does not add new task-level workflows.\", \"technology_data\": \"Harvey was evaluated against Microsoft Copilot and Repsol’s internal prototype before a pilot/rollout. Repsol’s dated interview describes a 150-person team, a Harvey Taskforce, champions and team-level use-case sharing/training. Technical integrations, model configuration and access controls are not disclosed.\", \"integration_human_role\": \"Repsol’s 2024 issuer announcement says its legal team is ultimately responsible for reviewing/validating Harvey proposals and deciding how to apply them. The dated 2026 interview describes a Repsol taskforce/champions, adoption support, training, beta testing and peer coaching. No automatic legal decision or filing authority is reported.\", \"deployment_scope\": \"A dated 2026 vendor interview confirms that Repsol’s legal team evaluated, piloted and rolled out Harvey across a stated 150-person team. The sources do not establish exact go-live date, percentage adoption in March 2026, or task-level production volume.\", \"partners\": [\"Harvey\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Harvey interview: Repsol tracks about 30 power users out of 150; approximately 6–7 meet both usage and qualitative-output definitions. These are adoption-support measures, not productivity outcomes. Repsol’s April 2024 issuer announcement estimated up to three hours per week of repetitive low-legal-content work saved per lawyer during implementation; this was an estimate, not an audited realized saving and predates the window.\", \"period\": \"Dated Harvey interview reports rollout across Repsol’s 150-person legal team; original agreement was announced in April 2024, outside this search window. The March 2026 source is a substantive in-window status update, not the original selection date.\", \"attribution\": \"Harvey vendor interview with Repsol Legal Counsel Javier Díez Tomillo; adoption and scope are named-customer statements, not independently verified.\", \"source_ids\": [\"SW5-001\", \"SW5-002\"]}], \"economics\": \"No implementation cost, operating cost, payback, or realized financial benefit disclosed. The up-to-three-hours-per-week estimate in the 2024 announcement is not an observed or audited cash saving.\", \"timeline\": [{\"date\": \"2026-03-30\", \"event\": \"Repsol announced its Harvey agreement on 2024-04-04, before the search window. A Harvey interview dated 2026-03-30 says Repsol’s 150-person legal team had completed evaluation, piloting and rollout; Javier Díez Tomillo says the team tracks about 30 power users, of whom about six or seven meet both quantitative and qualitative criteria. This is evidence of ongoing rollout/adoption by March 2026, not the date Harvey first entered production.\", \"source_ids\": [\"SW5-001\", \"SW5-002\"]}], \"limitations\": [\"The new dated source is a Harvey vendor interview, not a Repsol-issued performance report. It supports rollout across 150 people but not a completed adoption percentage or validated performance result. An undated Harvey Repsol customer page makes stronger claims (“96%” adoption; “4–6 hours” saved per lawyer/week, “two years later”) but has no verifiable publication date; treat those claims as an undated supplemental lead, not a dated in-window outcome.\"], \"assessment\": \"Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.\", \"sources\": [{\"title\": \"Harvey Power Users: “People Call Me Harvey”\", \"url\": \"https://www.harvey.ai/blog/harvey-power-users-people-call-me-harvey\", \"date\": \"2026-03-30\", \"type\": \"Harvey customer interview with Repsol Legal Counsel\", \"locator\": \"Byline/date and rollout scope lines 62–77; legal-team size, rollout lead, training and power users.\", \"review_status\": \"Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.\", \"publisher\": \"Harvey\", \"id\": \"SW5-001\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Repsol becomes the first Spanish in-house legal team to use Harvey\", \"url\": \"https://www.repsol.com/en/press-room/press-releases/2024/repsol-becomes-the-first-spanish-in-house-legal-team-to-use-harvey/index.cshtml\", \"date\": \"2024-04-04 (pre-window context only)\", \"type\": \"Repsol issuer press release\", \"locator\": \"Press release date and scope lines 675–689: legal document analysis, contract/regulation comparison, clause suggestions, translation, legal queries, summaries; human validation; estimated up to three hours/week saved per lawyer.\", \"review_status\": \"Full issuer page opened and read. Pre-window agreement and estimated time savings used only as history; not counted as an in-window deployment or measured outcome.\", \"publisher\": \"Repsol\", \"id\": \"SW5-002\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-001\", \"SW5-002\"], \"field_sources\": {\"workflow\": [\"SW5-001\", \"SW5-002\"], \"technology_data\": [\"SW5-001\", \"SW5-002\"], \"deployment_scope\": [\"SW5-001\", \"SW5-002\"], \"practical_relevance\": [\"SW5-001\", \"SW5-002\"]}, \"verification\": \"verified partner\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Harvey vendor interview with Repsol Legal Counsel Javier Díez Tomillo; adoption and scope are named-customer statements, not independently verified.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Repsol Legal Division; original issuer disclosure described more than 200 legal professionals worldwide. The March 2026 interview describes a 150-person rollout population; relationship between historical and current headcount is not reconciled.", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-238", "Company": "Spearpoint Resources", "Company_primary": "Spearpoint Resources", "Segment": "Upstream (non-operated interests)", "Theme": "Operational AI", "Initiative": "AI-assisted non-op revenue statement extraction and WolfePak export at Spearpoint Resources", "Extracted_finding": "Spearpoint’s small accounting team processes monthly non-operated revenue statements arriving in 27 purchaser formats. Joltly extracts structured line items at well/lease level—including production dates, volumes, prices, interests, taxes and deductions—then users review exceptions and export the results for WolfePak posting.", "Stage_at_disclosure": "In production since May 2026 per the vendor case study; output is exported in WolfePak-compatible format. The source does not say that Joltly posts transactions autonomously.", "Reported_metric": "Spearpoint Controller Elle Ross estimates Joltly gives the team 20–30 hours per week back across three entities. From May through August 2026 the case reports 82 statements processed, 5,441 pages and 56,358 line items across 27 purchaser formats. Time estimate is customer-reported; no baseline time study, quality/exception rate or independent validation is supplied.", "Interpretation_limit": "Vendor-authored case study; no audited results, accuracy/error rate, control testing, implementation cost, customer-specific pricing, or independent attribution. The output is reviewed/exported by people; do not describe it as fully autonomous posting. It is a named but small non-operated asset owner described without basin or asset-level details.", "Source_IDs": "SW5-003; OCT5-S053; OCT5-S054", "Source_date": "2026-09-05", "Source_type": "Joltly customer case study quoting Spearpoint controller", "Source_locator": "Date lines 9–14; workflow/architecture lines 24–47 and 130–154; May–August 2026 volumes lines 49–78; outcome quote lines 156–170.", "Source_URL": "https://www.joltly.io/blog/spear-point-revenue-accounting-case-study", "Evidence_status": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "Joltly says Spearpoint has been live since May 2026. The case study published 2026-09-05 reports data through end-August 2026: 82 statements, 5,441 pages, 56,358 line items across three entities and 27 purchaser formats.", "Context": "Concrete in-window named operator deployment in oil-and-gas revenue accounting; addresses purchaser-format variation and extreme statement-size variability with attributed staffing-capacity metrics and explicit human exception review.", "Implementation": "Upload or email-forward PDF statements without purchaser-specific templates. Joltly extracts each line item and retains coordinates linking figures to source-page locations. The accounting team reviews exceptions and exports in the format WolfePak expects. The case does not disclose models, training regimen, integration/API method, exception thresholds or posting automation.", "Assessment": "Concrete in-window named operator deployment in oil-and-gas revenue accounting; addresses purchaser-format variation and extreme statement-size variability with attributed staffing-capacity metrics and explicit human exception review.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-238", "Initiative_ID": "AI-238", "Title": "AI-assisted non-op revenue statement extraction and WolfePak export at Spearpoint Resources", "Date": "2026-09-05", "Finding": "Spearpoint’s small accounting team processes monthly non-operated revenue statements arriving in 27 purchaser formats. Joltly extracts structured line items at well/lease level—including production dates, volumes, prices, interests, taxes and deductions—then users review exceptions and export the results for WolfePak posting.", "Metric": "Spearpoint Controller Elle Ross estimates Joltly gives the team 20–30 hours per week back across three entities. From May through August 2026 the case reports 82 statements processed, 5,441 pages and 56,358 line items across 27 purchaser formats. Time estimate is customer-reported; no baseline time study, quality/exception rate or independent validation is supplied.", "Source_IDs": "SW5-003", "Evidence": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "Origin": "Fifth research swarm"}, {"ID": "OCT5-AI-238-1", "Initiative_ID": "AI-238", "Title": "Joltly Pricing — Operators & Non-Op / Royalty Owners", "Date": "Undated; accessed 2026-10-06", "Finding": "Joltly describes a monthly platform fee plus usage tied to documents, ACH payments, mailed checks and enabled workflows, with no seat fees. Pricing cards show $750 Independent Operator starter and $1,500 Midsize Operator under a “Monthly Contracts” label. Integrations include direct QuickBooks and Quorum On-Demand connections and file exports for PakEnergy and Integra. The vendor claims AP can go live in under one week and revenue in about two weeks with a sample statement.", "Metric": "Public pricing cards: $750 and $1,500 monthly plan amounts; included volume/modules and usage costs are undefined. These are not Spearpoint contract terms.", "Source_IDs": "OCT5-S053", "Source_URL": "https://www.joltly.io/pricing", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-238-2", "Initiative_ID": "AI-238", "Title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "Date": "2026-08-08", "Finding": "Joltly describes extraction of line-item JIBs and purchaser revenue statements, well/account mapping, exception flagging, line-level accounting output, deposit matching and posting revenue by well/product. This is vendor workflow explanation, not evidence of Spearpoint-specific control implementation.", "Metric": "No customer-specific additional result in this workflow article.", "Source_IDs": "OCT5-S054", "Source_URL": "https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}], "Latest_disclosure": "Undated; accessed 2026-10-06", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai238-joltly-context", "swarm5-ai-238"], "Research_search": "{\"packet_id\": \"oct6-executive-ai238-joltly-context\", \"review_date\": \"2026-10-06\", \"source_title\": \"Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month\", \"publisher\": \"Joltly\", \"publisher_note\": \"Source publisher is Joltly; do not substitute Spearpoint (the customer) as publisher.\", \"source_date\": \"2026-08-08\", \"source_type\": \"First-party vendor workflow article\", \"source_locator\": \"Sections describing JIB/revenue-statement processing and exceptions\", \"source_review\": \"Full original vendor pages reviewed on 2026-10-06.\", \"workflow\": \"Joltly describes line-item JIB and purchaser-statement extraction, well/account mapping, exception flagging, line-level accounting output, deposit matching and posting revenue by well/product. Pricing page adds monthly platform plus usage-based fee structure and published operator plan cards.\", \"technology_data\": \"Offer-side product descriptions only; no independently validated accuracy or customer-specific controls.\", \"deployment_scope\": \"Context for existing Spearpoint customer case AI-238 only; no new adoption record.\", \"practical_relevance\": \"Enables buy-versus-build and cost diligence with real document/payment volume, contract terms, exception effort and close controls.\", \"business_problem\": \"JIB and purchaser revenue-statement processing.\", \"economics\": \"Joltly public pricing cards show $750/$1,500 monthly plan prices and usage-based fees; these are not Spearpoint contract prices. Customer implementation and recurring costs remain undisclosed.\", \"result_label\": \"Vendor offer context, not customer contract or outcome\", \"attribution\": \"Vendor\", \"verification\": \"verified partner\", \"limitations\": [\"Do not attribute $750/$1,500 pricing, included volume or implementation timing to Spearpoint.\", \"Vendor workflow article is capability description, not observed Spearpoint controls or independent accuracy study.\", \"Existing profile’s outputs are reviewed/exported by people; do not call this autonomous posting.\"], \"timeline_date\": \"2026-10-06\", \"timeline_event\": \"Offer-side context reviewed 2026-10-06; does not change the existing customer implementation timeline.\", \"supplementary_source_urls\": [\"https://www.joltly.io/pricing\", \"https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas\"], \"title\": \"Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month\", \"stage\": \"In production since May 2026 per the vendor case study; output is exported in WolfePak-compatible format. The source does not say that Joltly posts transactions autonomously.\", \"key\": \"oct6-executive-ai238-joltly-context\", \"record_ids\": [\"AI-238\"], \"company\": \"Spearpoint Resources\", \"category\": \"Internal adoption\", \"department\": \"Finance, investment & strategy\", \"source_ids\": [\"OCT5-S054\", \"OCT5-S053\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S054\"], \"workflow\": [\"OCT5-S054\"], \"technology_data\": [\"OCT5-S054\"], \"deployment_scope\": [\"OCT5-S054\"], \"practical_relevance\": [\"OCT5-S054\"], \"economics\": [\"OCT5-S054\"], \"integration_human_role\": [\"OCT5-S054\"]}, \"disclosed_results\": [{\"metric\": \"Vendor offer context, not customer contract or outcome\", \"value\": \"No additional customer result from these offer-side pages. Existing AI-238 reports 20–30 hours/week returned by the customer; do not convert time into cash savings.\", \"period\": \"2026-08-08\", \"attribution\": \"Vendor\", \"source_ids\": [\"OCT5-S054\"]}], \"timeline\": [{\"date\": \"2026-10-06\", \"event\": \"Offer-side context reviewed 2026-10-06; does not change the existing customer implementation timeline.\", \"source_ids\": [\"OCT5-S054\"]}], \"sources\": [{\"id\": \"OCT5-S054\", \"title\": \"Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month\", \"url\": \"https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas\", \"date\": \"2026-08-08\", \"type\": \"First-party vendor workflow article\", \"publisher\": \"Joltly\", \"locator\": \"Sections describing JIB/revenue-statement processing and exceptions\", \"review_status\": \"Full original vendor pages reviewed on 2026-10-06.\"}, {\"id\": \"OCT5-S053\", \"title\": \"Joltly Pricing — Operators & Non-Op / Royalty Owners\", \"url\": \"https://www.joltly.io/pricing\", \"date\": \"Undated; accessed 2026-10-06\", \"type\": \"First-party vendor offer page\", \"publisher\": \"Joltly\", \"locator\": \"Pricing cards and FAQ: monthly platform fee, usage fees, integrations and implementation claims\", \"review_status\": \"full text reviewed\"}]}", "Research_status": "Updated · October gap review · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Spearpoint Resources; company location and basins not specified. The case says three entities hold non-op assets across multiple operators and basins.", "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-239", "Company": "INPEX Corporation", "Company_primary": "INPEX", "Segment": "Upstream & CCS", "Theme": "Operational AI", "Initiative": "Machine-learning surrogate accelerates INPEX CCS reservoir simulation", "Extracted_finding": "A high-speed machine-learning surrogate model provides rapid predictions of pressure, gas saturation and other reservoir parameters from geological and operating conditions, supporting reservoir-efficiency and pressure-rise evaluation plus optimization and sensitivity analysis for CCS decision-making.", "Stage_at_disclosure": "Model development stated. Project-level deployment or operational use is not established.", "Reported_metric": "The stated benefit is reduced simulation time and the ability to evaluate parameters quickly; no measured speedup, accuracy figure, validation result or project result is quantified.", "Interpretation_limit": "This is an engineering decision-support model, not evidence of AI control of an operating CCS asset. No specific field, injection project, go-live date, user role, validation method, accuracy, quantified time reduction or economic result is reported.", "Source_IDs": "SW5-004; SW5-005; SW5-006", "Source_date": "2025-07-04", "Source_type": "Operator integrated report", "Source_locator": "Printed p. 41 (PDF physical page 42), “Thorough Utilization of Digital Technology,” Case example 2: “Accelerating CCS simulation using machine learning”; official integrated-report index dates publication 4 July 2025.", "Source_URL": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf", "Evidence_status": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "The July 2025 report states INPEX had developed the surrogate model. The July 2026 report Growth Strategy/Digital Strategy section does not mention or update this CCS surrogate model. It gives no development date, CCS project deployment date, or evidence it was used in field operations.", "Context": "Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.", "Implementation": "INPEX says it developed the surrogate model to reduce the time required for CCS reservoir simulation. The report does not disclose model architecture, training data volume, simulator integration, validation set, or how technical teams review or approve its outputs.", "Assessment": "Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.", "Primary_department": "Reservoir engineering & field development", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Research & development", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-239", "Initiative_ID": "AI-239", "Title": "Machine-learning surrogate accelerates INPEX CCS reservoir simulation", "Date": "2025-07-04", "Finding": "A high-speed machine-learning surrogate model provides rapid predictions of pressure, gas saturation and other reservoir parameters from geological and operating conditions, supporting reservoir-efficiency and pressure-rise evaluation plus optimization and sensitivity analysis for CCS decision-making.", "Metric": "The stated benefit is reduced simulation time and the ability to evaluate parameters quickly; no measured speedup, accuracy figure, validation result or project result is quantified.", "Source_IDs": "SW5-004; SW5-005; SW5-006", "Evidence": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2025-07-04", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-239"], "Research_search": "{\"key\": \"swarm5-ai-239\", \"packet_id\": \"swarm5-ai-239\", \"record_ids\": [\"AI-239\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"INPEX Corporation\", \"title\": \"Machine-learning surrogate accelerates INPEX CCS reservoir simulation\", \"department\": \"Reservoir engineering & field development\", \"secondary_departments\": [\"IT, data & enterprise transformation\"], \"category\": \"Research & development\", \"stage\": \"Model development stated. Project-level deployment or operational use is not established.\", \"practical_relevance\": \"Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.\", \"business_problem\": \"\", \"workflow\": \"A high-speed machine-learning surrogate model provides rapid predictions of pressure, gas saturation and other reservoir parameters from geological and operating conditions, supporting reservoir-efficiency and pressure-rise evaluation plus optimization and sensitivity analysis for CCS decision-making.\", \"technology_data\": \"INPEX says it developed the surrogate model to reduce the time required for CCS reservoir simulation. The report does not disclose model architecture, training data volume, simulator integration, validation set, or how technical teams review or approve its outputs.\", \"integration_human_role\": \"The report says the analyses support decision-making; specific users, review steps and approval authority are not identified. No automated control is described.\", \"deployment_scope\": \"Model development stated. Project-level deployment or operational use is not established.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"The stated benefit is reduced simulation time and the ability to evaluate parameters quickly; no measured speedup, accuracy figure, validation result or project result is quantified.\", \"period\": \"Disclosure in Integrated Report 2024, published 4 July 2025; development and operational dates not specified.\", \"attribution\": \"Operator report; described model capability is not independently validated.\", \"source_ids\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"]}], \"economics\": \"No CCS project cost, saved engineering hours, capital avoided or payback disclosed.\", \"timeline\": [{\"date\": \"2025-07-04\", \"event\": \"The July 2025 report states INPEX had developed the surrogate model. The July 2026 report Growth Strategy/Digital Strategy section does not mention or update this CCS surrogate model. It gives no development date, CCS project deployment date, or evidence it was used in field operations.\", \"source_ids\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"]}], \"limitations\": [\"This is an engineering decision-support model, not evidence of AI control of an operating CCS asset. No specific field, injection project, go-live date, user role, validation method, accuracy, quantified time reduction or economic result is reported.\"], \"assessment\": \"Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.\", \"sources\": [{\"title\": \"INPEX Integrated Report 2024\", \"url\": \"https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf\", \"date\": \"2025-07-04\", \"type\": \"Operator integrated report\", \"locator\": \"Printed p. 41 (PDF physical page 42), “Thorough Utilization of Digital Technology,” Case example 2: “Accelerating CCS simulation using machine learning”; official integrated-report index dates publication 4 July 2025.\", \"review_status\": \"Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.\", \"publisher\": \"INPEX Corporation\", \"id\": \"SW5-004\", \"accessed\": \"2026-10-01\"}, {\"title\": \"INPEX Integrated Report 2025 — Growth Strategy section\", \"url\": \"https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf\", \"date\": \"2026-07-07\", \"type\": \"Operator integrated report (separate Growth Strategy PDF)\", \"locator\": \"Printed pp. 39–42 (PDF pp. 40–43), Technical Strategies and Digital Strategy sections; CCS surrogate was not discussed in the section reviewed.\", \"review_status\": \"Original 21-page issuer Growth Strategy PDF opened and reviewed, including full Digital Strategy section. It reports a separate deployed Ichthys O&M AI case and does not update or quantify the CCS surrogate model.\", \"publisher\": \"INPEX Corporation\", \"id\": \"SW5-005\", \"accessed\": \"2026-10-01\"}, {\"title\": \"INPEX Integrated Report index — publication date confirmation\", \"url\": \"https://www.inpex.com/english/ir/library/integrated_report/\", \"date\": \"Undated live index; accessed 2026-10-01\", \"type\": \"Issuer report index\", \"locator\": \"Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.\", \"review_status\": \"Full issuer index page opened and publication dates read.\", \"publisher\": \"INPEX Corporation\", \"id\": \"SW5-006\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"], \"field_sources\": {\"workflow\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"], \"technology_data\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"], \"deployment_scope\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"], \"practical_relevance\": [\"SW5-004\", \"SW5-005\", \"SW5-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator report; described model capability is not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Research & development", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "CCS reservoir simulation; specific project/site undisclosed.", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-240", "Company": "INPEX Corporation", "Company_primary": "INPEX", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Ichthys AI tool searches operational reports and structures maintenance history", "Extracted_finding": "An Ichthys-specific generative-AI tool converts daily reports into a searchable database and answers natural-language questions with past operating performance and troubleshooting cases. A connected function structures maintenance history by categorizing work descriptions, failure details and affected components.", "Stage_at_disclosure": "Developed and deployed at Ichthys; specific go-live date and adoption scale are not disclosed.", "Reported_metric": "INPEX reports that searches formerly taking several hours now complete in seconds to minutes; the report graphic gives “seconds to minutes.” It says AI structured approximately 110,000 maintenance records and achieved over 90% classification accuracy. No accuracy metric definition, validation sample, baseline method, error rate or time-study method is provided.", "Interpretation_limit": "Operator-reported implementation and process metrics. The reported >90% concerns classification accuracy, not overall answer accuracy. Faster information retrieval and record structuring are not quantified as downtime or production improvements. Engineer approval of search quality does not establish mandatory human review of every response or maintenance action.", "Source_IDs": "SW5-005; SW5-006", "Source_date": "2026-07-07", "Source_type": "Operator integrated report (separate Growth Strategy PDF)", "Source_locator": "Printed p. 40 (PDF physical page 17 of the 21-page Growth Strategy extract), “Technical Strategies,” Ichthys LNG O&M AI case: development and deployment, report retrieval and maintenance-record structuring; official report index confirms report publication 7 July 2026.", "Source_URL": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "Evidence_status": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "The July 2026 report says the Technical Headquarters developed and deployed an Ichthys-dedicated AI tool; the deployment/go-live date is not stated. The underlying operation reports cover the period since Ichthys began operating in 2018.", "Context": "A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.", "Implementation": "INPEX says the Technical Headquarters tailored and deployed the tool to site needs. It draws on Ichthys operation reports and maintenance records, including approximately 110,000 previously unstructured maintenance records. Model/provider, architecture, integration details, active user count and records inclusion dates are undisclosed.", "Assessment": "A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-240", "Initiative_ID": "AI-240", "Title": "Ichthys AI tool searches operational reports and structures maintenance history", "Date": "2026-07-07", "Finding": "An Ichthys-specific generative-AI tool converts daily reports into a searchable database and answers natural-language questions with past operating performance and troubleshooting cases. A connected function structures maintenance history by categorizing work descriptions, failure details and affected components.", "Metric": "INPEX reports that searches formerly taking several hours now complete in seconds to minutes; the report graphic gives “seconds to minutes.” It says AI structured approximately 110,000 maintenance records and achieved over 90% classification accuracy. No accuracy metric definition, validation sample, baseline method, error rate or time-study method is provided.", "Source_IDs": "SW5-005; SW5-006", "Evidence": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2026-07-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-240"], "Research_search": "{\"key\": \"swarm5-ai-240\", \"packet_id\": \"swarm5-ai-240\", \"record_ids\": [\"AI-240\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"INPEX Corporation\", \"title\": \"Ichthys AI tool searches operational reports and structures maintenance history\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"IT, data & enterprise transformation\"], \"category\": \"Internal adoption\", \"stage\": \"Developed and deployed at Ichthys; specific go-live date and adoption scale are not disclosed.\", \"practical_relevance\": \"A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.\", \"business_problem\": \"\", \"workflow\": \"An Ichthys-specific generative-AI tool converts daily reports into a searchable database and answers natural-language questions with past operating performance and troubleshooting cases. A connected function structures maintenance history by categorizing work descriptions, failure details and affected components.\", \"technology_data\": \"INPEX says the Technical Headquarters tailored and deployed the tool to site needs. It draws on Ichthys operation reports and maintenance records, including approximately 110,000 previously unstructured maintenance records. Model/provider, architecture, integration details, active user count and records inclusion dates are undisclosed.\", \"integration_human_role\": \"The report says the search is “approved by engineers” and describes tool use for onsite needs; it does not specify whether every output receives human review or who authorizes follow-up maintenance actions.\", \"deployment_scope\": \"Developed and deployed at Ichthys; specific go-live date and adoption scale are not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"INPEX reports that searches formerly taking several hours now complete in seconds to minutes; the report graphic gives “seconds to minutes.” It says AI structured approximately 110,000 maintenance records and achieved over 90% classification accuracy. No accuracy metric definition, validation sample, baseline method, error rate or time-study method is provided.\", \"period\": \"2025 Integrated Report published 7 July 2026; Ichthys AI tool deployment date is not specified.\", \"attribution\": \"Operator report; throughput/quality descriptions are operator-reported, not independently validated.\", \"source_ids\": [\"SW5-005\", \"SW5-006\"]}], \"economics\": \"No cost reduction, avoided downtime, maintenance savings or payback is quantified or attributed to this tool.\", \"timeline\": [{\"date\": \"2026-07-07\", \"event\": \"The July 2026 report says the Technical Headquarters developed and deployed an Ichthys-dedicated AI tool; the deployment/go-live date is not stated. The underlying operation reports cover the period since Ichthys began operating in 2018.\", \"source_ids\": [\"SW5-005\", \"SW5-006\"]}], \"limitations\": [\"Operator-reported implementation and process metrics. The reported >90% concerns classification accuracy, not overall answer accuracy. Faster information retrieval and record structuring are not quantified as downtime or production improvements. Engineer approval of search quality does not establish mandatory human review of every response or maintenance action.\"], \"assessment\": \"A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.\", \"sources\": [{\"title\": \"INPEX Integrated Report 2025 — Growth Strategy section\", \"url\": \"https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf\", \"date\": \"2026-07-07\", \"type\": \"Operator integrated report (separate Growth Strategy PDF)\", \"locator\": \"Printed p. 40 (PDF physical page 17 of the 21-page Growth Strategy extract), “Technical Strategies,” Ichthys LNG O&M AI case: development and deployment, report retrieval and maintenance-record structuring; official report index confirms report publication 7 July 2026.\", \"review_status\": \"Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.\", \"publisher\": \"INPEX Corporation\", \"id\": \"SW5-005\", \"accessed\": \"2026-10-01\"}, {\"title\": \"INPEX Integrated Report index — publication date confirmation\", \"url\": \"https://www.inpex.com/english/ir/library/integrated_report/\", \"date\": \"Undated live index; accessed 2026-10-01\", \"type\": \"Issuer report index\", \"locator\": \"Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.\", \"review_status\": \"Full issuer index page opened and publication dates read.\", \"publisher\": \"INPEX Corporation\", \"id\": \"SW5-006\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-005\", \"SW5-006\"], \"field_sources\": {\"workflow\": [\"SW5-005\", \"SW5-006\"], \"technology_data\": [\"SW5-005\", \"SW5-006\"], \"deployment_scope\": [\"SW5-005\", \"SW5-006\"], \"practical_relevance\": [\"SW5-005\", \"SW5-006\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Operator report; throughput/quality descriptions are operator-reported, not independently validated.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Ichthys LNG Project, Australia.", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-241", "Company": "Saipem", "Company_primary": "Saipem", "Segment": "Oilfield services / offshore production", "Theme": "Operational AI", "Initiative": "AI video safety alerts with human verification", "Extracted_finding": "Cameras identify selected unsafe conditions and notify trained safety personnel. Karimun examples include missing PPE and unsafe proximity to machinery.", "Stage_at_disclosure": "Company reports video analytics at 11 operational sites in 2025, after additions on two vessels and two yards. A December 2025 Livorno installation agreement is separate from completed installation evidence.", "Reported_metric": "11 operational sites reported for 2025. No attributable incident reduction or false-positive rate disclosed.", "Interpretation_limit": "Notices cover different entities/dates. The March 2025 vessel notice has inconsistent subsidiary naming; use it for the named vessel and stated controls only. No deployment-wide accuracy or causal safety result.", "Source_IDs": "SW5-007; SW5-008; SW5-009; SW5-010; SW5-011", "Source_date": "2026-05-20", "Source_type": "Issuer disclosure", "Source_locator": "Dated announcement and safety technology discussion", "Source_URL": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "Evidence_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "Vessel notice: 31 March 2025; corporate notice: 28 January 2026; sustainability release: 20 May 2026.", "Context": "Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.", "Implementation": "The January 2026 notice describes local software, a manufacturer-provided model and retraining if false alerts warrant it. Routine video is retained for 24 hours; incident and training retention follow separate rules. The March 2025 vessel notice separately specifies validation on unused data.", "Assessment": "Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-241", "Initiative_ID": "AI-241", "Title": "AI video safety alerts with human verification", "Date": "2026-05-20", "Finding": "Cameras identify selected unsafe conditions and notify trained safety personnel. Karimun examples include missing PPE and unsafe proximity to machinery.", "Metric": "11 operational sites reported for 2025. No attributable incident reduction or false-positive rate disclosed.", "Source_IDs": "SW5-007; SW5-008; SW5-009; SW5-010; SW5-011", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2026-05-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-241"], "Research_search": "{\"key\": \"swarm5-ai-241\", \"packet_id\": \"swarm5-ai-241\", \"record_ids\": [\"AI-241\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Saipem\", \"title\": \"AI video safety alerts with human verification\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Company reports video analytics at 11 operational sites in 2025, after additions on two vessels and two yards. A December 2025 Livorno installation agreement is separate from completed installation evidence.\", \"practical_relevance\": \"Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.\", \"business_problem\": \"\", \"workflow\": \"Cameras identify selected unsafe conditions and notify trained safety personnel. Karimun examples include missing PPE and unsafe proximity to machinery.\", \"technology_data\": \"The January 2026 notice describes local software, a manufacturer-provided model and retraining if false alerts warrant it. Routine video is retained for 24 hours; incident and training retention follow separate rules. The March 2025 vessel notice separately specifies validation on unused data.\", \"integration_human_role\": \"A person checks each alert and decides the response. The system supports existing supervision; it does not assume decision authority.\", \"deployment_scope\": \"Company reports video analytics at 11 operational sites in 2025, after additions on two vessels and two yards. A December 2025 Livorno installation agreement is separate from completed installation evidence.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"11 operational sites reported for 2025. No attributable incident reduction or false-positive rate disclosed.\", \"period\": \"Disclosure status; exact operating dates not stated.\", \"attribution\": \"Company-reported scope; no independent performance validation.\", \"source_ids\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"]}], \"economics\": \"No project cost, monetized benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2026-05-20\", \"event\": \"Vessel notice: 31 March 2025; corporate notice: 28 January 2026; sustainability release: 20 May 2026.\", \"source_ids\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"]}], \"limitations\": [\"Notices cover different entities/dates. The March 2025 vessel notice has inconsistent subsidiary naming; use it for the named vessel and stated controls only. No deployment-wide accuracy or causal safety result.\"], \"assessment\": \"Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.\", \"sources\": [{\"title\": \"Saipem publishes 2025 sustainability results\", \"url\": \"https://www.saipem.com/en/media/news/2025-sustainability-results-published\", \"date\": \"2026-05-20\", \"locator\": \"Dated announcement and safety technology discussion\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-007\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem Smart Cameras privacy notice\", \"url\": \"https://www.saipem.com/sites/default/files/2026-01/75_Privacy_Notice_Smart_Cameras.pdf\", \"date\": \"2026-01-28\", \"locator\": \"pp. 1–2: operating model, human verification, retraining and retention; date printed on all pages\", \"type\": \"Issuer operational notice\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-008\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem MISR Smart Cameras notice — Saipem 10000\", \"url\": \"https://www.saipem.com/sites/default/files/2025-04/Information%20notice%20Smart%20Cameras_Saipem%20MISR%20for%20Petroleum%20Services%20%28S.A.E.%29_0.pdf\", \"date\": \"2025-03-31\", \"locator\": \"pp. 1–3: named vessel, validation and responsible safety personnel\", \"type\": \"Issuer operational notice\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-009\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem Sustainability Overview 2025\", \"url\": \"https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf\", \"date\": \"2026\", \"locator\": \"Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.\", \"type\": \"Issuer sustainability report\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-010\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem Consolidated Sustainability Statement 2025\", \"url\": \"https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf\", \"date\": \"2026\", \"locator\": \"Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests\", \"type\": \"Issuer annual-report extract\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-011\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"], \"field_sources\": {\"workflow\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"], \"technology_data\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"], \"deployment_scope\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"], \"practical_relevance\": [\"SW5-007\", \"SW5-008\", \"SW5-009\", \"SW5-010\", \"SW5-011\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported scope; no independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Saipem 10000 vessel; Karimun yard, Indonesia; Livorno installation agreement, Italy", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "11 sites", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-242", "Company": "Saipem", "Company_primary": "Saipem", "Segment": "Oilfield services / offshore production", "Theme": "Operational AI", "Initiative": "AI-assisted drone surveys for dropped-object hazards", "Extracted_finding": "Drone imagery of elevated structures is processed by a trained AI algorithm to support dropped-object hazard surveys.", "Stage_at_disclosure": "Operational tests reported during 2025; routine fleet-wide deployment is not established.", "Reported_metric": "No measured detection accuracy, inspection productivity or avoided-loss result.", "Interpretation_limit": "The readiness graphic is not used as an outcome or deployment percentage. Asset count and operator approval procedures remain unspecified.", "Source_IDs": "SW5-007; SW5-010; SW5-011", "Source_date": "2026-05-20", "Source_type": "Issuer disclosure", "Source_locator": "Dated announcement and safety technology discussion", "Source_URL": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "Evidence_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "2025 activity reported in a document whose publication was announced on 20 May 2026.", "Context": "Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.", "Implementation": "High-resolution cameras feed image analysis; the report places this technology in the Karimun innovation programme.", "Assessment": "Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-242", "Initiative_ID": "AI-242", "Title": "AI-assisted drone surveys for dropped-object hazards", "Date": "2026-05-20", "Finding": "Drone imagery of elevated structures is processed by a trained AI algorithm to support dropped-object hazard surveys.", "Metric": "No measured detection accuracy, inspection productivity or avoided-loss result.", "Source_IDs": "SW5-007; SW5-010; SW5-011", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2026-05-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-242"], "Research_search": "{\"key\": \"swarm5-ai-242\", \"packet_id\": \"swarm5-ai-242\", \"record_ids\": [\"AI-242\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Saipem\", \"title\": \"AI-assisted drone surveys for dropped-object hazards\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Operational tests reported during 2025; routine fleet-wide deployment is not established.\", \"practical_relevance\": \"Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.\", \"business_problem\": \"\", \"workflow\": \"Drone imagery of elevated structures is processed by a trained AI algorithm to support dropped-object hazard surveys.\", \"technology_data\": \"High-resolution cameras feed image analysis; the report places this technology in the Karimun innovation programme.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"Operational tests reported during 2025; routine fleet-wide deployment is not established.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No measured detection accuracy, inspection productivity or avoided-loss result.\", \"period\": \"Disclosure status; exact operating dates not stated.\", \"attribution\": \"Company-reported scope; no independent performance validation.\", \"source_ids\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"]}], \"economics\": \"No project cost, monetized benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2026-05-20\", \"event\": \"2025 activity reported in a document whose publication was announced on 20 May 2026.\", \"source_ids\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"]}], \"limitations\": [\"The readiness graphic is not used as an outcome or deployment percentage. Asset count and operator approval procedures remain unspecified.\"], \"assessment\": \"Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.\", \"sources\": [{\"title\": \"Saipem publishes 2025 sustainability results\", \"url\": \"https://www.saipem.com/en/media/news/2025-sustainability-results-published\", \"date\": \"2026-05-20\", \"locator\": \"Dated announcement and safety technology discussion\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-007\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem Sustainability Overview 2025\", \"url\": \"https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf\", \"date\": \"2026\", \"locator\": \"Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.\", \"type\": \"Issuer sustainability report\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-010\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Saipem Consolidated Sustainability Statement 2025\", \"url\": \"https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf\", \"date\": \"2026\", \"locator\": \"Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests\", \"type\": \"Issuer annual-report extract\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-011\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"], \"field_sources\": {\"workflow\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"], \"technology_data\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"], \"deployment_scope\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"], \"practical_relevance\": [\"SW5-007\", \"SW5-010\", \"SW5-011\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported scope; no independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Karimun innovation programme, Indonesia; intended offshore and industrial inspection applications", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-243", "Company": "MODEC", "Company_primary": "MODEC", "Segment": "Oilfield services / offshore production", "Theme": "Operational AI", "Initiative": "Shape Reef Phase II safety-barrier and gas-leak modelling", "Extracted_finding": "Develop risk models for degraded safety barriers and gas-leak scenarios on offshore units.", "Stage_at_disclosure": "A 36-month development and testing partnership, not completed Phase II deployment.", "Reported_metric": "No validated accuracy or incident-reduction result.", "Interpretation_limit": "Existing Shape Reef use in Brazil, Mexico and Africa does not establish deployment of the new Phase II models.", "Source_IDs": "SW5-012", "Source_date": "2025-08-13", "Source_type": "Issuer project announcement", "Source_locator": "Announcement paragraphs on 36-month project, Phase II and existing Shape Reef", "Source_URL": "https://www.modec.com/news/2025/20250813_pr_shape-digital.html", "Evidence_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "Partnership announced 13 August 2025; no exact project start/end supplied.", "Context": "Adds an absent FPSO operator and explicit process-safety R&D workflow.", "Implementation": "The project combines computational-fluid-dynamics and field data, building on earlier Fuzzy BowTie and neural-network leak-detection work. MODEC contributes FPSO operating expertise.", "Assessment": "Adds an absent FPSO operator and explicit process-safety R&D workflow.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Research & development", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-243", "Initiative_ID": "AI-243", "Title": "Shape Reef Phase II safety-barrier and gas-leak modelling", "Date": "2025-08-13", "Finding": "Develop risk models for degraded safety barriers and gas-leak scenarios on offshore units.", "Metric": "No validated accuracy or incident-reduction result.", "Source_IDs": "SW5-012", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2025-08-13", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-243"], "Research_search": "{\"key\": \"swarm5-ai-243\", \"packet_id\": \"swarm5-ai-243\", \"record_ids\": [\"AI-243\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"MODEC\", \"title\": \"Shape Reef Phase II safety-barrier and gas-leak modelling\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Research & development\", \"stage\": \"A 36-month development and testing partnership, not completed Phase II deployment.\", \"practical_relevance\": \"Adds an absent FPSO operator and explicit process-safety R&D workflow.\", \"business_problem\": \"\", \"workflow\": \"Develop risk models for degraded safety barriers and gas-leak scenarios on offshore units.\", \"technology_data\": \"The project combines computational-fluid-dynamics and field data, building on earlier Fuzzy BowTie and neural-network leak-detection work. MODEC contributes FPSO operating expertise.\", \"integration_human_role\": \"Not disclosed.\", \"deployment_scope\": \"A 36-month development and testing partnership, not completed Phase II deployment.\", \"partners\": [\"Shell Brasil\", \"Shape Digital\", \"Unicamp\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No validated accuracy or incident-reduction result.\", \"period\": \"Disclosure status; exact operating dates not stated.\", \"attribution\": \"Company-reported scope; no independent performance validation.\", \"source_ids\": [\"SW5-012\"]}], \"economics\": \"Financed through ANP RD&I-clause funds; amount and commercial payback are undisclosed.\", \"timeline\": [{\"date\": \"2025-08-13\", \"event\": \"Partnership announced 13 August 2025; no exact project start/end supplied.\", \"source_ids\": [\"SW5-012\"]}], \"limitations\": [\"Existing Shape Reef use in Brazil, Mexico and Africa does not establish deployment of the new Phase II models.\"], \"assessment\": \"Adds an absent FPSO operator and explicit process-safety R&D workflow.\", \"sources\": [{\"title\": \"MODEC, Shell Brasil, Shape Digital and Unicamp safety-barrier AI project\", \"url\": \"https://www.modec.com/news/2025/20250813_pr_shape-digital.html\", \"date\": \"2025-08-13\", \"locator\": \"Announcement paragraphs on 36-month project, Phase II and existing Shape Reef\", \"type\": \"Issuer project announcement\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-012\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-012\"], \"field_sources\": {\"workflow\": [\"SW5-012\"], \"technology_data\": [\"SW5-012\"], \"deployment_scope\": [\"SW5-012\"], \"practical_relevance\": [\"SW5-012\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported scope; no independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Research & development", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Brazilian FPSO context; specific units not named", "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-244", "Company": "SBM Offshore", "Company_primary": "SBM Offshore", "Segment": "Oilfield services / offshore production", "Theme": "Operational AI", "Initiative": "FPSO maintenance data foundation and planned full-field AI integration", "Extracted_finding": "Unify engineering and operating data for maintenance prioritization and reliability analysis. A linked SLB alliance aims to extend insights from subsea systems through FPSO topsides.", "Stage_at_disclosure": "Cognite platform expansion announced across the Brazilian fleet; Q3 update describes deployment to part of the overall fleet. Full-field optimization and advanced agents remain prospective.", "Reported_metric": "No AI-attributed uptime or maintenance saving. Reported overall fleet uptime is not evidence of AI impact.", "Interpretation_limit": "Fleet expansion is not proof that every asset or AI use case is live. No vessel-by-vessel rollout, acceptance tests or quantified benefit.", "Source_IDs": "SW5-013; SW5-014; SW5-015", "Source_date": "2025-11-13", "Source_type": "Issuer disclosure", "Source_locator": "Digitalization and AI section; separates platform rollout from broader alliance", "Source_URL": "https://www.sbmoffshore.com/newsroom/sbm-offshore-third-quarter-2025-trading-update/", "Evidence_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "PE_relationship": null, "Coverage_bucket": "Fifth targeted gap review", "Time_scope": "Two linked announcements on 2 October 2025; status reiterated 13 November 2025. One programme profile prevents counting the shared data foundation twice.", "Context": "Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.", "Implementation": "Cognite Data Fusion supplies the data foundation; Atlas AI is a prospective extension. The SLB alliance names OptiSite and Lumi. The October release describes 18 months of foundation work before scaling.", "Assessment": "Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["IT, data & enterprise transformation", "Production & field operations"], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Fifth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW5-D-AI-244", "Initiative_ID": "AI-244", "Title": "FPSO maintenance data foundation and planned full-field AI integration", "Date": "2025-11-13", "Finding": "Unify engineering and operating data for maintenance prioritization and reliability analysis. A linked SLB alliance aims to extend insights from subsea systems through FPSO topsides.", "Metric": "No AI-attributed uptime or maintenance saving. Reported overall fleet uptime is not evidence of AI impact.", "Source_IDs": "SW5-013; SW5-014; SW5-015", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}], "Latest_disclosure": "2025-11-13", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm5-ai-244"], "Research_search": "{\"key\": \"swarm5-ai-244\", \"packet_id\": \"swarm5-ai-244\", \"record_ids\": [\"AI-244\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"SBM Offshore\", \"title\": \"FPSO maintenance data foundation and planned full-field AI integration\", \"department\": \"Maintenance & reliability\", \"secondary_departments\": [\"IT, data & enterprise transformation\", \"Production & field operations\"], \"category\": \"Internal adoption\", \"stage\": \"Cognite platform expansion announced across the Brazilian fleet; Q3 update describes deployment to part of the overall fleet. Full-field optimization and advanced agents remain prospective.\", \"practical_relevance\": \"Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.\", \"business_problem\": \"\", \"workflow\": \"Unify engineering and operating data for maintenance prioritization and reliability analysis. A linked SLB alliance aims to extend insights from subsea systems through FPSO topsides.\", \"technology_data\": \"Cognite Data Fusion supplies the data foundation; Atlas AI is a prospective extension. The SLB alliance names OptiSite and Lumi. The October release describes 18 months of foundation work before scaling.\", \"integration_human_role\": \"Maintenance, operations and engineering teams use contextualized information. Autonomous control and approval rules are not disclosed.\", \"deployment_scope\": \"Cognite platform expansion announced across the Brazilian fleet; Q3 update describes deployment to part of the overall fleet. Full-field optimization and advanced agents remain prospective.\", \"partners\": [\"Cognite\", \"SLB\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No AI-attributed uptime or maintenance saving. Reported overall fleet uptime is not evidence of AI impact.\", \"period\": \"Disclosure status; exact operating dates not stated.\", \"attribution\": \"Company-reported scope; no independent performance validation.\", \"source_ids\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"]}], \"economics\": \"No project cost, monetized benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2025-11-13\", \"event\": \"Two linked announcements on 2 October 2025; status reiterated 13 November 2025. One programme profile prevents counting the shared data foundation twice.\", \"source_ids\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"]}], \"limitations\": [\"Fleet expansion is not proof that every asset or AI use case is live. No vessel-by-vessel rollout, acceptance tests or quantified benefit.\"], \"assessment\": \"Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.\", \"sources\": [{\"title\": \"SBM Offshore Q3 2025 trading update\", \"url\": \"https://www.sbmoffshore.com/newsroom/sbm-offshore-third-quarter-2025-trading-update/\", \"date\": \"2025-11-13\", \"locator\": \"Digitalization and AI section; separates platform rollout from broader alliance\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-013\", \"accessed\": \"2026-10-01\"}, {\"title\": \"SBM Offshore and Cognite accelerate AI and data collaboration\", \"url\": \"https://www.sbmoffshore.com/newsroom/sbm-offshore-and-cognite-partner-to-accelerate-ai-and-data-collaboration-to-optimize-floating-production-operations/\", \"date\": \"2025-10-02\", \"locator\": \"Dated release: Brazilian fleet scope, 18-month foundation work and prospective maintenance uses\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-014\", \"accessed\": \"2026-10-01\"}, {\"title\": \"SLB and SBM Offshore digital alliance\", \"url\": \"https://www.sbmoffshore.com/newsroom/slb-and-sbm-offshore-agree-digital-alliance-to-transform-fpso-production-performance/\", \"date\": \"2025-10-02\", \"locator\": \"OptiSite, Cognite Data Fusion and Lumi; planned full-field integration\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original document opened and relevant full sections read; issuer claims are not independently validated.\", \"id\": \"SW5-015\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"], \"field_sources\": {\"workflow\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"], \"technology_data\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"], \"deployment_scope\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"], \"practical_relevance\": [\"SW5-013\", \"SW5-014\", \"SW5-015\"]}, \"verification\": \"verified primary\", \"cohort\": \"Fifth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported scope; no independent performance validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Brazilian fleet for Cognite expansion; broader offshore scope for planned SLB integration", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-245", "Company": "PTTEP", "Company_primary": "PTTEP", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "AI-assisted internal audit data review", "Extracted_finding": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "Stage_at_disclosure": "Project initiated in 2025; production deployment not established.", "Reported_metric": "No measured outcome disclosed. Faster review is an intended benefit.", "Interpretation_limit": "No model, deployment scale, accuracy, approval controls or measured productivity. Do not relabel the separately described audit software as AI.", "Source_IDs": "SW6-001; SW6-002", "Source_date": "2026", "Source_type": "Operator annual report", "Source_locator": "Printed p. 223; physical PDF page 212 of downloaded 557-page file, Oversight of Internal Audit Activities.", "Source_URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Evidence_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "PE_relationship": null, "Coverage_bucket": "Sixth targeted gap review", "Time_scope": "2026 report on 2025 activities; precise publication day not established.", "Context": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "Implementation": "Internal Audit Division initiated an AI project; model and integration details are undisclosed. Adjacent TeamMate+ and dashboards are separate systems, not identified as the AI implementation.", "Assessment": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Sixth research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-067"], "Disclosures": [{"ID": "SW6-D-AI-245", "Initiative_ID": "AI-245", "Title": "AI-assisted internal audit data review", "Date": "2026", "Finding": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "Metric": "No measured outcome disclosed. Faster review is an intended benefit.", "Source_IDs": "SW6-001; SW6-002", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm6-ai-245"], "Research_search": "{\"key\": \"swarm6-ai-245\", \"packet_id\": \"swarm6-ai-245\", \"record_ids\": [\"AI-245\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-067\"], \"company\": \"PTTEP\", \"title\": \"AI-assisted internal audit data review\", \"department\": \"Finance, investment & strategy\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Project initiated in 2025; production deployment not established.\", \"practical_relevance\": \"Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.\", \"business_problem\": \"\", \"workflow\": \"Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.\", \"technology_data\": \"Internal Audit Division initiated an AI project; model and integration details are undisclosed. Adjacent TeamMate+ and dashboards are separate systems, not identified as the AI implementation.\", \"integration_human_role\": \"Internal Audit Division owns the project; output validation and final decision rights are not described.\", \"deployment_scope\": \"Project initiated in 2025; production deployment not established.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No measured outcome disclosed. Faster review is an intended benefit.\", \"period\": \"2025 operating period unless stated otherwise\", \"attribution\": \"Company-reported; no initiative-specific independent assurance established.\", \"source_ids\": [\"SW6-001\", \"SW6-002\"]}], \"economics\": \"No initiative-level cost, benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"2026 report on 2025 activities; precise publication day not established.\", \"source_ids\": [\"SW6-001\", \"SW6-002\"]}], \"limitations\": [\"No model, deployment scale, accuracy, approval controls or measured productivity. Do not relabel the separately described audit software as AI.\"], \"assessment\": \"Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.\", \"sources\": [{\"title\": \"PTTEP 56-1 One Report 2025 — internal audit\", \"url\": \"https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025\", \"date\": \"2026; referenced in invitation dated 27 February 2026\", \"locator\": \"Printed p. 223; physical PDF page 212 of downloaded 557-page file, Oversight of Internal Audit Activities.\", \"type\": \"Operator annual report\", \"review_status\": \"Original PDF downloaded and relevant full pages read; source text and page locators independently checked.\", \"id\": \"SW6-001\", \"accessed\": \"2026-10-01\"}, {\"title\": \"PTTEP 2026 AGM invitation\", \"url\": \"https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-invitation-en.pdf\", \"date\": \"2026-02-27\", \"locator\": \"Physical PDF page 2: dated invitation lists the 2025 One Report as attachment.\", \"type\": \"Issuer meeting invitation\", \"review_status\": \"Original PDF opened; date and report attachment read.\", \"id\": \"SW6-002\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW6-001\", \"SW6-002\"], \"field_sources\": {\"workflow\": [\"SW6-001\", \"SW6-002\"], \"technology_data\": [\"SW6-001\", \"SW6-002\"], \"deployment_scope\": [\"SW6-001\", \"SW6-002\"], \"practical_relevance\": [\"SW6-001\", \"SW6-002\"]}, \"verification\": \"verified primary\", \"cohort\": \"Sixth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported; no initiative-specific independent assurance established.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-246", "Company": "MOL Group", "Company_primary": "MOL Group", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "MOL MOVE AI customer-offer personalization", "Extracted_finding": "Personalize loyalty-program experiences and offers.", "Stage_at_disclosure": "Successfully tested and being scaled in 2025 AGM material covering FY2024.", "Reported_metric": "No AI-attributed outcome. Nearly five million app downloads describe the broader platform.", "Interpretation_limit": "Eight-market program reach is not eight-market AI deployment. No test population, conversion uplift, cost or model controls disclosed.", "Source_IDs": "SW6-003", "Source_date": "2025", "Source_type": "Issuer AGM management report", "Source_locator": "Printed p. 27; physical PDF page 29, MOL MOVE paragraph.", "Source_URL": "https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf", "Evidence_status": "Original PDF opened; full relevant paragraph and surrounding pages read.", "PE_relationship": null, "Coverage_bucket": "Sixth targeted gap review", "Time_scope": "Disclosure in 2025 AGM material for meeting on 24 April; underlying FY2024 review. Exact publication and go-live dates not established.", "Context": "Personalize loyalty-program experiences and offers.", "Implementation": "MOL reports testing AI and scaling the capability within Consumer Services; no model or vendor disclosed.", "Assessment": "Personalize loyalty-program experiences and offers.", "Primary_department": "Trading & marketing", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Sixth research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW6-D-AI-246", "Initiative_ID": "AI-246", "Title": "MOL MOVE AI customer-offer personalization", "Date": "2025", "Finding": "Personalize loyalty-program experiences and offers.", "Metric": "No AI-attributed outcome. Nearly five million app downloads describe the broader platform.", "Source_IDs": "SW6-003", "Evidence": "Original PDF opened; full relevant paragraph and surrounding pages read.", "Origin": "Sixth research swarm"}], "Latest_disclosure": "2025", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm6-ai-246"], "Research_search": "{\"key\": \"swarm6-ai-246\", \"packet_id\": \"swarm6-ai-246\", \"record_ids\": [\"AI-246\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"MOL Group\", \"title\": \"MOL MOVE AI customer-offer personalization\", \"department\": \"Trading & marketing\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Successfully tested and being scaled in 2025 AGM material covering FY2024.\", \"practical_relevance\": \"Personalize loyalty-program experiences and offers.\", \"business_problem\": \"\", \"workflow\": \"Personalize loyalty-program experiences and offers.\", \"technology_data\": \"MOL reports testing AI and scaling the capability within Consumer Services; no model or vendor disclosed.\", \"integration_human_role\": \"Not specified.\", \"deployment_scope\": \"Successfully tested and being scaled in 2025 AGM material covering FY2024.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No AI-attributed outcome. Nearly five million app downloads describe the broader platform.\", \"period\": \"FY2024 performance review in 2025 AGM material.\", \"attribution\": \"Company-reported; no initiative-specific independent assurance established.\", \"source_ids\": [\"SW6-003\"]}], \"economics\": \"No initiative-level cost, benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2025\", \"event\": \"Disclosure in 2025 AGM material for meeting on 24 April; underlying FY2024 review. Exact publication and go-live dates not established.\", \"source_ids\": [\"SW6-003\"]}], \"limitations\": [\"Eight-market program reach is not eight-market AI deployment. No test population, conversion uplift, cost or model controls disclosed.\"], \"assessment\": \"Personalize loyalty-program experiences and offers.\", \"sources\": [{\"title\": \"MOL 2025 AGM documents — Consumer Services\", \"url\": \"https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf\", \"date\": \"2025 AGM material; meeting 24 April 2025\", \"locator\": \"Printed p. 27; physical PDF page 29, MOL MOVE paragraph.\", \"type\": \"Issuer AGM management report\", \"review_status\": \"Original PDF opened; full relevant paragraph and surrounding pages read.\", \"id\": \"SW6-003\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW6-003\"], \"field_sources\": {\"workflow\": [\"SW6-003\"], \"technology_data\": [\"SW6-003\"], \"deployment_scope\": [\"SW6-003\"], \"practical_relevance\": [\"SW6-003\"]}, \"verification\": \"verified primary\", \"cohort\": \"Sixth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported; no initiative-specific independent assurance established.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original PDF opened; full relevant paragraph and surrounding pages read.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-247", "Company": "Ecopetrol", "Company_primary": "Ecopetrol", "Segment": "Integrated", "Theme": "Operational AI", "Initiative": "EVO talent-profile and role matching", "Extracted_finding": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "Stage_at_disclosure": "Described among three consolidated HR technology enablers for 2025; user scale not disclosed.", "Reported_metric": "No measured outcome disclosed.", "Interpretation_limit": "Do not attach adjacent SAP onboarding or Signavio results to EVO. Bias testing, privacy controls and human override are unspecified.", "Source_IDs": "SW6-004", "Source_date": "2026", "Source_type": "Operator annual report", "Source_locator": "Printed p. 229; physical PDF page 230, TEC para el talento humano / EVO.", "Source_URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Evidence_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "PE_relationship": null, "Coverage_bucket": "Sixth targeted gap review", "Time_scope": "2026 report on 2025 activities; precise publication day not established.", "Context": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "Implementation": "In-house analytics and AI tool; technical architecture undisclosed.", "Assessment": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Sixth research swarm"], "Reference_IDs": [], "Related_IDs": ["AI-147"], "Disclosures": [{"ID": "SW6-D-AI-247", "Initiative_ID": "AI-247", "Title": "EVO talent-profile and role matching", "Date": "2026", "Finding": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "Metric": "No measured outcome disclosed.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm6-ai-247"], "Research_search": "{\"key\": \"swarm6-ai-247\", \"packet_id\": \"swarm6-ai-247\", \"record_ids\": [\"AI-247\"], \"existing_ids\": [], \"related_existing_ids\": [\"AI-147\"], \"company\": \"Ecopetrol\", \"title\": \"EVO talent-profile and role matching\", \"department\": \"Workforce & learning\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Described among three consolidated HR technology enablers for 2025; user scale not disclosed.\", \"practical_relevance\": \"Match employee profiles to roles, identify skill gaps and inform succession, selection and training.\", \"business_problem\": \"\", \"workflow\": \"Match employee profiles to roles, identify skill gaps and inform succession, selection and training.\", \"technology_data\": \"In-house analytics and AI tool; technical architecture undisclosed.\", \"integration_human_role\": \"Supports people-management decisions; automated final decisions are not established.\", \"deployment_scope\": \"Described among three consolidated HR technology enablers for 2025; user scale not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No measured outcome disclosed.\", \"period\": \"2025 operating period unless stated otherwise\", \"attribution\": \"Company-reported; no initiative-specific independent assurance established.\", \"source_ids\": [\"SW6-004\"]}], \"economics\": \"No initiative-level cost, benefit or payback disclosed.\", \"timeline\": [{\"date\": \"2026\", \"event\": \"2026 report on 2025 activities; precise publication day not established.\", \"source_ids\": [\"SW6-004\"]}], \"limitations\": [\"Do not attach adjacent SAP onboarding or Signavio results to EVO. Bias testing, privacy controls and human override are unspecified.\"], \"assessment\": \"Match employee profiles to roles, identify skill gaps and inform succession, selection and training.\", \"sources\": [{\"title\": \"Ecopetrol 2025 Integrated Management Report — EVO talent map\", \"url\": \"https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf\", \"date\": \"2026 AGM report; covers 2025\", \"locator\": \"Printed p. 229; physical PDF page 230, TEC para el talento humano / EVO.\", \"type\": \"Operator annual report\", \"review_status\": \"Original PDF downloaded and relevant full pages read; source text and page locators independently checked.\", \"id\": \"SW6-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW6-004\"], \"field_sources\": {\"workflow\": [\"SW6-004\"], \"technology_data\": [\"SW6-004\"], \"deployment_scope\": [\"SW6-004\"], \"practical_relevance\": [\"SW6-004\"]}, \"verification\": \"verified primary\", \"cohort\": \"Sixth gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Company-reported; no initiative-specific independent assurance established.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated, original-source checked and separately reviewed."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-248", "Company": "Continental Resources", "Company_primary": "Continental Resources", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "Agent-assisted drilling and completion data validation", "Extracted_finding": "UiPath interprets inconsistent well-data emails; a Databricks agent validates context and scores accuracy. Robots return errors to suppliers and deliver validated data to engineers. Award material also names WellView drilling-report integration.", "Stage_at_disclosure": "Operating. Undated UiPath customer case describes Continental’s agent/robot email intake, validation and engineer data delivery as an operating workflow. The 2025-09-29 award announcement dates recognition, not go-live or case metrics.", "Reported_metric": "Undated vendor case: 200+ emails/day, approximately five minutes saved per previously manual email, one reusable robot supporting 40+ processes. Undated award page: accuracy rose from about 70% to the high 90% range.", "Interpretation_limit": "2025-09-29 dates the award announcement only. Current workflow and metric dates are unknown; 40+ processes are not 40 AI initiatives. No test-set definition, error monitoring or independently verified cash savings.", "Source_IDs": "SW7-001; SW7-002; SW7-003; SW7-004", "Source_date": "Undated case; award announced 2025-09-29", "Source_type": "Vendor customer case", "Source_locator": "Full page opened/read. Workflow and transition from human validation to automated handling at lines 241-255; quoted Sr. IT Manager Justin Cornell at 257-259; results at 260-265; 200+ emails/40+ processes/~5 minutes at 228-237.", "Source_URL": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "Evidence_status": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "Award announcement on 2025-09-29; detailed case and winners page undated. Do not date their outcomes to the award release.", "Context": "Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.", "Implementation": "UiPath Agent Builder, Robots and Databricks; WellView named in award profile. Initial Action Center validation was later removed according to the current case; remaining exception controls are undisclosed.", "Assessment": "Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Seventh research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW7-D-AI-248", "Initiative_ID": "AI-248", "Title": "Agent-assisted drilling and completion data validation", "Date": "2025-09-29", "Finding": "UiPath interprets inconsistent well-data emails; a Databricks agent validates context and scores accuracy. Robots return errors to suppliers and deliver validated data to engineers. Award material also names WellView drilling-report integration.", "Metric": "Undated vendor case: 200+ emails/day, approximately five minutes saved per previously manual email, one reusable robot supporting 40+ processes. Undated award page: accuracy rose from about 70% to the high 90% range.", "Source_IDs": "SW7-001; SW7-002; SW7-003; SW7-004", "Evidence": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2025-09-29 (award only; case undated)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-248"], "Research_search": "{\"key\": \"swarm7-ai-248\", \"packet_id\": \"swarm7-ai-248\", \"record_ids\": [\"AI-248\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Continental Resources\", \"title\": \"Agent-assisted drilling and completion data validation\", \"department\": \"Drilling & completions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Named 2025 award recipient; current undated vendor case describes an operating workflow.\", \"practical_relevance\": \"Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.\", \"business_problem\": \"\", \"workflow\": \"UiPath interprets inconsistent well-data emails; a Databricks agent validates context and scores accuracy. Robots return errors to suppliers and deliver validated data to engineers. Award material also names WellView drilling-report integration.\", \"technology_data\": \"UiPath Agent Builder, Robots and Databricks; WellView named in award profile. Initial Action Center validation was later removed according to the current case; remaining exception controls are undisclosed.\", \"integration_human_role\": \"Case describes removal of initial manual validation; ongoing escalation and quality controls not disclosed.\", \"deployment_scope\": \"Named 2025 award recipient; current undated vendor case describes an operating workflow.\", \"partners\": [\"UiPath\", \"Databricks\"], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"Undated vendor case: 200+ emails/day, approximately five minutes saved per previously manual email, one reusable robot supporting 40+ processes. Undated award page: accuracy rose from about 70% to the high 90% range.\", \"period\": \"Current metrics undated; award cohort 2025.\", \"attribution\": \"Source-reported claims; no independent validation.\", \"source_ids\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"]}], \"economics\": \"Reported time capacity, not verified labor-cost reduction. Implementation and running costs absent.\", \"timeline\": [{\"date\": \"2025-09-29\", \"event\": \"Award announcement on 2025-09-29; detailed case and winners page undated. Do not date their outcomes to the award release.\", \"source_ids\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"]}], \"limitations\": [\"2025-09-29 dates the award announcement only. Current workflow and metric dates are unknown; 40+ processes are not 40 AI initiatives. No test-set definition, error monitoring or independently verified cash savings.\"], \"assessment\": \"Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.\", \"sources\": [{\"title\": \"Continental Resources Streamlines Operations with Automation: From fragmented emails to orchestrated operations at Continental Resources\", \"url\": \"https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation\", \"date\": \"No displayed publication/update date; accessed 2026-10-01\", \"locator\": \"Full page opened/read. Workflow and transition from human validation to automated handling at lines 241-255; quoted Sr. IT Manager Justin Cornell at 257-259; results at 260-265; 200+ emails/40+ processes/~5 minutes at 228-237.\", \"review_status\": \"Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.\", \"type\": \"Vendor customer case\", \"id\": \"SW7-001\", \"accessed\": \"2026-10-01\"}, {\"title\": \"Honoring the 2025 UiPath AI25 Winners\", \"url\": \"https://www.uipath.com/ai/ai25-awards/winners\", \"date\": \"Undated page; references 2025 award cohort; accessed 2026-10-01\", \"locator\": \"Full page opened/read; Continental listing at lines 294-302, reports WellView and Databricks integration and high-90% accuracy vs ~70%, near-equal manual-intervention reduction.\", \"review_status\": \"Full page read. No publication/update or metric-measurement date displayed.\", \"type\": \"Vendor customer case\", \"id\": \"SW7-002\", \"accessed\": \"2026-10-01\"}, {\"title\": \"UiPath Announces 2025 AI25 Awards Winners\", \"url\": \"https://www.uipath.com/newsroom/uipath-announces-2025-ai25-awards-winners\", \"date\": \"2025-09-29\", \"locator\": \"Full press-release content reviewed from issuer's newsroom/search result. Names Continental among 2025 AI25 award winners. Does not disclose workflow details or performance metrics.\", \"review_status\": \"Dated primary UiPath press release reviewed. Used solely to date the award announcement, not the operational metrics.\", \"type\": \"Vendor customer case\", \"id\": \"SW7-003\", \"accessed\": \"2026-10-01\"}, {\"title\": \"UiPath + Databricks: Data Intelligence Meets Agentic Automation\", \"url\": \"https://www.uipath.com/platform/agentic-automation/ai-ecosystem/databricks\", \"date\": \"Undated page; accessed 2026-10-01\", \"type\": \"Vendor partner page\", \"locator\": \"Continental customer panel: Databricks Genie Agents and UiPath Agents for email processing.\", \"review_status\": \"Original page reviewed; no date or incremental result.\", \"id\": \"SW7-004\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"], \"field_sources\": {\"workflow\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"], \"technology_data\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"], \"deployment_scope\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"], \"practical_relevance\": [\"SW7-001\", \"SW7-002\", \"SW7-003\", \"SW7-004\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Source-reported claims; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated and original-source reviewed; specific cross-checks described in research methodology."}, "Asset_location": "Not disclosed", "Normalized_stage": "Operating", "Stage_review": {"id": "AI-248", "stage": "Operating", "stage_basis": "Undated UiPath customer case describes Continental’s agent/robot email intake, validation and engineer data delivery as an operating workflow. The 2025-09-29 award announcement dates recognition, not go-live or case metrics.", "source_url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "source_date": "Undated case; separate 2025 award announcement 2025-09-29", "evidence_origin": "Vendor", "limits": "Vendor/customer case has no deployment date or period for reported 200+ emails/day, five minutes/email, accuracy, or 40+ process re-use. Award page is undated; metrics remain vendor-published/customer-attributed.", "source_id": "SW7-001"}, "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-249", "Company": "Devon Energy", "Company_primary": "Devon Energy", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "AI alerting from SCADA and decision-support-center data for operational / methane issues", "Extracted_finding": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "Stage_at_disclosure": "AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024. Sensor-system coverage is stated at 100% of production operations collectively; AI deployment scale itself is not stated.", "Reported_metric": "No AI-attributable emissions, repair-time, uptime or financial result disclosed. The 100% number is monitoring-technology coverage, not AI coverage or emissions reduction.", "Interpretation_limit": "Deployment scope, model, alert precision/recall, response time and impact are undisclosed. Avoid conflating the continuous-monitoring work-order automation with AI decision support.", "Source_IDs": "SW7-009", "Source_date": "2025 (report; covers calendar 2024 unless noted)", "Source_type": "Issuer disclosure", "Source_locator": "printed p. 15 / PDF p. 15, lines 1228-1261; full 90-page PDF read", "Source_URL": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "Evidence_status": "Original source opened; relevant passage read.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024.", "Context": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "Implementation": "SCADA plus decision-support-center data. 2024 continuous monitoring devices covered 46% of production; integrated OGI, aerial flyovers and continuous monitoring together gave comprehensive monitoring coverage across 100% of production operations. The report does not say that AI analyzes all these sources or runs over the full covered footprint.", "Assessment": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Seventh research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW7-D-AI-249", "Initiative_ID": "AI-249", "Title": "AI alerting from SCADA and decision-support-center data for operational / methane issues", "Date": "2025 (report; covers calendar 2024 unless noted)", "Finding": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "Metric": "No AI-attributable emissions, repair-time, uptime or financial result disclosed. The 100% number is monitoring-technology coverage, not AI coverage or emissions reduction.", "Source_IDs": "SW7-009", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2025 (report; covers calendar 2024 unless noted)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-249"], "Research_search": "{\"key\": \"swarm7-ai-249\", \"packet_id\": \"swarm7-ai-249\", \"record_ids\": [\"AI-249\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Devon Energy\", \"title\": \"AI alerting from SCADA and decision-support-center data for operational / methane issues\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024. Sensor-system coverage is stated at 100% of production operations collectively; AI deployment scale itself is not stated.\", \"practical_relevance\": \"AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.\", \"business_problem\": \"\", \"workflow\": \"AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.\", \"technology_data\": \"SCADA plus decision-support-center data. 2024 continuous monitoring devices covered 46% of production; integrated OGI, aerial flyovers and continuous monitoring together gave comprehensive monitoring coverage across 100% of production operations. The report does not say that AI analyzes all these sources or runs over the full covered footprint.\", \"integration_human_role\": \"Alert prompts staff to take action. Devon separately says detected-leak work orders are automated; source does not say AI autonomously creates or closes those work orders.\", \"deployment_scope\": \"AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024. Sensor-system coverage is stated at 100% of production operations collectively; AI deployment scale itself is not stated.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No AI-attributable emissions, repair-time, uptime or financial result disclosed. The 100% number is monitoring-technology coverage, not AI coverage or emissions reduction.\", \"period\": \"Period as described in source; metric measurement period not separately disclosed.\", \"attribution\": \"Issuer-reported; no independent validation.\", \"source_ids\": [\"SW7-009\"]}], \"economics\": \"None attributed to AI.\", \"timeline\": [{\"date\": \"2025 (report; covers calendar 2024 unless noted)\", \"event\": \"AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024.\", \"source_ids\": [\"SW7-009\"]}], \"limitations\": [\"Deployment scope, model, alert precision/recall, response time and impact are undisclosed. Avoid conflating the continuous-monitoring work-order automation with AI decision support.\"], \"assessment\": \"AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.\", \"sources\": [{\"title\": \"Devon Energy 2025 Sustainability Report\", \"date\": \"2025 (report; covers calendar 2024 unless noted)\", \"url\": \"https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf\", \"locator\": \"printed p. 15 / PDF p. 15, lines 1228-1261; full 90-page PDF read\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"SW7-009\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW7-009\"], \"field_sources\": {\"workflow\": [\"SW7-009\"], \"technology_data\": [\"SW7-009\"], \"deployment_scope\": [\"SW7-009\"], \"practical_relevance\": [\"SW7-009\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Issuer-reported; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated and original-source reviewed; specific cross-checks described in research methodology."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Original source opened; relevant passage read.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-250", "Company": "Devon Energy", "Company_primary": "Devon Energy", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "Hourly computer-vision checks of flare equipment function", "Extracted_finding": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "Stage_at_disclosure": "Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated. Wellsite-camera workflow; quantitative scale not disclosed.", "Reported_metric": "No performance or emissions metric attributed to this model. Devon separately reports flaring intensity but does not attribute that outcome to AI.", "Interpretation_limit": "No model accuracy, false-alert rate, equipment coverage, avoided emissions, uptime or financial attribution. General statement about capabilities, not proof of quantified impact.", "Source_IDs": "SW7-009", "Source_date": "2025 (report; covers calendar 2024 unless noted)", "Source_type": "Issuer disclosure", "Source_locator": "printed p. 9 / PDF p. 9, lines 713-738; full 90-page PDF read", "Source_URL": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "Evidence_status": "Original source opened; relevant passage read.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated.", "Context": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "Implementation": "Company data-science analytics team developed custom AI models over repeated camera imagery. Specific vision architecture, camera counts, site count and deployment start are not reported.", "Assessment": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Seventh research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW7-D-AI-250", "Initiative_ID": "AI-250", "Title": "Hourly computer-vision checks of flare equipment function", "Date": "2025 (report; covers calendar 2024 unless noted)", "Finding": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "Metric": "No performance or emissions metric attributed to this model. Devon separately reports flaring intensity but does not attribute that outcome to AI.", "Source_IDs": "SW7-009", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2025 (report; covers calendar 2024 unless noted)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-250"], "Research_search": "{\"key\": \"swarm7-ai-250\", \"packet_id\": \"swarm7-ai-250\", \"record_ids\": [\"AI-250\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Devon Energy\", \"title\": \"Hourly computer-vision checks of flare equipment function\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated. Wellsite-camera workflow; quantitative scale not disclosed.\", \"practical_relevance\": \"Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.\", \"business_problem\": \"\", \"workflow\": \"Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.\", \"technology_data\": \"Company data-science analytics team developed custom AI models over repeated camera imagery. Specific vision architecture, camera counts, site count and deployment start are not reported.\", \"integration_human_role\": \"Human operator is alerted or dispatched to act; no autonomous repair or control claim.\", \"deployment_scope\": \"Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated. Wellsite-camera workflow; quantitative scale not disclosed.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No performance or emissions metric attributed to this model. Devon separately reports flaring intensity but does not attribute that outcome to AI.\", \"period\": \"Period as described in source; metric measurement period not separately disclosed.\", \"attribution\": \"Issuer-reported; no independent validation.\", \"source_ids\": [\"SW7-009\"]}], \"economics\": \"None reported.\", \"timeline\": [{\"date\": \"2025 (report; covers calendar 2024 unless noted)\", \"event\": \"Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated.\", \"source_ids\": [\"SW7-009\"]}], \"limitations\": [\"No model accuracy, false-alert rate, equipment coverage, avoided emissions, uptime or financial attribution. General statement about capabilities, not proof of quantified impact.\"], \"assessment\": \"Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.\", \"sources\": [{\"title\": \"Devon Energy 2025 Sustainability Report\", \"date\": \"2025 (report; covers calendar 2024 unless noted)\", \"url\": \"https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf\", \"locator\": \"printed p. 9 / PDF p. 9, lines 713-738; full 90-page PDF read\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"SW7-009\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW7-009\"], \"field_sources\": {\"workflow\": [\"SW7-009\"], \"technology_data\": [\"SW7-009\"], \"deployment_scope\": [\"SW7-009\"], \"practical_relevance\": [\"SW7-009\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Issuer-reported; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated and original-source reviewed; specific cross-checks described in research methodology."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Original source opened; relevant passage read.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-251", "Company": "Diamondback Energy", "Company_primary": "Diamondback Energy", "Segment": "Upstream E&P", "Theme": "Operational AI", "Initiative": "Veriforce machine-learning contractor incident-risk scoring", "Extracted_finding": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "Stage_at_disclosure": "Continuing operating partnership during 2024; described in Diamondback's 2025 CSR. Contractors performing services for production operations; count and coverage percentage not stated.", "Reported_metric": "No Diamondback-specific incidents avoided or safety improvement attributed to the model. Its corporate safety statistics are not attributed to Veriforce scoring.", "Interpretation_limit": "Vendor product description can establish generic mechanism only; no Diamondback-specific performance or exact decision threshold. Do not infer the model's accuracy or causal safety benefit.", "Source_IDs": "SW7-010", "Source_date": "2025 (2024 performance reporting)", "Source_type": "Issuer disclosure", "Source_locator": "printed p. 20 / PDF p. 20, lines 1422-1438; full 51-page PDF read", "Source_URL": "https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418", "Evidence_status": "Original source opened; relevant passage read.", "PE_relationship": null, "Coverage_bucket": "Seventh targeted gap review", "Time_scope": "Continuing operating partnership during 2024; described in Diamondback's 2025 CSR.", "Context": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "Implementation": "Vendor Veriforce ML and predictive modeling applied to contractor safety data. Diamondback does not disclose features, algorithm, score thresholds, volumes or validation results.", "Assessment": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst classification; organizational ownership is not confirmed.", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Scope_type": "Internal adoption", "Ambition_assessment_basis": "Breadth and integration assessed separately from realized results.", "Review_tier": "Reviewed source", "Provenance": ["Seventh research swarm"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "SW7-D-AI-251", "Initiative_ID": "AI-251", "Title": "Veriforce machine-learning contractor incident-risk scoring", "Date": "2025 (2024 performance reporting)", "Finding": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "Metric": "No Diamondback-specific incidents avoided or safety improvement attributed to the model. Its corporate safety statistics are not attributed to Veriforce scoring.", "Source_IDs": "SW7-010", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}], "Latest_disclosure": "2025 (2024 performance reporting)", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["swarm7-ai-251"], "Research_search": "{\"key\": \"swarm7-ai-251\", \"packet_id\": \"swarm7-ai-251\", \"record_ids\": [\"AI-251\"], \"existing_ids\": [], \"related_existing_ids\": [], \"company\": \"Diamondback Energy\", \"title\": \"Veriforce machine-learning contractor incident-risk scoring\", \"department\": \"HSE & emissions\", \"secondary_departments\": [], \"category\": \"Internal adoption\", \"stage\": \"Continuing operating partnership during 2024; described in Diamondback's 2025 CSR. Contractors performing services for production operations; count and coverage percentage not stated.\", \"practical_relevance\": \"Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.\", \"business_problem\": \"\", \"workflow\": \"Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.\", \"technology_data\": \"Vendor Veriforce ML and predictive modeling applied to contractor safety data. Diamondback does not disclose features, algorithm, score thresholds, volumes or validation results.\", \"integration_human_role\": \"Diamondback uses scores to manage contractor risk and eligibility. The precise review, appeal or human override process is not disclosed.\", \"deployment_scope\": \"Continuing operating partnership during 2024; described in Diamondback's 2025 CSR. Contractors performing services for production operations; count and coverage percentage not stated.\", \"partners\": [], \"disclosed_results\": [{\"metric\": \"Disclosed evidence\", \"value\": \"No Diamondback-specific incidents avoided or safety improvement attributed to the model. Its corporate safety statistics are not attributed to Veriforce scoring.\", \"period\": \"Period as described in source; metric measurement period not separately disclosed.\", \"attribution\": \"Issuer-reported; no independent validation.\", \"source_ids\": [\"SW7-010\"]}], \"economics\": \"No cost, savings or return disclosed.\", \"timeline\": [{\"date\": \"2025 (2024 performance reporting)\", \"event\": \"Continuing operating partnership during 2024; described in Diamondback's 2025 CSR.\", \"source_ids\": [\"SW7-010\"]}], \"limitations\": [\"Vendor product description can establish generic mechanism only; no Diamondback-specific performance or exact decision threshold. Do not infer the model's accuracy or causal safety benefit.\"], \"assessment\": \"Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.\", \"sources\": [{\"title\": \"Diamondback Energy 2025 Corporate Sustainability Report\", \"date\": \"2025 (2024 performance reporting)\", \"url\": \"https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418\", \"locator\": \"printed p. 20 / PDF p. 20, lines 1422-1438; full 51-page PDF read\", \"type\": \"Issuer disclosure\", \"review_status\": \"Original source opened; relevant passage read.\", \"id\": \"SW7-010\", \"accessed\": \"2026-10-01\"}], \"source_ids\": [\"SW7-010\"], \"field_sources\": {\"workflow\": [\"SW7-010\"], \"technology_data\": [\"SW7-010\"], \"deployment_scope\": [\"SW7-010\"], \"practical_relevance\": [\"SW7-010\"]}, \"verification\": \"Original source reviewed\", \"cohort\": \"Seventh gap review · October 2026\", \"report_date\": \"2026-10-01\", \"evidence_note\": \"Issuer-reported; no independent validation.\"}", "Research_status": "Updated · October gap review", "Research_category": "Internal adoption", "Research_verification": "Original source reviewed", "Audit_flags": [], "Followup_review": {"date": "2026-10-01", "decision": "New profile", "note": "Deduplicated and original-source reviewed; specific cross-checks described in research methodology."}, "Asset_location": "Not disclosed", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "Original source reviewed", "claim_status": "Original source opened; relevant passage read.", "independent_validation": "Not established by this collection"}}, {"ID": "AI-252", "Company": "Imperial Oil", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "Upstream ML well dashboard and AI shovel-truck analytics in oil sands", "Extracted_finding": "Imperial's 2025 Investor Day deck lists a machine learning dashboard covering more than 4,500 wells and shovel and truck analytics leveraging AI among more than 50 digital initiatives it describes as deployed and expanding. The same slide lists generative AI for chatting with documents and data as an initiative still in development.", "Stage_at_disclosure": null, "Reported_metric": "Machine learning dashboard for > 4,500 wells", "Interpretation_limit": "The deck pairs digital tiles with dollar savings (e.g., advanced controls and analytics > $45M/yr) but does not attribute any savings figure specifically to the AI or ML items.", "Source_IDs": "OCT5-S001", "Source_date": "2025-04-17", "Source_type": "Investor day presentation (Form 8-K Exhibit 99.1)", "Source_locator": null, "Source_URL": "https://www.sec.gov/Archives/edgar/data/49938/000004993825000022/a2025irday-finalforexter.htm", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 52, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Imperial Oil", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-04-17", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#0"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Imperial previously had zero profiles. Same deck also lists \"Shovel and truck analytics leveraging AI\" and \"Generative AI – chat with documents and data\" (in development). Do not attach the $1.2B/$0.7B portfolio value or the $45M/yr advanced-controls figure to AI; the slide does not make that attribution. Investor Day presentation filed as Form 8-K Exhibit 99.1; classified as an operator filing rather than a press release.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-253", "Company": "ENEOS Holdings", "Segment": "Refining, LNG & chemicals", "Theme": "Operational AI", "Initiative": "Group AI organization and AI plant-operation expansion in medium-term plan", "Extracted_finding": "ENEOS's Fourth Medium-Term Management Plan (FY2025-FY2027) commits to launching a dedicated organization to promote AI across all operations, with manufacturing uses including expansion of automated plant operation, equipment failure prediction and AI image diagnosis for corrosion and leak detection. The plan cites its existing AI-based continuous autonomous operation of an atmospheric distillation unit at the Kawasaki Refinery (May 2024) as a current initiative.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "This is a plan; apart from the pre-window Kawasaki autonomous-operation cases, it does not report new deployments, scale or outcomes.", "Source_IDs": "OCT5-S002", "Source_date": "2025-05-12", "Source_type": "Company disclosure (medium-term management plan, TSE)", "Source_locator": null, "Source_URL": "https://www.hd.eneos.co.jp/english/news/release_information/upload_pdf/upload/20250512_05_04_2009288.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 63, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "ENEOS Holdings", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-05-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#1"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "ENEOS not previously in register. The Kawasaki ADU autonomous-operation launch (May 2024) predates the window; this record is the in-window plan to expand it. Japan's largest refiner fills a refining gap.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-254", "Company": "SK Innovation", "Segment": "Refining, LNG & chemicals", "Theme": "Operational AI", "Initiative": "Dual Brain AI anomaly detection across Ulsan refinery process units", "Extracted_finding": "SK Innovation describes a \"Dual Brain\" operating framework at its Ulsan Complex in which AI detects abnormal signs in processes and equipment and proposes responses, with field engineers making the final decision. The report says the system analyzes 98 interconnected process units and the company expects about 100 billion won a year in productivity gains.", "Stage_at_disclosure": null, "Reported_metric": "about 100 billion won ($72 million) annual productivity gains (expected)", "Interpretation_limit": "The 100 billion won figure is an expectation reported by a newspaper, not a realized or audited result.", "Source_IDs": "OCT5-S003", "Source_date": "2026-09-16", "Source_type": "News article (company statements at WAVE2026 expo)", "Source_locator": null, "Source_URL": "https://en.sedaily.com/finance/2026/09/16/sk-innovation-turns-to-ai-to-lift-refinery-output-by-100", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 55, "Primary_department": "Refining, LNG & process operations", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "SK Innovation", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-09-16", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#2"], "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "target", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Korea's largest refiner; not previously in register. Seek an SK Innovation press release or Ulsan CLX presentation to upgrade the grade. Metric wording comes from the article summary; check the exact wording before publishing.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-255", "Company": "GS Caltex", "Segment": "Refining, LNG & chemicals", "Theme": "Operational AI", "Initiative": "AI surveillance cameras for safety during Yeosu refinery turnaround", "Extracted_finding": "GS Caltex used AI-powered surveillance cameras during its first-half 2026 Yeosu turnaround to detect risks in real time, including whether staff wore protective gear and approached heavy equipment. The turnaround covered nine process units, with about 200 billion won invested and an average of 3,000 workers a day.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "No safety or efficiency outcome is attributed to the AI cameras; the 200 billion won is total turnaround spend, not AI spend.", "Source_IDs": "OCT5-S004", "Source_date": "2026-07-16", "Source_type": "News article", "Source_locator": null, "Source_URL": "https://www.koreaherald.com/article/10811139", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 48, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "GS Caltex", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-07-16", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#3"], "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Refining, LNG & chemicals", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Vice chairman quote in article: \"Our latest TA combines on-site expertise and digital and AI technologies to enhance both efficiency and safety.\" The MOVision valve-locator app and tube-cleaning robot are digital/robotic, not AI. Related lead: GS Caltex 8 billion won AI smart-manufacturing platform (Korea Herald 10419333), not opened.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-256", "Company": "Sonatrach (ENTP drilling subsidiary)", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "ENTP EYES AI-assisted real-time remote drilling-rig monitoring center", "Extracted_finding": "Sonatrach's drilling subsidiary ENTP launched \"ENTP EYES / Real Time – Decision Support Center\", a real-time remote monitoring system built on a solution that integrates artificial intelligence and uses cameras installed on rigs. The report says it was built by national expertise with Algerian startups to monitor operations and KPIs continuously.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "The article does not say what the AI does (e.g., video analytics vs. KPI analytics) and reports no outcome.", "Source_IDs": "OCT5-S005", "Source_date": "2025-12-22", "Source_type": "News article (CEO site visit)", "Source_locator": null, "Source_URL": "https://www.jeune-independant.net/daoudi-a-lentp-de-hassi-messaoud-lia-pour-optimiser-les-operations-de-forage/", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 51, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Sonatrach", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-12-22", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#4"], "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Sonatrach previously had zero profiles. Related lead: the Sonatrach News no. 49 (Q3 2025) staff award for AI-based predictive maintenance in the LQS (liquefaction & separation) activity.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-257", "Company": "Tourmaline Oil", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "Machine learning informs Montney well landing and completion design", "Extracted_finding": "On Tourmaline's Q2 2026 call, management credited improved Northeast BC Montney well performance partly to learnings from machine learning on landing and completion optimization. Higher completion intensity and longer laterals were cited alongside it.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Management names ML as one of several factors, so no share of the well-performance improvement can be attributed to ML.", "Source_IDs": "OCT5-S006", "Source_date": "2026-07-30", "Source_type": "Earnings call highlights (GuruFocus via Yahoo Finance)", "Source_locator": null, "Source_URL": "https://finance.yahoo.com/energy/articles/tourmaline-oil-corp-trmlf-q2-230031565.html", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 34, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Tourmaline Oil", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-07-30", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:coverage.json#5"], "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Canada's largest gas producer; previously zero profiles. Speaker: Jamie Heard, VP Capital Markets. The summary mentions a ~28% well-performance improvement, but the exact wording and basis were not verified and it has several causes, so it is left out of Reported_metric. Upgrade by checking against a full transcript. Grade set to D: secondary call summary; verify against the full transcript.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-258", "Company": "TotalEnergies", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "MethaneLive methane monitoring centre; agentic AI to target top emitters", "Extracted_finding": "TotalEnergies launched MethaneLive, a global centre that analyses real-time data from about 13,000 permanent methane sensors on all operated upstream sites using advanced algorithms. The company says this foundation now enables deployment of agentic AI solutions to target the highest-emitting equipment.", "Stage_at_disclosure": null, "Reported_metric": "35 fugitive methane emissions detected since MethaneLive's early-2026 launch (company-reported; attributed to real-time data analysis, not specifically to AI)", "Interpretation_limit": "The explicit AI element (agentic AI) is described as newly enabled, not as the source of the 35 detections; detection is credited to sensors, 'advanced algorithms' and real-time data analysis.", "Source_IDs": "OCT5-S007", "Source_date": "2026-06-18", "Source_type": "operator press release", "Source_locator": null, "Source_URL": "https://www.sec.gov/Archives/edgar/data/0000879764/000110465926079516/tm2619405d1_ex99-4.htm", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 41, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "TotalEnergies", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-06-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#0"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Original totalenergies.com press-release URL now returns HTTP 410; same text filed with SEC as 6-K exhibit. Same release says AI systems monitor nearly 3,000 pieces of equipment for early failure signs (maintenance, not emissions). Distinct from AI-078 (Cognite data foundation). Prior TotalEnergies STORM/MethaneLive page (2026-05-26) contained no AI wording.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-259", "Company": "Shell", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "T-Pulse AI CCTV monitoring of unsafe acts, conditions and leaks", "Extracted_finding": "Shell says it continues to expand deployment of T-Pulse, an AI-automated safety monitoring solution from Detect Technologies that uses CCTV to identify unsafe acts, unsafe conditions and potential leaks in real time. Shell reports sites achieved an average 70% reduction in the number of safety observations T-Pulse makes since deployment began.", "Stage_at_disclosure": null, "Reported_metric": "Average 70% reduction in the number of safety observations made by T-Pulse since the start of deployment (site average; company-reported; includes line-of-fire and material-handling categories)", "Interpretation_limit": "The 70% figure is a fall in tool-flagged observations, not injuries or leaks avoided; no site count, baseline period or independent validation is disclosed.", "Source_IDs": "OCT5-S008", "Source_date": "2026-05-13", "Source_type": "annual report", "Source_locator": null, "Source_URL": "https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section/promo.multi.stream/1779352356739/36306d968747b6079d3f800b2b1552a033856b5d/shell-integrated-annual-and-sustainability-report.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 51, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Shell", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-05-13", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#1"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Source_date is the PDF creation date (2026-05-13); report covers FY2025 and may have first been published earlier in 2026 – verify. Same paragraph reports AFDD fatigue/distraction devices in 20 countries with a 25% SMVI-rate reduction, but those devices are not described as AI, so they are excluded. Vendor: Detect Technologies (T-Pulse).", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-260", "Company": "bp", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "AI safety-observation pilot at Whiting refinery; LENS AI assistant", "Extracted_finding": "bp says that in 2025 it launched a pilot of a third-party AI-driven safety observation tool at its Whiting refinery to monitor high-risk situations and activities in real time in the field. It also added an AI Assistant to its LENS digital safety-learning tool to identify and summarise incident information.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "The report claims qualitative benefits (risk management, coaching, engagement) but no safety metric attributable to either AI tool.", "Source_IDs": "OCT5-S010", "Source_date": "2026-03-09", "Source_type": "sustainability report", "Source_locator": null, "Source_URL": "https://www.bp.com/api/files/6cqieuqhq4no/master/6ISrW0tRFKSHbaBT7yeVNf/18b0bc82f1dcf6b0b4d1df187477f8c8/bp-sustainability-report-2025.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 50, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "bp", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-03-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#3"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Vendor of the Whiting observation tool not named. Distinct from AI-136 (bpx Permian driver-hazard alerts, 2024 report). Source_date from PDF metadata (created 2026-03-05, modified 2026-03-09).", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-261", "Company": "EQT", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "AI-based ergonomic assessments to reduce musculoskeletal risk", "Extracted_finding": "EQT says that in 2025 its safety team began using AI technologies to perform ergonomic assessments that evaluate body positioning during common tasks. The aim is to find opportunities to reduce musculoskeletal risk.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "No vendor, number of assessments or injury outcome is disclosed, so effect on musculoskeletal incidents cannot be inferred.", "Source_IDs": "OCT5-S011", "Source_date": "2026-07-20", "Source_type": "sustainability report", "Source_locator": "Printed/PDF p.59, EHS Management System and Risk Identification; use began during 2025. Independently rechecked 9 October 2026.", "Source_URL": "https://cdn.prod.website-files.com/6751b0d7add5edb1f0141aef/6ab53525a5a0e486b918d352_EQT-2025-Sustainability-Report.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 33, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "EQT", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-07-20", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#4"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Release date from EQT IR news release (20 July 2026). The same report notes a new AI Usage Corporate Policy (governance, not HSE). The report's methane programme (Appalachian Methane Initiative, more than 15,000 aerial surveys) contains no explicit AI wording, so it is excluded.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-262", "Company": "Eni", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "Safety Presence AI predicting recurring hazards from safety-data weak signals", "Extracted_finding": "Eni reports using Safety Presence, an AI tool that predicts recurring hazardous situations from weak signals recorded in safety databases. Eni says it has generated about 800 alerts since deployment, which led to targeted preventive actions.", "Stage_at_disclosure": null, "Reported_metric": "About 800 alerts generated since deployment, leading to targeted preventive actions (company-reported; deployment date not stated)", "Interpretation_limit": "Alert counts measure tool output; Eni does not attribute any injury-rate change to Safety Presence or report its precision.", "Source_IDs": "OCT5-S012", "Source_date": "2026-06-03", "Source_type": "annual report", "Source_locator": null, "Source_URL": "https://report.eni.com/annual-report-2025/en/_assets/downloads/sustainability-eni-ar25.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 36, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Eni", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-06-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#5"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "The same statement (Asset integrity, p.76) describes a 2025 'processing tool based on artificial intelligence techniques' that supports systematic analysis of asset-integrity assessments; it is a candidate secondary record. Source_date is the PDF metadata date; Eni's AR2025 may have been published earlier in 2026. Not in AI-027.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-263", "Company": "Woodside Energy", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "HaloGuard machine-vision red-zone system on contracted drilling rig", "Extracted_finding": "Woodside said it had recently installed the HaloGuard system on a contracted drilling rig, combining wearables, real-time location and machine vision to track people and equipment on the rig floor. The system alarms when a person nears moving equipment and can stop that equipment.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "The source says 'machine vision' but not AI or ML, and reports no near-miss or incident data.", "Source_IDs": "OCT5-S013", "Source_date": "2025-04-07", "Source_type": "regulatory filing", "Source_locator": null, "Source_URL": "https://www.woodside.com/docs/default-source/asx-announcements/2025/021-sustainability-briefing-2025-transcript.pdf?sfvrsn=91b6b681_3", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 44, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Woodside", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-04-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#6"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "ASX-lodged investor-briefing transcript (p.7). Graded B because AI is implied by 'machine vision' rather than named. HaloGuard vendor not named in the source.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-264", "Company": "SLB", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "EPA-approved SOOFIE continuous methane monitoring with AI event detection", "Extracted_finding": "On 23 September 2025 EPA approved SLB's SOOFIE continuous monitoring network as an alternative test method for Subpart OOOOb periodic screening. EPA's letter and memo describe a proprietary AI-based event-detection algorithm and a hybrid physics-based and machine-learning quantification approach.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Approval only makes the method eligible for compliance screening; it shows neither operator adoption nor emissions reduced.", "Source_IDs": "OCT5-S014", "Source_date": "2025-09-23", "Source_type": "regulatory filing", "Source_locator": null, "Source_URL": "https://www.epa.gov/system/files/documents/2025-09/slb-soofie-approval-letter-and-justification-memo.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 39, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "SLB", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-09-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#7"], "Record_type": "vendor_product", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "independent", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Initially submitted 14 March 2025 under ChampionX and resubmitted under SLB on 28 August 2025. A separate EPA approval of an SLB methane lidar camera method (19 August 2025) refers only to a 'plume detection algorithm', not AI. Older 2021 ConocoPhillips SOOFIE ML material is pre-window.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-265", "Company": "Clean Connect AI", "Segment": "Technology vendor", "Theme": "Operational AI", "Initiative": "EPA-approved LeakFinder autonomous OGI with computer-vision methane detection", "Extracted_finding": "On 13 May 2025 EPA approved Clean Connect AI's LeakFinder, an autonomous OGI camera on the Minerva platform, as alternative test method MATM-011 for OOOOb periodic screening. EPA's description says proprietary computer-vision models detect methane/VOC emission events and upload them to a cloud dashboard.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Regulatory approval of a vendor method; it names no operator deployments and reports no field emissions outcomes.", "Source_IDs": "OCT5-S015", "Source_date": "2025-05-13", "Source_type": "regulatory filing", "Source_locator": null, "Source_URL": "https://www.epa.gov/system/files/documents/2025-05/clean-connect-alternative-test-method-signed.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 44, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Clean Connect AI", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-05-13", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Commercial AI product", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#8"], "Record_type": "vendor_product", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "independent", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Submission dated 27 July 2024; approval letter signed 13 May 2025. Vendor technology, not operator adoption; include only if the register tracks vendor capability.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-266", "Company": "Percepto", "Segment": "Technology vendor", "Theme": "Operational AI", "Initiative": "AI Emission Detector analysing autonomous-drone OGI video for methane", "Extracted_finding": "Percepto launched an AI Emission Detector in which AI analyses OGI video from autonomous scheduled drone missions to flag potential methane emissions at component level. Percepto says the system is already deployed across major U.S. basins but names no operators.", "Stage_at_disclosure": null, "Reported_metric": "Field-validated performance down to 100 g/hr at 90% confidence (vendor-reported detection threshold)", "Interpretation_limit": "Vendor claim with unnamed customers; the detection threshold is a performance spec, not an emissions-reduction outcome.", "Source_IDs": "OCT5-S016", "Source_date": "2025-06-16", "Source_type": "vendor press release", "Source_locator": null, "Source_URL": "https://percepto.com/percepto-launches-ai-emission-detector-for-remote-drone-based-methane-surveying/", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 40, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Percepto", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-06-16", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Commercial AI product", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:emissions.json#9"], "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Technology vendor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Secondary coverage suggests Chevron is a Percepto partner, but the release does not name it; not attributed.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-267", "Company": "Chevron / Arkestro", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "Arkestro predictive AI procurement expanded across global business units", "Extracted_finding": "Arkestro says Chevron expanded use of its AI-based predictive procurement platform to speed supplier cycle times and manage sourcing cost volatility. Deployment spans U.S. teams and international teams in Buenos Aires, Angola, the Philippines, Australia and Thailand.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Vendor release; the >20% price-reduction figure is a cross-customer claim, not a Chevron result.", "Source_IDs": "OCT5-S017", "Source_date": "2025-10-15", "Source_type": "Vendor press release with named customer quote", "Source_locator": null, "Source_URL": "https://arkestro.com/press-releases/arkestro-announces-customer-expansion-with-chevron-driving-predictive-procurement-transformation-across-global-markets/", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 37, "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Chevron", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-10-15", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#0"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "U.S. plus five international locations; business units not specified.", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Quote is from Joshua Flowers, Senior Manager, SCM IT Category Management, Chevron. Release also notes Aramco Ventures co-led Arkestro's $36M round.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-268", "Company": "Equinor / Cognite", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "AI anomaly detection on robot inspection data at Sture and Mongstad", "Extracted_finding": "Equinor streams visual and audio data from Boston Dynamics Spot and Adigo VÅK robots into Cognite Data Fusion, where AI flags equipment anomalies. The pilot moved from Technology Center Mongstad to the Sture oil terminal and Mongstad refinery.", "Stage_at_disclosure": null, "Reported_metric": "Spot has run more than 1000 autonomous inspection rounds", "Interpretation_limit": "Vendor-authored account; outcomes (faster fault detection, time savings) are qualitative and not quantified.", "Source_IDs": "OCT5-S018", "Source_date": "2025-09-19", "Source_type": "Trade-press article based on vendor customer story", "Source_locator": null, "Source_URL": "https://www.technologyrecord.com/article/cognite-and-equinor-leverage-robotics-and-ai-to-improve-field-operations", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 38, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Equinor", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-09-19", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#1"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Cognite-hosted customer story is undated: https://www.cognite.com/en/resources/customer-stories/cognite-and-equinor-leverage-robotics-and-ai-to-reinvent-safer-smarter-field-operations . Cognite is being acquired by Schneider Electric (announced 2026-06-30).", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-269", "Company": "Aker BP / Cognite / NVIDIA", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "Cognite–NVIDIA NV-Tesseract anomaly detection across wells and equipment", "Extracted_finding": "Aker BP deployed Cognite's platform with integrated NVIDIA NV-Tesseract time-series models to detect anomalies earlier on wells, heat exchangers and other critical equipment. The aim is to move from alarm-based monitoring to predictive maintenance that reduces unplanned downtime.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "No quantified results; deployment claim originates with the vendor announcement.", "Source_IDs": "OCT5-S019", "Source_date": "2026-03-23", "Source_type": "Trade press (World Oil)", "Source_locator": null, "Source_URL": "https://worldoil.com/news/2026/3/23/ai-anomaly-detection-gains-traction-offshore-as-aker-bp-deploys-cognite-nvidia-solution/", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 38, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Aker BP", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-03-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#2"], "Record_type": "company_adoption", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Distinct workflow from AI-049 (Atlas AI root-cause agents) and AI-121 (Cognite–NVIDIA with Celanese); consider linking as related.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-270", "Company": "ONGC / SLB", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "SLB Lumi data and AI platform across all ONGC assets", "Extracted_finding": "ONGC awarded SLB a major digital contract to deploy the Lumi data and AI platform as an enterprise solution across all its assets. SLB says it will create a unified data foundation for advanced analytics and AI-enabled workflows.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Contract award reported by the supplier; no rollout milestones or results disclosed.", "Source_IDs": "OCT5-S020", "Source_date": "2026-04-24", "Source_type": "Vendor earnings release (SEC 8-K exhibit)", "Source_locator": null, "Source_URL": "https://www.sec.gov/Archives/edgar/data/0000087347/000119312526174940/d111579dex99.htm", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 38, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "ONGC", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-04-24", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#3"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Separate from AI-156 (ONGC DrillOps) and AI-225 (Tachyus water-injection optimization).", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-271", "Company": "Pampa Energía / Halliburton", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "Halliburton DecisionSpace 365 AI-enabled digital transformation in Vaca Muerta", "Extracted_finding": "Pampa Energía signed a multi-year agreement with Halliburton covering digital orchestration, high-resolution reservoir modeling, logistics optimization and energy-efficiency management. Halliburton will deploy DecisionSpace 365 cloud applications and describes the programme as putting industrial-grade AI into team workflows.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "AI language comes from a Halliburton executive quote; specific models and deployment dates are not given.", "Source_IDs": "OCT5-S021", "Source_date": "2026-06-09", "Source_type": "Vendor press release", "Source_locator": null, "Source_URL": "https://www.halliburton.com/en/about-us/press-release/pampa-energia-selects-halliburton-to-support-enterprise-digital-transformation", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 37, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Pampa Energía", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-06-09", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#4"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Quote from Tony Antoun, SVP Landmark, Halliburton.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-272", "Company": "Aramco / NVIDIA", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "NVIDIA MoU for industrial AI computing, AI hub and robotics", "Extracted_finding": "Aramco signed an MoU with NVIDIA to develop industrial AI computing infrastructure, an AI hub and enterprise platforms, and an engineering and robotics centre of excellence. It was one of 34 agreements with US companies signed at the May 2025 Saudi-US Investment Forum.", "Stage_at_disclosure": null, "Reported_metric": null, "Interpretation_limit": "Non-binding MoU; no named operational workflow or deployment yet.", "Source_IDs": "OCT5-S022", "Source_date": "2025-05-18", "Source_type": "Trade press (JPT) summarising operator announcement", "Source_locator": null, "Source_URL": "https://jpt.spe.org/aramco-signs-mous-worth-up-to-90-billion-for-lng-ai-emissions-reduction-and-asset-management-initiatives", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": null, "Implementation": null, "Assessment": null, "Detail_words": 43, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-05-18", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Strategy only", "Research_verification": "secondary only", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:vendors.json#5"], "Record_type": "strategy", "Evidence_grade": "D", "Evidence_label": "D · Secondary / unverified", "Attribution": "secondary", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Aramco announced the package on 2025-05-13. Comparable to AI-070 (Aramco–Microsoft MoU).", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "secondary only", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-273", "Company": "ADNOC Onshore / AIQ", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "Autonomous gas-lift control across 200 onshore producer wells", "Extracted_finding": null, "Stage_at_disclosure": "Deployed across 200 ADNOC Onshore producer wells; operating start and measurement duration unstated.", "Reported_metric": "Abstract reports up to 2% oil-production increase and 5–30% lift-gas optimization at well level; these are not fleet-average results.", "Interpretation_limit": "Abstract only: baseline, measurement period and evaluation details unconfirmed. No implementation cost or cash benefit disclosed. JPT identifies the underlying paper as not peer-reviewed.", "Source_IDs": "OCT5-S025", "Source_date": "2025-11-03", "Source_type": "SPE technical paper authored by ADNOC Onshore and AIQ staff", "Source_locator": "SPE-230130-MS abstract, deployment results and challenges paragraphs", "Source_URL": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/D021S043R003/793441", "Evidence_status": "Original technical-paper abstract reviewed; full paper unavailable", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": "Allocate lift gas across wells while respecting facility limits. Distinct from the broader AiPSO platform (AI-011).", "Implementation": "Centralized process control works with cloud analytics. Field teams validate sensors, adjust injection and log overrides; cloud workflows retrain models and update operating envelopes. Sensor/valve failures required fallback logic. Missing measurements reduced model accuracy; inferential models, synthetic data and transfer learning addressed gaps.", "Assessment": "Control, retraining and fallback practices make this a useful operating case; reported gains do not establish returns.", "Detail_words": 76, "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "ADNOC", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-011"], "Disclosures": [], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:z-followup.json#1"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "200 wells", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Operator/AIQ-authored abstract; original publisher indexed text reviewed. Full-paper access remains open.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original technical-paper abstract reviewed; full paper unavailable", "independent_validation": "Not established by this collection"}}, {"ID": "AI-274", "Company": "Northern Oil and Gas (NOG)", "Segment": "Independent E&P", "Theme": "Enterprise AI", "Initiative": "Drakkar AI and data platform for investment and portfolio decisions", "Extracted_finding": null, "Stage_at_disclosure": "Described as a company-wide platform in the February 2026 deck; launch date, user count and workflow coverage are not supplied.", "Reported_metric": "NOG claims improved ROI and lower G&A, without a quantified, attributable result.", "Interpretation_limit": "No model architecture, decision controls, validation, attributable savings or system cost disclosed. AFE inputs do not establish automated JIB or AFE accounting.", "Source_IDs": "OCT5-S026", "Source_date": "2026-02-25", "Source_type": "Operator investor presentation", "Source_locator": "Slide 21, Drakkar; cover date February 25, 2026", "Source_URL": "https://s203.q4cdn.com/492547254/files/doc_financials/2025/q4/NOG-Earnings-Presentation-Q4-YE-2025-FINAL.pdf", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": "NOG holds non-operated interests and compares operators, well economics and investments. Drakkar supports those decisions.", "Implementation": "A company-wide platform consolidates expenditure approvals, field reports, production metrics and financials. NOG describes real-time analytics with agentic AI capabilities, bespoke applications and cross-team communication. Portfolio analysis considers productivity, cost structure and operating expenses.", "Assessment": "A named non-operator example extends coverage from field decisions to investment decisions. Broad benefit claims need a baseline and costs.", "Detail_words": 69, "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Northern Oil and Gas (NOG)", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-02-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:z-followup.json#2"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Added in the 5 October 2026 research review; classifications supplied by the reviewing researcher.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-275", "Company": "Enbridge", "Segment": "Midstream", "Theme": "Operational AI", "Initiative": "Pipeline deformation prioritization with machine learning", "Extracted_finding": null, "Stage_at_disclosure": "Under development for Liquids Pipelines; production rollout not established.", "Reported_metric": "No model-performance or attributable reliability outcome disclosed.", "Interpretation_limit": "Accuracy, false-positive/negative rates, deployment scope and costs are undisclosed. Five inspection segments in the report belong to a separate distribution-risk program.", "Source_IDs": "OCT5-S027", "Source_date": "2026-05-28", "Source_type": "Operator sustainability report", "Source_locator": "Report printed page 28, LP: Using machine learning to prioritize activities", "Source_URL": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "Evidence_status": "Primary source text reviewed (5 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "5 October 2026 research review", "Time_scope": null, "Context": "Many inspection-reported deformations are harmless; engineers must identify potentially injurious features.", "Implementation": "Enbridge is developing a model trained on historical risk assessments, in-line inspections and field non-destructive examinations. Engineers assess and mitigate prioritized features.", "Assessment": "A specific engineering handoff is disclosed, but development status must remain clear.", "Detail_words": 45, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Enbridge", "Review_tier": "Reviewed source", "Provenance": ["5 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2026-05-28", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 5 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["oct5:z-followup.json#3"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Midstream", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Added in the 5 October 2026 research review; classifications supplied by the reviewing researcher.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (5 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-276", "Company": "Quantum Capital Group", "Segment": "Private equity & investors", "Theme": "Operational AI", "Initiative": "Vickery Energy StrataSource AI land-title research with Quantum", "Extracted_finding": "Vickery Energy, identified in Quantum’s 2026 report, uses Azure OpenAI to read deeds, leases and conveyances for ownership, heirship and lease availability. Custom Azure Document Intelligence OCR handles scanned historical records, including handwriting.", "Stage_at_disclosure": "Operating use described by source; go-live date and full rollout scope not established.", "Reported_metric": "Microsoft reports almost 10,000 documents in the largest production run, completed in 40 hours versus an estimated 40 business days of landmen’s work.", "Interpretation_limit": "The sponsor reports 640 parcels and 9,520 documents in ~40 hours versus an estimated ~40 business days. Difficult records go to human review. Accuracy, complete costs and replacement of legal title opinions are not established.", "Source_IDs": "OCT5-S028; OCT9-QCG-2026", "Source_date": "2026-04-30", "Source_type": "Microsoft customer story", "Source_locator": "From the field to the back office and beyond", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "The 2026 Quantum report names Vickery Energy and StrataSource AI, used in West Virginia leasing. Earlier Microsoft reporting left the operator unnamed.", "Implementation": "Vickery Energy, identified in Quantum’s 2026 report, uses Azure OpenAI to read deeds, leases and conveyances for ownership, heirship and lease availability. Custom Azure Document Intelligence OCR handles scanned historical records, including handwriting.", "Assessment": "A specific workflow is disclosed; costs and comparable net-cash returns remain unreported.", "Detail_words": 108, "Primary_department": "Land & legal", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Quantum Capital Group", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-052", "AI-054"], "Disclosures": [{"ID": "OCT9-AI-276", "Initiative_ID": "AI-276", "Title": "Vickery Energy StrataSource AI land-title research with Quantum", "Date": "2026", "Finding": "The 2026 Quantum report names Vickery Energy and StrataSource AI, used in West Virginia leasing. Earlier Microsoft reporting left the operator unnamed.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}], "Latest_disclosure": "2026-04-30", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/gaps-v2.json#0"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Private equity & investors", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Vendor account. Operator identity, rollout date, accuracy and review protocol are undisclosed. The roughly $6,500 title-opinion price is context, not measured savings or proof that legal opinions are replaced.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Operating_company": "Vickery Energy", "Current_source_check": "2026-10-09", "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-277", "Company": "PRIO", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "Copilot-assisted HR job descriptions and competency mapping", "Extracted_finding": "PRIO records and transcribes manager interviews. Copilot combines custom prompts, a competency model and company context to produce job descriptions and competency maps. The source says more than 15 careers had been covered. Use began about a year before the April 2025 disclosure, implying a start around April 2024; this is an in-window report of continuing use.", "Stage_at_disclosure": "Operating use described by source; go-live date and full rollout scope not established.", "Reported_metric": "PRIO staff report more than 70 hours saved per area: approximately 50 consultant hours plus 26 manager hours previously, versus about four hours for the new process. A claimed 99% “assertiveness” has no stated validation method.", "Interpretation_limit": "Vendor-published company quotes; approximate baseline. Number of areas, aggregate hours, quality-test denominator, subscription costs and realized cash savings are not supplied. Historical-origin implementation predates the reporting window.", "Source_IDs": "OCT5-S029", "Source_date": "2025-04-15", "Source_type": "Microsoft customer story", "Source_locator": "PRIO HR workflow and statements by Marcelo Ludwig", "Source_URL": "https://news.microsoft.com/pt-br/prio-utiliza-inteligencia-artificial-no-rh-e-economiza-mais-de-70-horas-por-setor/", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "PRIO records and transcribes manager interviews. Copilot combines custom prompts, a competency model and company context to produce job descriptions and competency maps. The source says more than 15 careers had been covered. Use began about a year before the April 2025 disclosure, implying a start around April 2024; this is an in-window report of continuing use.", "Implementation": "PRIO records and transcribes manager interviews. Copilot combines custom prompts, a competency model and company context to produce job descriptions and competency maps. The source says more than 15 careers had been covered. Use began about a year before the April 2025 disclosure, implying a start around April 2024; this is an in-window report of continuing use.", "Assessment": "A specific workflow is disclosed; costs and comparable net-cash returns remain unreported.", "Detail_words": 183, "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "PRIO", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [], "Latest_disclosure": "2025-04-15", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/gaps-v2.json#1"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Vendor-published company quotes; approximate baseline. Number of areas, aggregate hours, quality-test denominator, subscription costs and realized cash savings are not supplied. Historical-origin implementation predates the reporting window.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-278", "Company": "Kiewit", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "T-Pulse safety-camera pilot at an offshore fabrication yard", "Extracted_finding": "Kiewit Offshore Services tested Detect Technologies T-Pulse on 18 cameras in three Ingleside fabrication-yard areas. Live video generated hazard alerts and reports for site leadership. Staff customized the rules to Kiewit’s safeguards and site-specific practices; six additional cameras were planned.", "Stage_at_disclosure": "Completed pilot; further expansion was planned.", "Reported_metric": "Kiewit reports 91.7% accuracy, 64 verified false positives and a 99% reduction in false positives after site-specific customization.", "Interpretation_limit": "Completed yard pilot, not an offshore-platform or fleet rollout. Duration, total alert count, definition of accuracy and false-negative rate are unstated; the 99% reduction has no denominator. Site tuning is a limitation, not a failed or cancelled deployment.", "Source_IDs": "OCT5-S030", "Source_date": "2025-04-07", "Source_type": "Company-owned pilot account", "Source_locator": "KOS Ingleside pilot; accuracy and customization paragraphs", "Source_URL": "https://www.kiewit.com/newsroom/a-different-kind-of-pilot/", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "Kiewit Offshore Services tested Detect Technologies T-Pulse on 18 cameras in three Ingleside fabrication-yard areas. Live video generated hazard alerts and reports for site leadership. Staff customized the rules to Kiewit’s safeguards and site-specific practices; six additional cameras were planned.", "Implementation": "Kiewit Offshore Services tested Detect Technologies T-Pulse on 18 cameras in three Ingleside fabrication-yard areas. Live video generated hazard alerts and reports for site leadership. Staff customized the rules to Kiewit’s safeguards and site-specific practices; six additional cameras were planned.", "Assessment": "A specific workflow is disclosed; costs and comparable net-cash returns remain unreported.", "Detail_words": 132, "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Kiewit", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-259"], "Disclosures": [], "Latest_disclosure": "2025-04-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/gaps-v2.json#2"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Completed yard pilot, not an offshore-platform or fleet rollout. Duration, total alert count, definition of accuracy and false-negative rate are unstated; the 99% reduction has no denominator. Site tuning is a limitation, not a failed or cancelled deployment.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-279", "Company": "Canadian Natural Resources", "Segment": "Independent E&P", "Theme": "Operational AI", "Initiative": "AI-assisted oil-sands reliability and hydrogen optimization", "Extracted_finding": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations. Separate reliability work covers pump maintenance and piping assessments; an in-house conveyor alignment algorithm was executing in Q4 2025.", "Stage_at_disclosure": "Operating use described by source; go-live date and full rollout scope not established.", "Reported_metric": "Company reports approximately 3% less unplanned downtime in 2025 versus 2024, maintenance intervals extended from six to ten months, and approximately 80% fewer man-hours for piping AI integrity assessments. Separately, upgrading hydrogen optimization: approximately CAD 8 million annual savings, company-reported; costs not disclosed.", "Interpretation_limit": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate. Reliability asset/pump counts and whether the 3% change is relative or percentage points remain unspecified. The separate 400,000 bbl/year flotation-automation claim is not counted as an AI outcome.", "Source_IDs": "OCT5-S031; WATCH17AM-CNRL", "Source_date": "2025-11-07", "Source_type": "Operator investor presentation", "Source_locator": "PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention", "Source_URL": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "Evidence_status": "Original operator slides reviewed; benefits remain company-reported", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": "Presentation dated 7 November 2025; exact posting and deployment dates unknown. File path is dated December 2025.", "Context": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.", "Implementation": "The presentation names the task but does not explain the model, inputs, vendor or approval process. Separate reliability dashboards and smart monitoring support pump and piping work. Conveyor completion is not established.", "Assessment": "A specific workflow is disclosed; costs and comparable net-cash returns remain unreported.", "Detail_words": 123, "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Canadian Natural Resources", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-098"], "Disclosures": [{"ID": "WATCH17AM-AI-279", "Initiative_ID": "AI-279", "Title": "CNRL puts an annual savings figure on hydrogen optimization", "Date": "2025-11-07", "Finding": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations. The presentation names the task but does not explain the model, inputs, vendor or approval process.", "Metric": "The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.", "Source_IDs": "WATCH17AM-CNRL", "Source_URL": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "Evidence": "Original operator source read; results are not independently validated", "Origin": "17 October morning source review", "Assessment": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate."}], "Latest_disclosure": "2025-11-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch17am-ai-279"], "Research_search": "{\"packet_id\": \"watch17am-ai-279\", \"key\": \"watch17am-ai-279\", \"record_ids\": [\"AI-279\"], \"company\": \"Canadian Natural Resources\", \"title\": \"CNRL puts an annual savings figure on hydrogen optimization\", \"category\": \"Internal adoption\", \"department\": \"Refining, LNG & process operations\", \"report_date\": \"2026-10-17\", \"verification\": \"verified primary\", \"stage\": \"Operating use reported; site and start date unspecified\", \"business_problem\": \"CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.\", \"workflow\": \"The presentation names the task but does not explain the model, inputs, vendor or approval process.\", \"technology_data\": \"The presentation names the task but does not explain the model, inputs, vendor or approval process.\", \"deployment_scope\": \"Presentation dated 7 November 2025; exact posting and deployment dates unknown. File path is dated December 2025.\", \"integration_human_role\": \"Hydrogen approval and control rights are not disclosed. Do not transfer conveyor control rights to hydrogen optimization.\", \"economics\": \"The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.\", \"practical_relevance\": \"How much hydrogen was avoided, against which baseline, and what does the system cost to run?\", \"assessment\": \"No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.\", \"limitations\": [\"No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.\"], \"source_ids\": [\"WATCH17AM-CNRL\"], \"sources\": [{\"id\": \"WATCH17AM-CNRL\", \"title\": \"Investor Open House · 7 November 2025\", \"url\": \"https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50\", \"date\": \"2025-11-07\", \"publisher\": \"Canadian Natural Resources\", \"locator\": \"PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention\", \"review_status\": \"Original source read by editor\", \"supports\": [\"CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.\", \"The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.\"], \"accessed\": \"2026-10-17\"}], \"disclosed_results\": [{\"metric\": \"Company-reported annual hydrogen savings; not net ROI\", \"value\": \"Approximately CAD 8 million per year\", \"period\": \"Measurement period not disclosed\", \"attribution\": \"operator\", \"source_ids\": [\"WATCH17AM-CNRL\"]}], \"timeline\": [{\"date\": \"2025-11-07\", \"event\": \"CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.\", \"source_ids\": [\"WATCH17AM-CNRL\"]}], \"field_sources\": {\"business_problem\": [\"WATCH17AM-CNRL\"], \"workflow\": [\"WATCH17AM-CNRL\"], \"technology_data\": [\"WATCH17AM-CNRL\"], \"deployment_scope\": [\"WATCH17AM-CNRL\"], \"integration_human_role\": [\"WATCH17AM-CNRL\"], \"economics\": [\"WATCH17AM-CNRL\"]}}", "Research_status": "Expanded 17 October 2026 · hydrogen workflow inside existing oil-sands profile", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/gaps-v2.json#3", "research/watch-2026-10-17-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.", "Outcome_counted": true, "Outcome_count_note": "Existing counted result profile; hydrogen is a separate reported savings claim, not a second profile or all-cost ROI.", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Original operator slides reviewed", "independent_validation": "Not established by this collection"}, "Current_source_check": "2026-10-17"}, {"ID": "AI-280", "Company": "Baker Hughes", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "GenAI technical-support assistant for artificial-lift field engineers", "Extracted_finding": "The eight-week prototype processed over 200 document pages and reportedly answered 85% of questions successfully. This is pilot evaluation, not demonstrated production rollout or labor savings.", "Stage_at_disclosure": "Internal ALS prototype/pilot; portal integration and scaling planned.", "Reported_metric": "Eight-week prototype; 200+ pages processed; 85% reported answer success. Estimated 10,000 hours/year is existing support workload, not saved time.", "Interpretation_limit": "Partner-reported; test denominator and quality method absent. No deployed user count, spend, recurring cost, realized hours saved or payback. Undated client page is linked by a 17 July 2025 EPAM-hosted media summary; deployment date is unknown. Separate four-week Leucipa assistant targets clients, not this internal workflow.", "Source_IDs": "OCT5-S038; OCT5-S039", "Source_date": "Undated case; referenced 2025-07-17", "Source_type": "EPAM delivery-partner client case study", "Source_locator": "Artificial Lift Systems assistant; Results; supporting EPAM newsroom item dated 17 July 2025", "Source_URL": "https://www.epam.com/services/client-work/leveraging-genai-to-increase-productivity", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "Baker Hughes estimates its artificial-lift specialists handle 1,600 support requests and spend 10,000 hours annually answering routine questions.", "Implementation": "EPAM piloted a retrieval assistant over technical bulletins, manuals and procedures using Textract, Titan embeddings, OpenSearch and Claude on AWS Bedrock. Subject-matter experts tested and refined answers; wider document coverage and ALS-portal integration remained planned.", "Assessment": "Useful build-time and evaluation evidence; routine adoption and financial return remain unverified.", "Detail_words": 91, "Primary_department": "Workforce & learning", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Baker Hughes", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-020", "AI-112"], "Disclosures": [{"ID": "OCT5-AI-280-1", "Initiative_ID": "AI-280", "Title": "Dated media reference to the ALS assistant case", "Date": "2025-07-17", "Finding": "EPAM’s 17 July 2025 media summary links the undated original client case and repeats the pilot evaluation and planned integration. It supports an in-window public account; it does not establish the go-live date or independent verification.", "Metric": "", "Source_IDs": "OCT5-S039", "Source_URL": "https://www.epam.com/about/newsroom/in-the-news/2025/how-ai-accelerates-oil-production-and-saves-thousands-of-hours-of-support", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}], "Latest_disclosure": "Undated case; referenced 2025-07-17", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": [], "Research_search": null, "Research_status": "Added 6 October 2026 · Supplemented 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/three-rounds-review.json#0"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Single_source_vendor": true, "Review_note": "Partner-reported; test denominator and quality method absent. No deployed user count, spend, recurring cost, realized hours saved or payback. Undated client page is linked by a 17 July 2025 EPAM-hosted media summary; deployment date is unknown. Separate four-week Leucipa assistant targets clients, not this internal workflow.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified partner", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-281", "Company": "Aramco", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "Aramco Intelligent Live Earth Model for AI-assisted geosteering", "Extracted_finding": "Aramco’s FY2025 annual report names an AI geosteering workflow and reports approximately 10% increased reservoir contact and an average two-day reduction in drilling time per well.", "Stage_at_disclosure": "FY2025 report presents the model as an impactful AI solution used in geosteering; first deployment date, well count and rollout denominator are not stated.", "Reported_metric": "Aramco reports approximately 10% increased reservoir contact and two fewer average drilling days per well. No well count, comparison period, measurement method or project cost is disclosed.", "Interpretation_limit": "Operator-reported operating metrics in an annual report. Scope, baseline, time period, sample size and attribution method are not provided; no independent measurement reviewed. Do not convert two days per well to monetary benefit without well cost/time economics.", "Source_IDs": "OCT5-S050; OCT5-S043", "Source_date": "2026-03-10", "Source_type": "Operator annual report", "Source_locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “Aramco Intelligent Live Earth Model”", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "Geosteering uses subsurface interpretation during drilling to optimize well placement and reservoir contact.", "Implementation": "Aramco describes its Intelligent Live Earth Model as leveraging AI to optimize well placement and maximize reservoir contact in geosteering operations.", "Assessment": "A named upstream AI workflow with quantified operational measures; evidence supports the reported metric, not an underwritten net return.", "Detail_words": 79, "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-037"], "Disclosures": [], "Latest_disclosure": "2026-03-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai281"], "Research_search": "{\"review_date\": \"2026-10-06\", \"verification\": \"verified primary\", \"source_review\": \"Relevant original issuer source independently reviewed on 2026-10-06.\", \"attribution\": \"Operator account\", \"result_label\": \"Issuer-reported disclosure and scope limits\", \"packet_id\": \"oct6-executive-ai281\", \"title\": \"Aramco Intelligent Live Earth Model for AI-assisted geosteering\", \"source_locator\": \"Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “Aramco Intelligent Live Earth Model”\", \"workflow\": \"Aramco describes its Intelligent Live Earth Model as using AI to optimize well placement and maximize reservoir contact in geosteering operations.\", \"technology_data\": \"Company report identifies the workflow and use but gives no model architecture, data feeds or control/approval design.\", \"deployment_scope\": \"Listed among the company’s FY2025 impactful AI solutions; first deployment date, well count, regions and rollout denominator are not stated.\", \"practical_relevance\": \"A named upstream geosteering workflow links subsurface interpretation to well placement and drilling time.\", \"economics\": \"No well cost, implementation/sustaining cost, monetized benefit or payback disclosed.\", \"limitations\": [\"Operator-reported annual-report metrics. Baseline, time period, sample size, measurement method and well count are not stated.\", \"Do not translate two days per well into savings without rig/day cost and incremental drilling scope.\", \"No independent verification or project costs reviewed.\"], \"timeline_date\": \"2026-03-10\", \"timeline_event\": \"FY2025 Annual Report describes the workflow; launch date and measurement period are not disclosed.\", \"stage\": \"FY2025 report presents the model as an impactful AI solution used in geosteering; first deployment date, well count and rollout denominator are not stated.\", \"key\": \"oct6-executive-ai281\", \"record_ids\": [\"AI-281\"], \"company\": \"Aramco\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"source_ids\": [\"OCT5-S043\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S043\"], \"workflow\": [\"OCT5-S043\"], \"technology_data\": [\"OCT5-S043\"], \"deployment_scope\": [\"OCT5-S043\"], \"practical_relevance\": [\"OCT5-S043\"], \"economics\": [\"OCT5-S043\"], \"integration_human_role\": [\"OCT5-S043\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and scope limits\", \"value\": \"Aramco reports approximately 10% increased reservoir contact and an average two-day reduction in drilling time per well.\", \"period\": \"2026-03-10\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S043\"]}], \"timeline\": [{\"date\": \"2026-03-10\", \"event\": \"FY2025 Annual Report describes the workflow; launch date and measurement period are not disclosed.\", \"source_ids\": [\"OCT5-S043\"]}], \"sources\": [{\"id\": \"OCT5-S043\", \"title\": \"Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value\", \"url\": \"https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf\", \"date\": \"2026-03-10\", \"type\": \"Operator annual report\", \"publisher\": \"Aramco\", \"locator\": \"Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “Aramco Intelligent Live Earth Model”\", \"review_status\": \"Relevant original issuer source independently reviewed on 2026-10-06.\"}]}", "Research_status": "Added 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/zz-executive-run.json#7"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Issuer-report claim; use only as company-reported operating outcome. Metric period/sample/method/cost are undisclosed.", "Outcome_counted": true, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-282", "Company": "Aramco", "Segment": "NOC", "Theme": "Operational AI", "Initiative": "Aramco PE.AI advanced petrophysical prediction and synthetic well logs", "Extracted_finding": "Aramco’s FY2025 annual report identifies PE.AI advanced petrophysical prediction as a named AI workflow for synthetic well-log generation.", "Stage_at_disclosure": "Named as an impactful AI solution; launch date, operating asset count and volume are not disclosed.", "Reported_metric": "No quantified operational or financial result disclosed; issuer says the workflow reduces the need for traditional well logging and additional pilot holes.", "Interpretation_limit": "Qualitative issuer-reported benefit only. No logging/pilot-hole count, accuracy, validation benchmark, baseline, savings or deployment scope disclosed. Do not treat avoided-work claim as measured value.", "Source_IDs": "OCT5-S051; OCT5-S043", "Source_date": "2026-03-10", "Source_type": "Operator annual report", "Source_locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “PE.AI advanced petrophysical prediction”", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "Petrophysical prediction helps characterize subsurface formations from well data.", "Implementation": "Aramco describes PE.AI as an AI-powered platform that generates synthetic well logs. The report says this reduces the need for traditional well logging and drilling additional pilot holes.", "Assessment": "Adds a distinct, company-named subsurface prediction workflow; only qualitative benefits are published.", "Detail_words": 67, "Primary_department": "Exploration & geoscience", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "Aramco", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-037", "AI-077"], "Disclosures": [], "Latest_disclosure": "2026-03-10", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai282"], "Research_search": "{\"review_date\": \"2026-10-06\", \"verification\": \"verified primary\", \"source_review\": \"Relevant original issuer source independently reviewed on 2026-10-06.\", \"attribution\": \"Operator account\", \"result_label\": \"Issuer-reported disclosure and scope limits\", \"packet_id\": \"oct6-executive-ai282\", \"title\": \"Aramco PE.AI advanced petrophysical prediction and synthetic well logs\", \"source_locator\": \"Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “PE.AI advanced petrophysical prediction”\", \"workflow\": \"Aramco says PE.AI generates synthetic well logs and reduces the need for traditional well logging and drilling additional pilot holes.\", \"technology_data\": \"Named AI-powered platform in company report; model family, evaluation set, accuracy, quality controls and human review are not disclosed.\", \"deployment_scope\": \"Named among FY2025 impactful AI solutions; deployment start date, assets, wells and workflow rollout status are not specified.\", \"practical_relevance\": \"Potentially changes petrophysical data acquisition and subsurface interpretation requirements.\", \"economics\": \"No quantified cost avoidance, operating result, implementation cost or return disclosed.\", \"limitations\": [\"No denominator, accuracy benchmark, validation protocol, avoided logging/pilot-hole count, cost, or realized value.\", \"Do not convert qualitative “reduces the need” wording into a measured savings result.\"], \"timeline_date\": \"2026-03-10\", \"timeline_event\": \"FY2025 Annual Report names the solution; operating start and scope are not disclosed.\", \"stage\": \"Named as an impactful AI solution; launch date, operating asset count and volume are not disclosed.\", \"key\": \"oct6-executive-ai282\", \"record_ids\": [\"AI-282\"], \"company\": \"Aramco\", \"category\": \"Internal adoption\", \"department\": \"Exploration & geoscience\", \"source_ids\": [\"OCT5-S043\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S043\"], \"workflow\": [\"OCT5-S043\"], \"technology_data\": [\"OCT5-S043\"], \"deployment_scope\": [\"OCT5-S043\"], \"practical_relevance\": [\"OCT5-S043\"], \"economics\": [\"OCT5-S043\"], \"integration_human_role\": [\"OCT5-S043\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and scope limits\", \"value\": \"Qualitative issuer claim that synthetic logs reduce the need for traditional logging and additional pilot holes; no quantified outcome.\", \"period\": \"2026-03-10\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S043\"]}], \"timeline\": [{\"date\": \"2026-03-10\", \"event\": \"FY2025 Annual Report names the solution; operating start and scope are not disclosed.\", \"source_ids\": [\"OCT5-S043\"]}], \"sources\": [{\"id\": \"OCT5-S043\", \"title\": \"Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value\", \"url\": \"https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf\", \"date\": \"2026-03-10\", \"type\": \"Operator annual report\", \"publisher\": \"Aramco\", \"locator\": \"Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “PE.AI advanced petrophysical prediction”\", \"review_status\": \"Relevant original issuer source independently reviewed on 2026-10-06.\"}]}", "Research_status": "Added 6 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/zz-executive-run.json#8"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "unclear", "Scale_note": "", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Issuer report names the AI workflow but does not establish launch date, scope, accuracy or quantified outcome.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-283", "Company": "TotalEnergies", "Segment": "Integrated & supermajor", "Theme": "Operational AI", "Initiative": "Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D", "Extracted_finding": "TotalEnergies announced a >€100m Pangea 5 HPC contract. It is a multi-workload compute enabler, not an AI-model deployment or an AI-only investment.", "Stage_at_disclosure": "Contract signed; first commissioning planned for 2027.", "Reported_metric": "Issuer-disclosed infrastructure investment >€100m and sixfold compute-capacity increase. It forecasts approximately 40% lower energy consumption at equal performance and a fivefold reduction in cooling-system consumption.", "Interpretation_limit": "Whole infrastructure contract amount is not allocated to AI; compute supports seismic and other digital/scientific workloads. Energy/cooling figures are forecasts, not measured savings. Do not add the full amount to AI-specific investment, do not sum with Mistral program commitment, and do not count forecast efficiency as realized ROI.", "Source_IDs": "OCT5-S052", "Source_date": "2026-05-06", "Source_type": "Operator press release", "Source_locator": "Release paragraphs 38–50: contract, investment, compute, workload and forecast efficiency", "Source_URL": "https://totalenergies.com/newsroom/totalenergies-develops-pangea-5-a-next-generation-supercomputer-469512/?lang=eng", "Evidence_status": "Primary source text reviewed (6 October 2026 research)", "PE_relationship": null, "Coverage_bucket": "6 October 2026 research review", "Time_scope": null, "Context": "High-performance computing supports seismic imaging, subsurface engineering and multiple scientific/digital workloads.", "Implementation": "TotalEnergies signed a design-and-installation contract with Dell Technologies and NVIDIA for Pangea 5 at its CSTJF in Pau. The company says it will multiply computing power sixfold and support advanced seismic engineering and AI R&D; first commissioning is planned for 2027.", "Assessment": "Material strategy/infrastructure context behind subsurface AI; separately classified outside operator-adoption counts and AI-specific spend.", "Detail_words": 88, "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": null, "Ambition_shortlist_category": null, "Scope_type": null, "Ambition_assessment_basis": null, "Company_primary": "TotalEnergies", "Review_tier": "Reviewed source", "Provenance": ["6 October 2026 research review"], "Reference_IDs": [], "Related_IDs": ["AI-077", "AI-078"], "Disclosures": [], "Latest_disclosure": "2026-05-06", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct6-executive-ai283"], "Research_search": "{\"review_date\": \"2026-10-06\", \"verification\": \"verified primary\", \"source_review\": \"Relevant original issuer source independently reviewed on 2026-10-06.\", \"attribution\": \"Operator account\", \"result_label\": \"Issuer-reported disclosure and scope limits\", \"packet_id\": \"oct6-executive-ai283\", \"title\": \"Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D\", \"source_locator\": \"Release paragraphs 38–50: signed contract, investment, compute capacity, workloads and efficiency forecasts\", \"workflow\": \"Pangea 5 is a supercomputer contract supporting advanced seismic engineering and multiple scientific/digital workloads, including AI research and development.\", \"technology_data\": \"TotalEnergies names Dell Technologies and NVIDIA as collaborators; system architecture/utilization and workload allocation are not reported.\", \"deployment_scope\": \"Contract signed; system at CSTJF in Pau; first commissioning planned in 2027.\", \"practical_relevance\": \"A material compute enabler behind subsurface imaging and AI R&D, separately classified as infrastructure rather than operator adoption.\", \"economics\": \"Company reports >€100m whole-system investment. No AI-only share, operating cost, compute utilization, model value or return disclosed.\", \"limitations\": [\"Infrastructure contract supports several scientific/digital uses; amount is not AI-specific spend.\", \"Energy and cooling savings are forecasts, not realized values.\", \"Do not add >€100m to the separate TotalEnergies–Mistral joint program investment or count as AI ROI.\"], \"timeline_date\": \"2026-05-06\", \"timeline_event\": \"Contract announced 2026-05-06; first commissioning targeted in 2027.\", \"stage\": \"Contract signed; first commissioning planned for 2027.\", \"key\": \"oct6-executive-ai283\", \"record_ids\": [\"AI-283\"], \"company\": \"TotalEnergies\", \"category\": \"Strategy / enabling technology\", \"department\": \"IT, data & enterprise transformation\", \"source_ids\": [\"OCT5-S052\"], \"report_date\": \"2026-10-06\", \"field_sources\": {\"business_problem\": [\"OCT5-S052\"], \"workflow\": [\"OCT5-S052\"], \"technology_data\": [\"OCT5-S052\"], \"deployment_scope\": [\"OCT5-S052\"], \"practical_relevance\": [\"OCT5-S052\"], \"economics\": [\"OCT5-S052\"], \"integration_human_role\": [\"OCT5-S052\"]}, \"disclosed_results\": [{\"metric\": \"Issuer-reported disclosure and scope limits\", \"value\": \"Sixfold computing-power increase is planned. Company forecasts approximately 40% lower energy use at equal performance and fivefold lower cooling-system consumption.\", \"period\": \"2026-05-06\", \"attribution\": \"Operator account\", \"source_ids\": [\"OCT5-S052\"]}], \"timeline\": [{\"date\": \"2026-05-06\", \"event\": \"Contract announced 2026-05-06; first commissioning targeted in 2027.\", \"source_ids\": [\"OCT5-S052\"]}], \"sources\": [{\"id\": \"OCT5-S052\", \"title\": \"Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D\", \"url\": \"https://totalenergies.com/newsroom/totalenergies-develops-pangea-5-a-next-generation-supercomputer-469512/?lang=eng\", \"date\": \"2026-05-06\", \"type\": \"Operator press release\", \"publisher\": \"TotalEnergies\", \"locator\": \"Release paragraphs 38–50: signed contract, investment, compute capacity, workloads and efficiency forecasts\", \"review_status\": \"Relevant original issuer source independently reviewed on 2026-10-06.\"}]}", "Research_status": "Added 6 October 2026", "Research_category": "Strategy / enabling technology", "Research_verification": "verified primary", "Audit_flags": [], "Audit_review": null, "Audit_outcome": null, "Normalized_stage": null, "HQ": null, "Research_origin": ["research:oct6/zz-executive-run.json#9"], "Record_type": "strategy", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Single_source_vendor": false, "Review_note": "Multi-workload HPC strategy enabler; >€100m whole-system investment, not an AI-specific project-cost disclosure.", "Outcome_counted": false, "Outcome_count_note": "", "Financial_benefits_disclosed": false, "Evidence_review": {"source_read": "verified primary", "claim_status": "Primary source text reviewed (6 October 2026 research)", "independent_validation": "Not established by this collection"}}, {"ID": "AI-284", "Company": "BPX Energy", "Company_primary": "BPX Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Daily lift recommendations across a 441-well Permian pilot", "Extracted_finding": "A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.", "Stage_at_disclosure": "Pilot on a 441-well asset; result population limited to wells adopting recommendations", "Reported_metric": "Authors report average 6% incremental oil on wells that adopted recommendations, plus more than 20% lower average field gas-lift use. The number of adopting wells is not disclosed.", "Interpretation_limit": "The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established.", "Source_IDs": "OCT10-AI-284-SWPSC", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Source_locator": "SWPSC 2026051 abstract: method, 441-well asset and reported results", "Source_URL": "https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin", "Evidence_status": "Original technical sources reviewed; operator/vendor-reported results", "PE_relationship": null, "Coverage_bucket": "10 October 2026 field-case review", "Time_scope": "2026 source; exact conference publication and deployment days unknown.", "Context": "Lift settings and intervention choices must account for changing well conditions and shared gas-injection constraints.", "Implementation": "Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints. Operators receive daily recommendations. The paper does not establish autonomous actuation or disclose how many wells adopted the recommendations. Permian asset with 441 gas-lift and ESP wells. BPX is bp’s US onshore business; this is a distinct field case, not a count of bp-wide deployment.", "Assessment": "A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst workflow classification; organizational ownership is not confirmed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["10 October 2026 Luna research; parent source and adversarial review"], "Reference_IDs": [], "Related_IDs": ["AI-083"], "Disclosures": [{"ID": "OCT10-AI-284", "Initiative_ID": "AI-284", "Title": "Daily lift recommendations across a 441-well Permian pilot", "Date": "2026", "Finding": "Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.", "Metric": "Authors report average 6% incremental oil on wells that adopted recommendations, plus more than 20% lower average field gas-lift use. The number of adopting wells is not disclosed.", "Source_IDs": "OCT10-AI-284-SWPSC", "Source_URL": "https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established."}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct10-field-ai-284"], "Research_search": "{\"packet_id\": \"oct10-field-ai-284\", \"key\": \"oct10-field-ai-284\", \"record_ids\": [\"AI-284\"], \"company\": \"BPX Energy\", \"title\": \"Daily lift recommendations across a 441-well Permian pilot\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-10\", \"verification\": \"verified partner\", \"stage\": \"Pilot on a 441-well asset; result population limited to wells adopting recommendations\", \"business_problem\": \"Lift settings and intervention choices must account for changing well conditions and shared gas-injection constraints.\", \"workflow\": \"Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.\", \"technology_data\": \"Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.\", \"integration_human_role\": \"Operators receive daily recommendations. The paper does not establish autonomous actuation or disclose how many wells adopted the recommendations.\", \"deployment_scope\": \"Permian asset with 441 gas-lift and ESP wells. BPX is bp’s US onshore business; this is a distinct field case, not a count of bp-wide deployment.\", \"economics\": \"Lower gas-lift use could reduce resource demand, but no dollar saving, implementation cost or net return is disclosed.\", \"practical_relevance\": \"A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.\", \"assessment\": \"A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.\", \"limitations\": [\"The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established.\"], \"source_ids\": [\"OCT10-AI-284-SWPSC\"], \"partners\": [\"Xecta\"], \"disclosed_results\": [{\"metric\": \"Author-reported field results\", \"value\": \"Authors report average 6% incremental oil on wells that adopted recommendations, plus more than 20% lower average field gas-lift use. The number of adopting wells is not disclosed.\", \"period\": \"Duration not disclosed\", \"attribution\": \"Operator/vendor account; not independently evaluated\", \"source_ids\": [\"OCT10-AI-284-SWPSC\"]}], \"timeline\": [{\"date\": \"2026\", \"event\": \"Pilot on a 441-well asset; result population limited to wells adopting recommendations. Source date is not the deployment start.\", \"source_ids\": [\"OCT10-AI-284-SWPSC\"]}], \"sources\": [{\"id\": \"OCT10-AI-284-SWPSC\", \"title\": \"Daily lift recommendations across a 441-well Permian pilot\", \"url\": \"https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin\", \"date\": \"2026\", \"type\": \"Operator/vendor-coauthored technical paper\", \"publisher\": \"Southwestern Petroleum Short Course\", \"locator\": \"SWPSC 2026051 abstract: method, 441-well asset and reported results\", \"review_status\": \"Relevant original source text reviewed; performance remains author-reported, not independently validated.\"}], \"field_sources\": {\"business_problem\": [\"OCT10-AI-284-SWPSC\"], \"workflow\": [\"OCT10-AI-284-SWPSC\"], \"technology_data\": [\"OCT10-AI-284-SWPSC\"], \"integration_human_role\": [\"OCT10-AI-284-SWPSC\"], \"deployment_scope\": [\"OCT10-AI-284-SWPSC\"], \"economics\": [\"OCT10-AI-284-SWPSC\"]}}", "Research_status": "Added 10 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/oct10-next-swarm/parent-review.json"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "pilot", "Scale_note": "", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established.", "Outcome_counted": true, "Outcome_count_note": "", "Current_source_check": "2026-10-10", "Evidence_review": {"source_read": "Original technical sources reviewed", "claim_status": "Original technical sources reviewed; operator/vendor-reported results", "independent_validation": "Not established"}, "Detail_words": 77}, {"ID": "AI-285", "Company": "Burk Royalty", "Company_primary": "Burk Royalty", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Pump-by-exception monitoring and control at a West Texas lease", "Extracted_finding": "An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.", "Stage_at_disclosure": "Operating small-site system; eight-month evaluation", "Reported_metric": "Over eight months, authors report 43% lower electricity consumption and daily strokes; electricity expense of $660 to $376 per well-month; routine field time of 126 to about 18 hours per month (86% lower); and 26% lower hot-oiling cost. Production stayed around 19 BOPD versus an expected 4.25% annual decline. Authors calculate about $65,000 incremental profit at $65/bbl WTI and capital recovery in slightly over one month.", "Interpretation_limit": "The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established.", "Source_IDs": "OCT10-AI-285-SWPSC; OCT10-BURK-PDF", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Source_locator": "SWPSC 2026052 abstract; linked full paper PDF pp. 3, 7–9 (site configuration, eight-month results, economics and projections)", "Source_URL": "https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control", "Evidence_status": "Original technical sources reviewed; operator/vendor-reported results", "PE_relationship": null, "Coverage_bucket": "10 October 2026 field-case review", "Time_scope": "2026 source; exact conference publication and deployment days unknown.", "Context": "A remote satellite lease required routine visits and manual pumping checks, with risks from equipment problems and tank overfill.", "Implementation": "Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions. Software adjusts setpoints and can shut equipment down for protection; staff review alerts and respond to exceptions. This is an integrated instrumentation, control and work-process change, not a stand-alone chatbot. West Texas satellite lease. The full paper describes five sucker-rod producers and one injector; its HTML abstract says six producers and one injector. The source population conflict remains unresolved.", "Assessment": "An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst workflow classification; organizational ownership is not confirmed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["10 October 2026 Luna research; parent source and adversarial review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "OCT10-AI-285", "Initiative_ID": "AI-285", "Title": "Pump-by-exception monitoring and control at a West Texas lease", "Date": "2026", "Finding": "Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.", "Metric": "Over eight months, authors report 43% lower electricity consumption and daily strokes; electricity expense of $660 to $376 per well-month; routine field time of 126 to about 18 hours per month (86% lower); and 26% lower hot-oiling cost. Production stayed around 19 BOPD versus an expected 4.25% annual decline. Authors calculate about $65,000 incremental profit at $65/bbl WTI and capital recovery in slightly over one month.", "Source_IDs": "OCT10-AI-285-SWPSC; OCT10-BURK-PDF", "Source_URL": "https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established."}], "Latest_disclosure": "2026", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct10-field-ai-285"], "Research_search": "{\"packet_id\": \"oct10-field-ai-285\", \"key\": \"oct10-field-ai-285\", \"record_ids\": [\"AI-285\"], \"company\": \"Burk Royalty\", \"title\": \"Pump-by-exception monitoring and control at a West Texas lease\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-10\", \"verification\": \"verified partner\", \"stage\": \"Operating small-site system; eight-month evaluation\", \"business_problem\": \"A remote satellite lease required routine visits and manual pumping checks, with risks from equipment problems and tank overfill.\", \"workflow\": \"Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.\", \"technology_data\": \"Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.\", \"integration_human_role\": \"Software adjusts setpoints and can shut equipment down for protection; staff review alerts and respond to exceptions. This is an integrated instrumentation, control and work-process change, not a stand-alone chatbot.\", \"deployment_scope\": \"West Texas satellite lease. The full paper describes five sucker-rod producers and one injector; its HTML abstract says six producers and one injector. The source population conflict remains unresolved.\", \"economics\": \"Some operating costs are disclosed: electricity expense falls from $660 to $376 per well-month, alongside hot-oiling and workload measures. Authors calculate ~$65,000 incremental profit over eight months at $65/bbl WTI and payback slightly over one month. The capital amount and full recurring-cost bridge are undisclosed, so this is partial cost evidence, not a validated all-in ROI. Potential avoided spill-cleanup cost is an event estimate, not recurring savings.\", \"practical_relevance\": \"An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.\", \"assessment\": \"An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.\", \"limitations\": [\"The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established.\"], \"source_ids\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"partners\": [\"Amplified Industries\"], \"disclosed_results\": [{\"metric\": \"Author-reported field results\", \"value\": \"Over eight months, authors report 43% lower electricity consumption and daily strokes; electricity expense of $660 to $376 per well-month; routine field time of 126 to about 18 hours per month (86% lower); and 26% lower hot-oiling cost. Production stayed around 19 BOPD versus an expected 4.25% annual decline. Authors calculate about $65,000 incremental profit at $65/bbl WTI and capital recovery in slightly over one month.\", \"period\": \"Eight months\", \"attribution\": \"Operator/vendor account; not independently evaluated\", \"source_ids\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"]}], \"timeline\": [{\"date\": \"2026\", \"event\": \"Operating small-site system; eight-month evaluation. Source date is not the deployment start.\", \"source_ids\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"]}], \"sources\": [{\"id\": \"OCT10-AI-285-SWPSC\", \"title\": \"Pump-by-exception monitoring and control at a West Texas lease\", \"url\": \"https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control\", \"date\": \"2026\", \"type\": \"Operator/vendor-coauthored technical paper\", \"publisher\": \"Southwestern Petroleum Short Course\", \"locator\": \"SWPSC 2026052 abstract; linked full paper PDF pp. 3, 7–9 (site configuration, eight-month results, economics and projections)\", \"review_status\": \"Relevant original source text reviewed; performance remains author-reported, not independently validated.\"}, {\"id\": \"OCT10-BURK-PDF\", \"title\": \"Field Optimization Reimagined: full paper\", \"url\": \"https://www.swpshortcourse.org/sites/default/files/papers/2026052%20-%20Field%20Optimization%20Reimagined%20Data%20At%20The%20Core%20Exceptions%20At%20The%20Center%2C%20People%20In%20Control.pdf\", \"date\": \"2026\", \"type\": \"Operator/vendor-coauthored technical paper\", \"publisher\": \"Southwestern Petroleum Short Course\", \"locator\": \"PDF pp. 3, 7–9; figure on p. 8\", \"review_status\": \"Relevant original source text reviewed; performance remains author-reported, not independently validated.\"}], \"field_sources\": {\"business_problem\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"workflow\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"technology_data\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"integration_human_role\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"deployment_scope\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"], \"economics\": [\"OCT10-AI-285-SWPSC\", \"OCT10-BURK-PDF\"]}}", "Research_status": "Added 10 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/oct10-next-swarm/parent-review.json"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": true, "Single_source_vendor": true, "Review_note": "The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established.", "Outcome_counted": true, "Outcome_count_note": "", "Current_source_check": "2026-10-10", "Evidence_review": {"source_read": "Original technical sources reviewed", "claim_status": "Original technical sources reviewed; operator/vendor-reported results", "independent_validation": "Not established"}, "Detail_words": 95}, {"ID": "AI-286", "Company": "Expand Energy", "Company_primary": "Expand Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Autonomous liquid unloading on nine Haynesville gas wells", "Extracted_finding": "A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.", "Stage_at_disclosure": "Operating field test on nine wells across eight pads; 63–83-day results", "Reported_metric": "SLB’s JPT account reports 70–139% cumulative gas-production gains over 63–83 days on nine wells across eight pads, with daily gains up to 350 Mscf/d. More than 80 MMscf additional gas per well-year is an estimate, not a year of observed production.", "Interpretation_limit": "Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established.", "Source_IDs": "OCT10-AI-286-SWPSC; OCT10-EXPAND-JPT", "Source_date": "2026-07-01", "Source_type": "Supplier-authored technical account", "Source_locator": "SWPSC 2026001 abstract; JPT 1 July 2026, Haynesville Field Results", "Source_URL": "https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments", "Evidence_status": "Original technical sources reviewed; operator/vendor-reported results", "PE_relationship": null, "Coverage_bucket": "10 October 2026 field-case review", "Time_scope": "2026 source; exact conference publication and deployment days unknown.", "Context": "Liquid loading in intermittent gas wells requires timely shut-ins and reopening to restore flow without unnecessary production loss.", "Implementation": "An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight. The system makes local shut-in and choke decisions. The sources do not describe a language model or a general-purpose agent; this is specialist production control. Nine Haynesville wells across eight pads. The related July 2026 JPT account reports results over 63–83 days; it describes the same deployment, not an independent replication.", "Assessment": "A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst workflow classification; organizational ownership is not confirmed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["10 October 2026 Luna research; parent source and adversarial review"], "Reference_IDs": [], "Related_IDs": ["AI-194", "AI-195"], "Disclosures": [{"ID": "OCT10-AI-286", "Initiative_ID": "AI-286", "Title": "Autonomous liquid unloading on nine Haynesville gas wells", "Date": "2026-07-01", "Finding": "An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.", "Metric": "SLB’s JPT account reports 70–139% cumulative gas-production gains over 63–83 days on nine wells across eight pads, with daily gains up to 350 Mscf/d. More than 80 MMscf additional gas per well-year is an estimate, not a year of observed production.", "Source_IDs": "OCT10-AI-286-SWPSC; OCT10-EXPAND-JPT", "Source_URL": "https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established."}], "Latest_disclosure": "2026-07-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct10-field-ai-286"], "Research_search": "{\"packet_id\": \"oct10-field-ai-286\", \"key\": \"oct10-field-ai-286\", \"record_ids\": [\"AI-286\"], \"company\": \"Expand Energy\", \"title\": \"Autonomous liquid unloading on nine Haynesville gas wells\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-10\", \"verification\": \"verified partner\", \"stage\": \"Operating field test on nine wells across eight pads; 63–83-day results\", \"business_problem\": \"Liquid loading in intermittent gas wells requires timely shut-ins and reopening to restore flow without unnecessary production loss.\", \"workflow\": \"An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.\", \"technology_data\": \"An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.\", \"integration_human_role\": \"The system makes local shut-in and choke decisions. The sources do not describe a language model or a general-purpose agent; this is specialist production control.\", \"deployment_scope\": \"Nine Haynesville wells across eight pads. The related July 2026 JPT account reports results over 63–83 days; it describes the same deployment, not an independent replication.\", \"economics\": \"The reported short-period production gains and estimated annual gas volume are not net financial returns. Full implementation and running costs are not disclosed in the reviewed primary passages.\", \"practical_relevance\": \"A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.\", \"assessment\": \"A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.\", \"limitations\": [\"Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established.\"], \"source_ids\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"Author-reported field results\", \"value\": \"SLB’s JPT account reports 70–139% cumulative gas-production gains over 63–83 days on nine wells across eight pads, with daily gains up to 350 Mscf/d. More than 80 MMscf additional gas per well-year is an estimate, not a year of observed production.\", \"period\": \"63–83 days\", \"attribution\": \"Operator/vendor account; not independently evaluated\", \"source_ids\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"]}], \"timeline\": [{\"date\": \"2026-07-01\", \"event\": \"Operating field test on nine wells across eight pads; 63–83-day results. Source date is not the deployment start.\", \"source_ids\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"]}], \"sources\": [{\"id\": \"OCT10-AI-286-SWPSC\", \"title\": \"Autonomous liquid unloading on nine Haynesville gas wells\", \"url\": \"https://www.swpshortcourse.org/paper/2026001-autonomous-edge-based-optimization-liquid-loading-intermittent-gas-wells\", \"date\": \"2026\", \"type\": \"Operator/vendor-coauthored technical paper\", \"publisher\": \"Southwestern Petroleum Short Course\", \"locator\": \"SWPSC 2026001 abstract; JPT 1 July 2026, Haynesville Field Results\", \"review_status\": \"Relevant original source text reviewed; performance remains author-reported, not independently validated.\"}, {\"id\": \"OCT10-EXPAND-JPT\", \"title\": \"Edge Computing and IIoT Enable Autonomous Artificial Lift and Well Optimization\", \"url\": \"https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments\", \"date\": \"2026-07-01\", \"type\": \"Supplier-authored technical account\", \"publisher\": \"Journal of Petroleum Technology / SPE\", \"locator\": \"Field Results: Haynesville Basin\", \"review_status\": \"Relevant original source text reviewed; performance remains author-reported, not independently validated.\"}], \"field_sources\": {\"business_problem\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"workflow\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"technology_data\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"integration_human_role\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"deployment_scope\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"], \"economics\": [\"OCT10-AI-286-SWPSC\", \"OCT10-EXPAND-JPT\"]}}", "Research_status": "Added 10 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/oct10-next-swarm/parent-review.json"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established.", "Outcome_counted": true, "Outcome_count_note": "", "Current_source_check": "2026-10-10", "Evidence_review": {"source_read": "Original technical sources reviewed", "claim_status": "Original technical sources reviewed; operator/vendor-reported results", "independent_validation": "Not established"}, "Detail_words": 92}, {"ID": "AI-287", "Company": "Flywheel Energy", "Company_primary": "Flywheel Energy", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "An internal AI platform for employees to build business apps", "Extracted_finding": "Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.", "Stage_at_disclosure": "Existing company-reported platform with growing adoption; user and app counts undisclosed. Analyst-role workflows are hiring intentions.", "Reported_metric": "Growing adoption is described qualitatively. No user/app count, quantified business result, complete cost or net return is disclosed.", "Interpretation_limit": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect.", "Source_IDs": "OCT12-FLY-PLATFORM; OCT12-FLY-ACCOUNTING; OCT12-FLY-FEED; OCT12-FLY-QUORUM; OCT12-FLY-INTERN; OCT12-FLY-COMPANY", "Source_date": "2026", "Source_type": "Employer recruitment disclosure", "Source_locator": "Job Summary and Key Responsibilities", "Source_URL": "https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750", "Evidence_status": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "PE_relationship": "Flywheel identifies itself as an operating partner of Stone Ridge Energy; precise legal ownership is not established here.", "Coverage_bucket": "12 October 2026 targeted Flywheel review", "Time_scope": "2026 recruitment disclosures; exact dates unknown. Quorum source is undated; internship responsibilities concern Summer 2027.", "Context": "Let business teams improve their own workflows without asking the central software team to build every application.", "Implementation": "Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included. Separately, Flywheel advertises land and accounting AI analyst roles. The accounting role specifies AI-assisted workflows, document extraction, API/data-pipeline integration, acquisition-data normalization and migration, retries, fallbacks and exception queues. It calls for human review of ownership or legal determinations. These are staffing responsibilities, not confirmation that every workflow is deployed. OpenAI/Azure OpenAI are examples of desired candidate experience, not confirmed vendors. Quorum names Flywheel as an On Demand Land customer for more than five years. Its testimonial describes easier data movement and access for nontechnical users. Quorum’s newer AI features are general product capabilities; Flywheel-specific use is not established. A search-indexed Summer 2027 internship description proposes LLM, agentic and MCP integration prototypes, plus ML work on production forecasting and predictive maintenance. The direct posting expired/redirected during review; these prospective tasks are not operating field AI.", "Assessment": "Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Land & legal", "Finance, investment & strategy"], "Department_assignment_basis": "Source-described business teams; analytical department mapping.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["12 October 2026 Luna research; parent original-source review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "OCT12-FLY-PLATFORM-DISC", "Initiative_ID": "AI-287", "Title": "Existing internal AI app platform", "Date": "2026", "Finding": "Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-PLATFORM", "Source_URL": "https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750", "Evidence": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-HIRING-DISC", "Initiative_ID": "AI-287", "Title": "Land and accounting AI roles: intended automation work", "Date": "2026", "Finding": "Separately, Flywheel advertises land and accounting AI analyst roles. The accounting role specifies AI-assisted workflows, document extraction, API/data-pipeline integration, acquisition-data normalization and migration, retries, fallbacks and exception queues. It calls for human review of ownership or legal determinations. These are staffing responsibilities, not confirmation that every workflow is deployed. OpenAI/Azure OpenAI are examples of desired candidate experience, not confirmed vendors.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-ACCOUNTING; OCT12-FLY-FEED", "Source_URL": "https://www.linkedin.com/jobs/view/ai-analyst-at-flywheel-energy-llc-4472009913", "Evidence": "Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-CONTEXT-DISC", "Initiative_ID": "AI-287", "Title": "Quorum land software: adjacent technology context", "Date": "Undated", "Finding": "Quorum names Flywheel as an On Demand Land customer for more than five years. Its testimonial describes easier data movement and access for nontechnical users. Quorum’s newer AI features are general product capabilities; Flywheel-specific use is not established.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-QUORUM", "Source_URL": "https://www.quorumsoftware.com/solutions/upstream-on-demand/land/", "Evidence": "Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-INTERN-DISC", "Initiative_ID": "AI-287", "Title": "Summer 2027 prototype agenda: indexed hiring lead", "Date": "2026", "Finding": "A search-indexed Summer 2027 internship description proposes LLM, agentic and MCP integration prototypes, plus ML work on production forecasting and predictive maintenance. The direct posting expired/redirected during review; these prospective tasks are not operating field AI.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-INTERN", "Source_URL": "https://www.linkedin.com/jobs/view/ai-intern-summer-2027-at-flywheel-energy-llc-4461575049", "Evidence": "Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}], "Latest_disclosure": "2026", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct12-flywheel-app-platform"], "Research_search": "{\"packet_id\": \"oct12-flywheel-app-platform\", \"key\": \"oct12-flywheel-app-platform\", \"record_ids\": [\"AI-287\"], \"company\": \"Flywheel Energy\", \"title\": \"Internal AI app platform and business-team automation\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-12\", \"verification\": \"verified primary\", \"stage\": \"Company-reported platform use; adoption unquantified\", \"business_problem\": \"Let business teams improve their own workflows without asking the central software team to build every application.\", \"workflow\": \"Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.\", \"technology_data\": \"Custom AI-enabled webapp platform; model, provider and implementation stack not disclosed.\", \"integration_human_role\": \"Employees build apps; engineering owns the platform, training, QA and transition of important apps to central support. This is not autonomous executive decision-making.\", \"deployment_scope\": \"Nontechnical business teammates and basin-team training; no basin, user or app population is quantified.\", \"economics\": \"Growing adoption is described qualitatively. No user/app count, quantified business result, complete cost or net return is disclosed.\", \"practical_relevance\": \"Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.\", \"assessment\": \"Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.\", \"limitations\": [\"The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect.\"], \"source_ids\": [\"OCT12-FLY-PLATFORM\", \"OCT12-FLY-ACCOUNTING\", \"OCT12-FLY-FEED\", \"OCT12-FLY-QUORUM\", \"OCT12-FLY-INTERN\", \"OCT12-FLY-COMPANY\"], \"partners\": [], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026\", \"event\": \"Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.\", \"source_ids\": [\"OCT12-FLY-PLATFORM\"]}], \"sources\": [{\"id\": \"OCT12-FLY-PLATFORM\", \"title\": \"Flywheel AI Platform Engineer\", \"url\": \"https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750\", \"date\": \"2026\", \"type\": \"Employer recruitment disclosure\", \"publisher\": \"Flywheel Energy / LinkedIn\", \"locator\": \"Job Summary and Key Responsibilities\", \"review_status\": \"Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.\"}, {\"id\": \"OCT12-FLY-ACCOUNTING\", \"title\": \"Flywheel AI Analyst — accounting automation\", \"url\": \"https://www.linkedin.com/jobs/view/ai-analyst-at-flywheel-energy-llc-4472009913\", \"date\": \"2026\", \"type\": \"Employer recruitment disclosure\", \"publisher\": \"Flywheel Energy / LinkedIn\", \"locator\": \"Role Summary; Automation; Data Management; Required Qualifications\", \"review_status\": \"Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments.\"}, {\"id\": \"OCT12-FLY-FEED\", \"title\": \"Flywheel recruitment update: land, accounting and AI platform\", \"url\": \"https://www.linkedin.com/company/flywheel-energy\", \"date\": \"2026\", \"type\": \"Employer recruitment disclosure\", \"publisher\": \"Flywheel Energy / LinkedIn\", \"locator\": \"AI recruitment update naming AI Analyst–Land, AI Analyst–Accounting and AI Platform Engineer\", \"review_status\": \"Company-authored LinkedIn feed update reviewed 12 October 2026; relative date only. Feed position may change.\"}, {\"id\": \"OCT12-FLY-QUORUM\", \"title\": \"Quorum On Demand Land — Flywheel testimonial\", \"url\": \"https://www.quorumsoftware.com/solutions/upstream-on-demand/land/\", \"date\": \"Undated\", \"type\": \"Vendor customer testimonial\", \"publisher\": \"Quorum Software\", \"locator\": \"Proven for High-Growth Land Organizations; Flywheel customer testimonial\", \"review_status\": \"Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel.\"}, {\"id\": \"OCT12-FLY-INTERN\", \"title\": \"Flywheel AI Intern — Summer 2027\", \"url\": \"https://www.linkedin.com/jobs/view/ai-intern-summer-2027-at-flywheel-energy-llc-4461575049\", \"date\": \"2026\", \"type\": \"Employer recruitment disclosure\", \"publisher\": \"Flywheel Energy / LinkedIn\", \"locator\": \"Indexed employer description: About the AI Internship and What You Will Do\", \"review_status\": \"Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source.\"}, {\"id\": \"OCT12-FLY-COMPANY\", \"title\": \"Flywheel Energy — company and operating partner context\", \"url\": \"https://www.flywheelenergy.com/\", \"date\": \"Undated\", \"type\": \"Company website\", \"publisher\": \"Flywheel Energy\", \"locator\": \"Private E&P description; Stone Ridge footer\", \"review_status\": \"Official company page reviewed 12 October 2026. Identifies Flywheel as an operating partner of Stone Ridge Energy; does not establish a precise legal ownership schedule or internal AI adoption.\"}], \"field_sources\": {\"business_problem\": [\"OCT12-FLY-PLATFORM\"], \"workflow\": [\"OCT12-FLY-PLATFORM\"], \"technology_data\": [\"OCT12-FLY-PLATFORM\"], \"integration_human_role\": [\"OCT12-FLY-PLATFORM\"], \"deployment_scope\": [\"OCT12-FLY-PLATFORM\"], \"economics\": [\"OCT12-FLY-PLATFORM\"]}, \"hiring_context\": \"Separately, Flywheel advertises land and accounting AI analyst roles. The accounting role specifies AI-assisted workflows, document extraction, API/data-pipeline integration, acquisition-data normalization and migration, retries, fallbacks and exception queues. It calls for human review of ownership or legal determinations. These are staffing responsibilities, not confirmation that every workflow is deployed. OpenAI/Azure OpenAI are examples of desired candidate experience, not confirmed vendors.\", \"adjacent_technology\": \"Quorum names Flywheel as an On Demand Land customer for more than five years. Its testimonial describes easier data movement and access for nontechnical users. Quorum’s newer AI features are general product capabilities; Flywheel-specific use is not established.\", \"prospective_research\": \"A search-indexed Summer 2027 internship description proposes LLM, agentic and MCP integration prototypes, plus ML work on production forecasting and predictive maintenance. The direct posting expired/redirected during review; these prospective tasks are not operating field AI.\"}", "Research_status": "Added 12 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/flywheel-oct12/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "Company-reported use; unquantified adoption", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect.", "Outcome_counted": false, "Outcome_count_note": "No numerical result; hiring scope and qualitative adoption do not count as results.", "Current_source_check": "2026-10-12", "Evidence_review": {"source_read": "Primary platform and accounting job descriptions and company feed reviewed; internship text indexed-only.", "claim_status": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "independent_validation": "Not established"}, "Detail_words": 196}, {"ID": "AI-288", "Company": "CNX Resources", "Company_primary": "CNX Resources", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Safety cameras that flag unsafe conditions", "Extracted_finding": "Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.", "Stage_at_disclosure": "Company describes current use; start date unknown", "Reported_metric": "No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.", "Interpretation_limit": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch.", "Source_IDs": "OCT14-CNX-SAFETY", "Source_date": "Undated", "Source_type": "Company website", "Source_locator": "Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "Evidence_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 targeted research", "Time_scope": "The current company page describes use, but gives no publication date or deployment start. Its 2025 safety milestones do not date the AI workflow.", "Context": "Identify unsafe conditions and improve safety decisions.", "Implementation": "Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "Assessment": "Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["14 October 2026 Luna research and editor original-source check"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "OCT14-AI-288", "Initiative_ID": "AI-288", "Title": "Safety cameras that flag unsafe conditions", "Date": "Undated", "Finding": "CNX says it uses AI video analytics to identify unsafe conditions, coach employees in real time and inform training. HSE++ also analyzes earlier incidents and safety observations to find trends. Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "Metric": "No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.", "Source_IDs": "OCT14-CNX-SAFETY", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "Evidence": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 targeted review", "Assessment": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch."}], "Latest_disclosure": "Undated", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14-ai-288"], "Research_search": "{\"packet_id\": \"oct14-ai-288\", \"key\": \"oct14-ai-288\", \"record_ids\": [\"AI-288\"], \"company\": \"CNX Resources\", \"title\": \"Safety cameras that flag unsafe conditions\", \"category\": \"Internal adoption\", \"department\": \"HSE & emissions\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Company describes current use; start date unknown\", \"business_problem\": \"Identify unsafe conditions and improve safety decisions.\", \"workflow\": \"Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.\", \"technology_data\": \"Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.\", \"deployment_scope\": \"The current company page describes use, but gives no publication date or deployment start. Its 2025 safety milestones do not date the AI workflow.\", \"economics\": \"No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.\", \"practical_relevance\": \"Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.\", \"assessment\": \"Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.\", \"integration_human_role\": \"Which unsafe conditions trigger useful interventions, how many alerts are rejected, and did incident rates change on comparable sites?\", \"limitations\": [\"Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch.\"], \"source_ids\": [\"OCT14-CNX-SAFETY\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"Undated\", \"event\": \"CNX says it uses AI video analytics to identify unsafe conditions, coach employees in real time and inform training. HSE++ also analyzes earlier incidents and safety observations to find trends.\", \"source_ids\": [\"OCT14-CNX-SAFETY\"]}], \"sources\": [{\"id\": \"OCT14-CNX-SAFETY\", \"title\": \"CNX Workplace Health and Safety\", \"url\": \"https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/\", \"date\": \"Undated\", \"type\": \"Company website\", \"publisher\": \"CNX Resources\", \"locator\": \"Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections\", \"review_status\": \"Relevant original source text reviewed 14 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"OCT14-CNX-SAFETY\"], \"workflow\": [\"OCT14-CNX-SAFETY\"], \"technology_data\": [\"OCT14-CNX-SAFETY\"], \"deployment_scope\": [\"OCT14-CNX-SAFETY\"], \"economics\": [\"OCT14-CNX-SAFETY\"], \"integration_human_role\": [\"OCT14-CNX-SAFETY\"], \"practical_relevance\": [\"OCT14-CNX-SAFETY\"]}}", "Research_status": "Added 14 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/gap-swarm-oct14/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "Deployment population undisclosed", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch.", "Outcome_counted": false, "Outcome_count_note": "No achieved numerical AI outcome disclosed for this workflow.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original text reviewed", "claim_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 33}, {"ID": "AI-289", "Company": "Aker BP", "Company_primary": "Aker BP", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Agents for risk assessment and alarm reviews", "Extracted_finding": "This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.", "Stage_at_disclosure": "Implementation announced; operating results not established", "Reported_metric": "No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.", "Interpretation_limit": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed.", "Source_IDs": "OCT14-AKER-ABB", "Source_date": "2026-05-21", "Source_type": "Vendor announcement", "Source_locator": "Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits", "Source_URL": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "Evidence_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 targeted research", "Time_scope": "Published 21 May 2026. Aker BP signed on to implement; completion date, site scope and sustained live use are not established.", "Context": "Coordinate operational risk assessment and alarm reviews across industrial systems.", "Implementation": "The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "Assessment": "This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.", "Primary_department": "HSE & emissions", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["14 October 2026 Luna research and editor original-source check"], "Reference_IDs": [], "Related_IDs": ["AI-049"], "Disclosures": [{"ID": "OCT14-AI-289", "Initiative_ID": "AI-289", "Title": "Agents for risk assessment and alarm reviews", "Date": "2026-05-21", "Finding": "Cognite and ABB name Aker BP as the first customer signed on to implement an agentic offering for risk assessment and alarm rationalization. The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "Metric": "No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.", "Source_IDs": "OCT14-AKER-ABB", "Source_URL": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "Evidence": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 targeted review", "Assessment": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed."}], "Latest_disclosure": "2026-05-21", "Has_pending_disclosures": true, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14-ai-289"], "Research_search": "{\"packet_id\": \"oct14-ai-289\", \"key\": \"oct14-ai-289\", \"record_ids\": [\"AI-289\"], \"company\": \"Aker BP\", \"title\": \"Agents for risk assessment and alarm reviews\", \"category\": \"Internal adoption\", \"department\": \"HSE & emissions\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Implementation announced; operating results not established\", \"business_problem\": \"Coordinate operational risk assessment and alarm reviews across industrial systems.\", \"workflow\": \"The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.\", \"technology_data\": \"The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.\", \"deployment_scope\": \"Published 21 May 2026. Aker BP signed on to implement; completion date, site scope and sustained live use are not established.\", \"economics\": \"No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.\", \"practical_relevance\": \"This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.\", \"assessment\": \"This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.\", \"integration_human_role\": \"Which risk and alarm reviews are live, what actions can agents take, and where are human approval and safety-system boundaries?\", \"limitations\": [\"Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed.\"], \"source_ids\": [\"OCT14-AKER-ABB\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-05-21\", \"event\": \"Cognite and ABB name Aker BP as the first customer signed on to implement an agentic offering for risk assessment and alarm rationalization.\", \"source_ids\": [\"OCT14-AKER-ABB\"]}], \"sources\": [{\"id\": \"OCT14-AKER-ABB\", \"title\": \"Cognite and ABB collaborate on industrial agentic applications\", \"url\": \"https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver\", \"date\": \"2026-05-21\", \"type\": \"Vendor announcement\", \"publisher\": \"Cognite\", \"locator\": \"Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits\", \"review_status\": \"Relevant original source text reviewed 14 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"OCT14-AKER-ABB\"], \"workflow\": [\"OCT14-AKER-ABB\"], \"technology_data\": [\"OCT14-AKER-ABB\"], \"deployment_scope\": [\"OCT14-AKER-ABB\"], \"economics\": [\"OCT14-AKER-ABB\"], \"integration_human_role\": [\"OCT14-AKER-ABB\"], \"practical_relevance\": [\"OCT14-AKER-ABB\"]}}", "Research_status": "Added 14 October 2026", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/gap-swarm-oct14/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "Deployment population undisclosed", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed.", "Outcome_counted": false, "Outcome_count_note": "No achieved numerical AI outcome disclosed for this workflow.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original text reviewed", "claim_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 44}, {"ID": "AI-290", "Company": "PetroChina", "Company_primary": "PetroChina", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "A model that diagnoses rod-pumped wells in Xinjiang", "Extracted_finding": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.", "Stage_at_disclosure": "Company reports scaled use on more than 3,800 wells", "Reported_metric": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "Interpretation_limit": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain.", "Source_IDs": "OCT14B-CNPC-XINJIANG", "Source_date": "2026-04-15", "Source_type": "Company news article", "Source_locator": "Article dated 15 April 2026; first four body paragraphs", "Source_URL": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 second targeted review", "Time_scope": "15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.", "Context": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.", "Implementation": "A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "Assessment": "This connects a large industrial model to a specific engineering job: diagnose abnormal wells and explain what to inspect. The published results do not establish autonomous operation or a net return.", "Primary_department": "Production & field operations", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["14 October 2026 second Luna pass; editor original-source checks"], "Reference_IDs": [], "Related_IDs": ["AI-081", "AI-236"], "Disclosures": [{"ID": "OCT14B-AI-290", "Initiative_ID": "AI-290", "Title": "A model that diagnoses rod-pumped wells in Xinjiang", "Date": "2026-04-15", "Finding": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions. A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "Metric": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "Source_IDs": "OCT14B-CNPC-XINJIANG", "Source_URL": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain."}], "Latest_disclosure": "2026-04-15", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14b-ai-290"], "Research_search": "{\"packet_id\": \"oct14b-ai-290\", \"key\": \"oct14b-ai-290\", \"record_ids\": [\"AI-290\"], \"company\": \"PetroChina\", \"title\": \"A model that diagnoses rod-pumped wells in Xinjiang\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Company reports scaled use on more than 3,800 wells\", \"business_problem\": \"CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.\", \"workflow\": \"A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.\", \"technology_data\": \"A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.\", \"deployment_scope\": \"15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.\", \"integration_human_role\": \"Staff review the outputs and act on exceptions; autonomous control is not established.\", \"economics\": \"Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.\", \"practical_relevance\": \"This connects a large industrial model to a specific engineering job: diagnose abnormal wells and explain what to inspect. The published results do not establish autonomous operation or a net return.\", \"assessment\": \"This connects a large industrial model to a specific engineering job: diagnose abnormal wells and explain what to inspect. The published results do not establish autonomous operation or a net return.\", \"limitations\": [\"Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain.\"], \"source_ids\": [\"OCT14B-CNPC-XINJIANG\"], \"disclosed_results\": [{\"metric\": \"Source-reported workflow results\", \"value\": \"Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.\", \"period\": \"15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.\", \"attribution\": \"CNPC / China Petroleum News; not independently validated\", \"source_ids\": [\"OCT14B-CNPC-XINJIANG\"]}], \"timeline\": [{\"date\": \"2026-04-15\", \"event\": \"CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.\", \"source_ids\": [\"OCT14B-CNPC-XINJIANG\"]}], \"sources\": [{\"id\": \"OCT14B-CNPC-XINJIANG\", \"title\": \"CNPC: industrial large model for rod-pumped-well production optimization\", \"url\": \"https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml\", \"date\": \"2026-04-15\", \"type\": \"Company news article\", \"publisher\": \"CNPC / China Petroleum News\", \"locator\": \"Article dated 15 April 2026; first four body paragraphs\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"OCT14B-CNPC-XINJIANG\"], \"workflow\": [\"OCT14B-CNPC-XINJIANG\"], \"technology_data\": [\"OCT14B-CNPC-XINJIANG\"], \"deployment_scope\": [\"OCT14B-CNPC-XINJIANG\"], \"integration_human_role\": [\"OCT14B-CNPC-XINJIANG\"], \"economics\": [\"OCT14B-CNPC-XINJIANG\"], \"practical_relevance\": [\"OCT14B-CNPC-XINJIANG\"]}}", "Research_status": "Added 14 October 2026 · second targeted pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/gap-swarm-oct14b/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "scaled", "Scale_note": "More than 3,800 operating wells; 78,000-well data set is separate", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain.", "Outcome_counted": true, "Outcome_count_note": "Counts one source-reported numerical-result profile, not independently validated performance or net cash ROI.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 57}, {"ID": "AI-291", "Company": "CNX Resources", "Company_primary": "CNX Resources", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Finding methane-release precursors and prioritizing repairs", "Extracted_finding": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.", "Stage_at_disclosure": "Company describes current use; exact start date unknown", "Reported_metric": "CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.", "Interpretation_limit": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority.", "Source_IDs": "OCT14B-CNX-METHANE", "Source_date": "Undated", "Source_type": "Company website", "Source_locator": "2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 second targeted review", "Time_scope": "Undated current company page, with explicit 2026 program and Q1 2026 progress. No model launch date or deployment population is given.", "Context": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.", "Implementation": "Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "Assessment": "Methane response involves prediction, source identification and repair. The next useful evidence is whether this integration prevents releases or shortens the work needed to fix them.", "Primary_department": "HSE & emissions", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["14 October 2026 second Luna pass; editor original-source checks"], "Reference_IDs": [], "Related_IDs": ["AI-288"], "Disclosures": [{"ID": "OCT14B-AI-291", "Initiative_ID": "AI-291", "Title": "Finding methane-release precursors and prioritizing repairs", "Date": "Undated", "Finding": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory. Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "Metric": "CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.", "Source_IDs": "OCT14B-CNX-METHANE", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority."}], "Latest_disclosure": "Undated", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14b-ai-291"], "Research_search": "{\"packet_id\": \"oct14b-ai-291\", \"key\": \"oct14b-ai-291\", \"record_ids\": [\"AI-291\"], \"company\": \"CNX Resources\", \"title\": \"Finding methane-release precursors and prioritizing repairs\", \"category\": \"Internal adoption\", \"department\": \"HSE & emissions\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Company describes current use; exact start date unknown\", \"business_problem\": \"CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.\", \"workflow\": \"Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.\", \"technology_data\": \"Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.\", \"deployment_scope\": \"Undated current company page, with explicit 2026 program and Q1 2026 progress. No model launch date or deployment population is given.\", \"integration_human_role\": \"Staff review the outputs and act on exceptions; autonomous control is not established.\", \"economics\": \"CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.\", \"practical_relevance\": \"Methane response involves prediction, source identification and repair. The next useful evidence is whether this integration prevents releases or shortens the work needed to fix them.\", \"assessment\": \"Methane response involves prediction, source identification and repair. The next useful evidence is whether this integration prevents releases or shortens the work needed to fix them.\", \"limitations\": [\"This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority.\"], \"source_ids\": [\"OCT14B-CNX-METHANE\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"Undated\", \"event\": \"CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.\", \"source_ids\": [\"OCT14B-CNX-METHANE\"]}], \"sources\": [{\"id\": \"OCT14B-CNX-METHANE\", \"title\": \"CNX Commitment to Methane Reduction\", \"url\": \"https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/\", \"date\": \"Undated\", \"type\": \"Company website\", \"publisher\": \"CNX Resources\", \"locator\": \"2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"field_sources\": {\"business_problem\": [\"OCT14B-CNX-METHANE\"], \"workflow\": [\"OCT14B-CNX-METHANE\"], \"technology_data\": [\"OCT14B-CNX-METHANE\"], \"deployment_scope\": [\"OCT14B-CNX-METHANE\"], \"integration_human_role\": [\"OCT14B-CNX-METHANE\"], \"economics\": [\"OCT14B-CNX-METHANE\"], \"practical_relevance\": [\"OCT14B-CNX-METHANE\"]}}", "Research_status": "Added 14 October 2026 · second targeted pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/gap-swarm-oct14b/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "AI-specific population undisclosed", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority.", "Outcome_counted": false, "Outcome_count_note": "No numerical AI-specific outcome disclosed.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 50}, {"ID": "AI-292", "Company": "HG Oil Holdings", "Company_primary": "HG Oil Holdings", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "An invoice assistant with human exception review", "Extracted_finding": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.", "Stage_at_disclosure": "Undated vendor-described use; benefits not independently verified", "Reported_metric": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "Interpretation_limit": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata.", "Source_IDs": "OCT14B-HG-INVOICES; OCT14B-HG-IDENTITY; OCT14B-HG-CONTROLS", "Source_date": "Undated", "Source_type": "Vendor customer case", "Source_locator": "Before; What we built; Operational handoffs; Outcomes", "Source_URL": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 second targeted review", "Time_scope": "Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.", "Context": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.", "Implementation": "Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "Assessment": "Invoice processing is a practical AI use case for smaller operators. Time released to other work is valuable capacity; it becomes a cash benefit only when the business avoids costs or improves an outcome.", "Primary_department": "Finance, investment & strategy", "Secondary_departments": [], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["14 October 2026 second Luna pass; editor original-source checks"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "OCT14B-AI-292", "Initiative_ID": "AI-292", "Title": "An invoice assistant with human exception review", "Date": "Undated", "Finding": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review. Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "Metric": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "Source_IDs": "OCT14B-HG-INVOICES", "Source_URL": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata."}], "Latest_disclosure": "Undated", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14b-ai-292"], "Research_search": "{\"packet_id\": \"oct14b-ai-292\", \"key\": \"oct14b-ai-292\", \"record_ids\": [\"AI-292\"], \"company\": \"HG Oil Holdings\", \"title\": \"An invoice assistant with human exception review\", \"category\": \"Internal adoption\", \"department\": \"Finance, investment & strategy\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Undated vendor-described use; benefits not independently verified\", \"business_problem\": \"Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.\", \"workflow\": \"Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.\", \"technology_data\": \"Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.\", \"deployment_scope\": \"Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.\", \"integration_human_role\": \"Staff review the outputs and act on exceptions; autonomous control is not established.\", \"economics\": \"Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.\", \"practical_relevance\": \"Invoice processing is a practical AI use case for smaller operators. Time released to other work is valuable capacity; it becomes a cash benefit only when the business avoids costs or improves an outcome.\", \"assessment\": \"Invoice processing is a practical AI use case for smaller operators. Time released to other work is valuable capacity; it becomes a cash benefit only when the business avoids costs or improves an outcome.\", \"limitations\": [\"Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata.\"], \"source_ids\": [\"OCT14B-HG-INVOICES\", \"OCT14B-HG-IDENTITY\", \"OCT14B-HG-CONTROLS\"], \"disclosed_results\": [{\"metric\": \"Source-reported workflow results\", \"value\": \"Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.\", \"period\": \"Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.\", \"attribution\": \"Preisser Solutions; not independently validated\", \"source_ids\": [\"OCT14B-HG-INVOICES\"]}], \"timeline\": [{\"date\": \"Undated\", \"event\": \"Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.\", \"source_ids\": [\"OCT14B-HG-INVOICES\"]}], \"sources\": [{\"id\": \"OCT14B-HG-INVOICES\", \"title\": \"HG Oil Holdings: AI invoice processing\", \"url\": \"https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing\", \"date\": \"Undated\", \"type\": \"Vendor customer case\", \"publisher\": \"Preisser Solutions\", \"locator\": \"Before; What we built; Operational handoffs; Outcomes\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}, {\"id\": \"OCT14B-HG-IDENTITY\", \"title\": \"HG Oil Holdings About\", \"url\": \"https://www.hgoilholdings.com/about\", \"date\": \"Undated\", \"type\": \"Company website\", \"publisher\": \"HG Oil Holdings\", \"locator\": \"Company description; current company-wide wells are not AI scope\", \"review_status\": \"Original page reviewed 14 October 2026; adjacent context, not independent outcome validation.\"}, {\"id\": \"OCT14B-HG-CONTROLS\", \"title\": \"Preisser AI invoice processing for small businesses\", \"url\": \"https://preissersolutions.com/use-cases/ai-invoice-processing-small-business\", \"date\": \"Undated\", \"type\": \"Vendor product page\", \"publisher\": \"Preisser Solutions\", \"locator\": \"Human payment approval; names HG as reference; general product controls\", \"review_status\": \"Original page reviewed 14 October 2026; adjacent context, not independent outcome validation.\"}], \"field_sources\": {\"business_problem\": [\"OCT14B-HG-INVOICES\"], \"workflow\": [\"OCT14B-HG-INVOICES\"], \"technology_data\": [\"OCT14B-HG-INVOICES\"], \"deployment_scope\": [\"OCT14B-HG-INVOICES\"], \"integration_human_role\": [\"OCT14B-HG-INVOICES\", \"OCT14B-HG-CONTROLS\"], \"economics\": [\"OCT14B-HG-INVOICES\"], \"practical_relevance\": [\"OCT14B-HG-INVOICES\"]}}", "Research_status": "Added 14 October 2026 · second targeted pass", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/gap-swarm-oct14b/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "AI-specific population undisclosed", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata.", "Outcome_counted": true, "Outcome_count_note": "Counts one source-reported numerical-result profile, not independently validated performance or net cash ROI.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 60}, {"ID": "AI-293", "Company": "Aramco", "Company_primary": "Aramco", "Segment": "Upstream", "Theme": "Geological knowledge", "Initiative": "Connecting rock samples to a geological knowledge platform", "Extracted_finding": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.", "Stage_at_disclosure": "Company describes use; additional ML work is underway", "Reported_metric": "No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.", "Interpretation_limit": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.", "Source_IDs": "OCT14C-ARAMCO-CORE", "Source_date": "2026-08-25", "Source_type": "Company magazine article", "Source_locator": "25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration", "Source_URL": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 targeted research", "Time_scope": "Published 25 August 2026. The center’s 150+ employees and archive of more than one million feet of samples are facility context, not AI deployment or validation denominators.", "Context": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.", "Implementation": "Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "Assessment": "Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?", "Primary_department": "Exploration & geoscience", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["Third October 14 targeted review"], "Reference_IDs": [], "Related_IDs": ["AI-281", "AI-282"], "Disclosures": [{"ID": "OCT14C-AI-293", "Initiative_ID": "AI-293", "Title": "Connecting rock samples to a geological knowledge platform", "Date": "2026-08-25", "Finding": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists. Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "Metric": "No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.", "Source_IDs": "OCT14C-ARAMCO-CORE", "Source_URL": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established."}], "Latest_disclosure": "2026-08-25", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["oct14c-ai-293"], "Research_search": "{\"packet_id\": \"oct14c-ai-293\", \"key\": \"oct14c-ai-293\", \"record_ids\": [\"AI-293\"], \"company\": \"Aramco\", \"title\": \"Connecting rock samples to a geological knowledge platform\", \"category\": \"Internal adoption\", \"department\": \"Exploration & geoscience\", \"report_date\": \"2026-10-14\", \"verification\": \"verified primary\", \"stage\": \"Company describes use; additional ML work is underway\", \"business_problem\": \"Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.\", \"workflow\": \"Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.\", \"technology_data\": \"Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.\", \"deployment_scope\": \"Published 25 August 2026. The center’s 150+ employees and archive of more than one million feet of samples are facility context, not AI deployment or validation denominators.\", \"integration_human_role\": \"Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.\", \"economics\": \"No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.\", \"practical_relevance\": \"Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?\", \"assessment\": \"Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?\", \"limitations\": [\"Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.\"], \"source_ids\": [\"OCT14C-ARAMCO-CORE\"], \"sources\": [{\"id\": \"OCT14C-ARAMCO-CORE\", \"title\": \"Understanding the subsurface: inside Aramco’s Core Laboratories Center\", \"url\": \"https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center\", \"date\": \"2026-08-25\", \"type\": \"Company magazine article\", \"publisher\": \"Aramco\", \"locator\": \"25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-08-25\", \"event\": \"Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.\", \"source_ids\": [\"OCT14C-ARAMCO-CORE\"]}], \"field_sources\": {\"business_problem\": [\"OCT14C-ARAMCO-CORE\"], \"workflow\": [\"OCT14C-ARAMCO-CORE\"], \"technology_data\": [\"OCT14C-ARAMCO-CORE\"], \"deployment_scope\": [\"OCT14C-ARAMCO-CORE\"], \"economics\": [\"OCT14C-ARAMCO-CORE\"]}}", "Research_status": "Clarified 14 October 2026 · third pass", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/gap-swarm-oct14c/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "AI population undisclosed; facility context is separate", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.", "Outcome_counted": false, "Outcome_count_note": "No quantified AI outcome supplied.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original source reviewed", "claim_status": "Company describes use; additional ML work is underway", "independent_validation": "Not established"}, "Detail_words": 35}, {"ID": "AI-294", "Company": "INPEX", "Company_primary": "INPEX", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "Employees build agents for research, rules and management meetings", "Extracted_finding": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.", "Stage_at_disclosure": "Shared employee agents in use; active-agent population unknown", "Reported_metric": "Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.", "Interpretation_limit": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.", "Source_IDs": "WATCH14PM-INPEX", "Source_date": "2026-09-01", "Source_type": "Vendor customer story", "Source_locator": "Research tool; Building a foundation for change; Looking beyond estimated annual value", "Source_URL": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "Evidence_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "PE_relationship": null, "Coverage_bucket": "14 October 2026 targeted research", "Time_scope": "Published 1 September 2026; Copilot became the core platform in 2025. Exact measurement windows, active-agent count, survey size and value calculation are not supplied.", "Context": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.", "Implementation": "SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "Assessment": "Which agents do people keep using, and does better research or meeting preparation change a decision?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Workforce & learning", "Finance, investment & strategy"], "Department_assignment_basis": "Analyst functional mapping from the source; organizational owner not confirmed.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "No achieved result or quantified scale established for this workflow.", "Review_tier": "Reviewed source", "Provenance": ["Scheduled research watch / 14 October evening"], "Reference_IDs": [], "Related_IDs": ["AI-240"], "Disclosures": [{"ID": "WATCH14PM-AI-294", "Initiative_ID": "AI-294", "Title": "Employees build agents for research, rules and management meetings", "Date": "2026-09-01", "Finding": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions. SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "Metric": "Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.", "Source_IDs": "WATCH14PM-INPEX", "Source_URL": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method."}], "Latest_disclosure": "2026-09-01", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch14pm-ai-294"], "Research_search": "{\"packet_id\": \"watch14pm-ai-294\", \"key\": \"watch14pm-ai-294\", \"record_ids\": [\"AI-294\"], \"company\": \"INPEX\", \"title\": \"Employees build agents for research, rules and management meetings\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-14\", \"verification\": \"verified partner\", \"stage\": \"Shared employee agents in use; active-agent population unknown\", \"business_problem\": \"INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.\", \"workflow\": \"SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.\", \"technology_data\": \"SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.\", \"deployment_scope\": \"Published 1 September 2026; Copilot became the core platform in 2025. Exact measurement windows, active-agent count, survey size and value calculation are not supplied.\", \"integration_human_role\": \"Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.\", \"economics\": \"Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.\", \"practical_relevance\": \"Which agents do people keep using, and does better research or meeting preparation change a decision?\", \"assessment\": \"Which agents do people keep using, and does better research or meeting preparation change a decision?\", \"limitations\": [\"Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.\"], \"source_ids\": [\"WATCH14PM-INPEX\"], \"sources\": [{\"id\": \"WATCH14PM-INPEX\", \"title\": \"INPEX transforms business processes with Microsoft 365 Copilot and AI agents\", \"url\": \"https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot\", \"date\": \"2026-09-01\", \"type\": \"Vendor customer story\", \"publisher\": \"Microsoft / named INPEX speakers\", \"locator\": \"Research tool; Building a foundation for change; Looking beyond estimated annual value\", \"review_status\": \"Relevant original source reviewed 14 October 2026; claims remain source-reported.\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-09-01\", \"event\": \"INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.\", \"source_ids\": [\"WATCH14PM-INPEX\"]}], \"field_sources\": {\"business_problem\": [\"WATCH14PM-INPEX\"], \"workflow\": [\"WATCH14PM-INPEX\"], \"technology_data\": [\"WATCH14PM-INPEX\"], \"deployment_scope\": [\"WATCH14PM-INPEX\"], \"economics\": [\"WATCH14PM-INPEX\"]}}", "Research_status": "Reviewed 14 October 2026 · evening watch", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/watch-2026-10-14-pm/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "scaled", "Scale_note": "1,450 paid licenses and 1,000+ agents reported; active use unspecified", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.", "Outcome_counted": false, "Outcome_count_note": "License/agent counts are adoption; survey perceptions and company-wide estimated value are not independently measured workflow outcomes. No new numerical-result count.", "Current_source_check": "2026-10-14", "Evidence_review": {"source_read": "Relevant original text reviewed", "claim_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "independent_validation": "Not established"}, "Detail_words": 33}, {"ID": "AI-295", "Company": "Kuwait Oil Company", "Company_primary": "Kuwait Oil Company", "Segment": "NOC", "Theme": "Drilling knowledge", "Initiative": "SANAD helps drilling engineers find the procedure they need", "Extracted_finding": "KOC describes SANAD, a drilling knowledge assistant developed with SLB and piloted by its North & West Kuwait Drilling Group.", "Stage_at_disclosure": "Pilot and validation; later field test reported, wider rollout unverified", "Reported_metric": "A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.", "Interpretation_limit": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "Source_IDs": "WATCH18AM-KOC-DIGEST; WATCH18AM-KOC-INTERVIEW", "Source_date": "2026-09-29", "Source_type": "Operator implementation account and attributed engineer interview", "Source_locator": "Official Digest: PDF pages 28–31 / printed 26–29; September interview: field-test and decision-responsibility paragraphs", "Source_URL": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "Evidence_status": "Original KOC publication read; later attributed engineer interview reviewed. Outcomes not independently validated.", "PE_relationship": null, "Coverage_bucket": "18 October 2026 targeted research", "Time_scope": "Development reportedly began March 2025. KOC describes a 2025–26 pilot; its April–June 2026 issue has no verified posting day. The interview was published 29 September 2026. Field-test dates and well ID are not supplied.", "Context": "Find relevant drilling procedures, well knowledge and lessons without relying on repeated manual searches.", "Implementation": "The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision. LLM and retrieval-augmented generation over selected drilling procedures and records. SLB supported model training and data processing; KOC Innovation & Digitalization provided technical support. OpenWELL context was tested during the pilot; full integration is not established.", "Assessment": "The useful step is a drilling answer grounded in the right procedure and checked by an engineer. The public evidence establishes a bounded pilot and an attributed later field test, not a company-wide brain or isolated AI savings.", "Primary_department": "Drilling & completions", "Secondary_departments": ["IT, data & enterprise transformation"], "Department_assignment_basis": "KOC names Drilling Group North & West Kuwait as pilot owner; functional mapping by editor.", "Scope_type": "Defined operational or business workflow", "Ambition_shortlist_category": "Not selected for the seven-category shortlist", "Ambition_assessment_basis": "Pilot validation is described; broader modules remain a roadmap and no AI-isolated benefit is established.", "Review_tier": "Reviewed source", "Provenance": ["18 October 2026 Luna research and editor original-source check"], "Reference_IDs": [], "Related_IDs": ["AI-085"], "Disclosures": [{"ID": "WATCH18AM-AI-295", "Initiative_ID": "AI-295", "Title": "SANAD helps drilling engineers find the procedure they need", "Date": "2026-09-29", "Finding": "KOC describes SANAD, a drilling knowledge assistant developed with SLB and piloted by its North & West Kuwait Drilling Group. The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision.", "Metric": "A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.", "Source_IDs": "WATCH18AM-KOC-DIGEST; WATCH18AM-KOC-INTERVIEW", "Source_URL": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "Evidence": "Original KOC publication read; later attributed engineer interview reviewed. Outcomes not independently validated.", "Origin": "18 October 2026 targeted source review", "Assessment": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085."}], "Latest_disclosure": "2026-09-29", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch18am-ai-295"], "Research_search": "{\"packet_id\": \"watch18am-ai-295\", \"key\": \"watch18am-ai-295\", \"record_ids\": [\"AI-295\"], \"company\": \"Kuwait Oil Company\", \"title\": \"SANAD helps drilling engineers find the procedure they need\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"report_date\": \"2026-10-18\", \"verification\": \"verified primary\", \"stage\": \"Pilot and validation; later field test reported, wider rollout unverified\", \"business_problem\": \"Find relevant drilling procedures, well knowledge and lessons without relying on repeated manual searches.\", \"workflow\": \"The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision.\", \"technology_data\": \"LLM and retrieval-augmented generation over selected drilling procedures and records. SLB supported model training and data processing; KOC Innovation & Digitalization provided technical support. OpenWELL context was tested during the pilot; full integration is not established.\", \"deployment_scope\": \"Development reportedly began March 2025. KOC describes a 2025–26 pilot; its April–June 2026 issue has no verified posting day. The interview was published 29 September 2026. Field-test dates and well ID are not supplied. The pilot belonged to Drilling Group North & West Kuwait. The reported field test concerns one Lower Burgan production well; its identifier is undisclosed.\", \"integration_human_role\": \"KOC says its standards team validated every pilot output against official procedures. Engineering judgment and responsibility remain with the specialist. Retrieval and recommendations do not establish authority to control a rig.\", \"economics\": \"A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.\", \"practical_relevance\": \"Can engineers trace each answer to the right procedure, and does it improve decisions after review time and errors are included?\", \"assessment\": \"The useful step is a drilling answer grounded in the right procedure and checked by an engineer. The public evidence establishes a bounded pilot and an attributed later field test, not a company-wide brain or isolated AI savings.\", \"limitations\": [\"No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.\", \"SOP savings are not an isolated SANAD effect. No complete cost or net ROI established.\"], \"source_ids\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"sources\": [{\"id\": \"WATCH18AM-KOC-DIGEST\", \"title\": \"SANAD AI: The Generative AI Drilling Assistant\", \"url\": \"https://www.kockw.com/sites/EN/Periodicals/Digest%20April%20-%20June%202026.pdf#page=28\", \"date\": \"2026-Q2\", \"publisher\": \"Kuwait Oil Company\", \"locator\": \"PDF pages 28–31 / printed 26–29; 14 SOPs on PDF 30; validation, ownership and roadmap on PDF 31\", \"access\": \"Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.\", \"excerpt\": \"Every output was validated\", \"review_status\": \"Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.\", \"accessed\": \"2026-10-18\"}, {\"id\": \"WATCH18AM-KOC-INTERVIEW\", \"title\": \"KOC’s Sanad uses AI to unlock drilling knowledge\", \"url\": \"https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/\", \"date\": \"2026-09-29\", \"publisher\": \"KUNA interview, republished by Kuwait Times\", \"locator\": \"Timestamp; lower-Burgan field test, SOP-related savings, RAG description and human responsibility\", \"access\": \"Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.\", \"excerpt\": \"Engineering judgment and responsibility remain with the specialist\", \"review_status\": \"Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.\", \"accessed\": \"2026-10-18\"}], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-Q2\", \"event\": \"KOC Digest describes the 2025–26 pilot, 14 North Kuwait SOPs, standards-team validation and future modules. Exact posting day unknown.\", \"source_ids\": [\"WATCH18AM-KOC-DIGEST\"]}, {\"date\": \"2026-09-29\", \"event\": \"Attributed engineer interview reports a Lower Burgan field test. SOP-related savings do not isolate AI contribution.\", \"source_ids\": [\"WATCH18AM-KOC-INTERVIEW\"]}], \"field_sources\": {\"business_problem\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"workflow\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"technology_data\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"deployment_scope\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"integration_human_role\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"economics\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"], \"practical_relevance\": [\"WATCH18AM-KOC-DIGEST\", \"WATCH18AM-KOC-INTERVIEW\"]}}", "Research_status": "Added 18 October 2026 · supervised pilot and attributed field-test account", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/watch-2026-10-18-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Primary publication", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "Pilot grounded in 14 SOPs; later one-well field test reported. Active-user and rollout population undisclosed.", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "Outcome_counted": false, "Outcome_count_note": "13 days and KD 97,000 relate to an attributed field test / SOP application; no isolated AI benefit. Excluded from counted numerical AI results and ROI.", "Current_source_check": "2026-10-18", "Evidence_review": {"source_read": "verified primary", "claim_status": "Bounded pilot and attributed field test", "independent_validation": "Not established by this collection"}, "Detail_words": 165}, {"ID": "AI-296", "Company": "U.S. Energy Development Corporation", "Company_primary": "U.S. Energy Development Corporation", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Closed-loop ESP optimization on 12 Permian wells", "Extracted_finding": "U.S. Energy and ROAM-AI describe automatic ESP adjustments on 12 Permian wells. Engineers define the operating limits and handle exceptions.", "Stage_at_disclosure": "Operating field deployment; operator/vendor-reported results", "Reported_metric": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "Interpretation_limit": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case. The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion. Return definition, full initial/recurring costs, realized cash, adjustment for other operating changes and matched controls are undisclosed. The article says engineering work changed toward exceptions, not that headcount or total workload fell.", "Source_IDs": "WATCH20PM-USEDC", "Source_date": "2026-07", "Source_type": "Operator/vendor-coauthored field account", "Source_locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "Source_URL": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "Evidence_status": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.", "PE_relationship": null, "Coverage_bucket": "20 October evening named-field review", "Time_scope": "Started December 2025. July 2026 article; exact publication day unknown. Six-month results, not a year or the operator’s full fleet. This older case was missed in earlier research.", "Context": "Well conditions change faster than an engineering team can manually review and execute every small ESP adjustment.", "Implementation": "U.S. Energy and ROAM-AI describe automatic ESP adjustments on 12 Permian wells. Engineers define the operating limits and handle exceptions. Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed. 12 Permian Basin wells; activated December 2025. Six-month evaluation against a 90-day pre-activation baseline; exact activation and publication days unknown.", "Assessment": "Shows how execution permissions and exception handling can matter as much as recommendation quality. Test the baseline and all-in economics before copying the workflow.", "Primary_department": "Production & field operations", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst workflow mapping; operator organizational structure not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["20 October Luna discovery; editor publisher-source and adversarial review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH20PM-AI-296", "Initiative_ID": "AI-296", "Title": "Twelve wells moved from recommendations to automatic adjustments", "Date": "2026-07", "Finding": "U.S. Energy and ROAM-AI describe automatic ESP adjustments on 12 Permian wells. Engineers define the operating limits and handle exceptions. Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed.", "Metric": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "Source_IDs": "WATCH20PM-USEDC", "Source_URL": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "Evidence": "Original operator/vendor account read; results not independently validated", "Origin": "20 October evening Luna research; editor original-source review", "Assessment": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case."}], "Latest_disclosure": "2026-07", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch20pm-ai-296"], "Research_search": "{\"packet_id\": \"watch20pm-ai-296\", \"key\": \"watch20pm-ai-296\", \"record_ids\": [\"AI-296\"], \"company\": \"U.S. Energy Development Corporation\", \"title\": \"Twelve wells moved from recommendations to automatic adjustments\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-20\", \"verification\": \"verified partner\", \"stage\": \"Operating field deployment; operator/vendor-reported results\", \"business_problem\": \"Well conditions change faster than an engineering team can manually review and execute every small ESP adjustment.\", \"workflow\": \"Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed.\", \"technology_data\": \"Physics-aware AI, live operating data, edge hardware and bounded ESP setpoint control. No language model identified.\", \"integration_human_role\": \"Engineers define operating envelopes and review exceptions. In-envelope adjustments execute automatically; exact override, shutdown and fallback procedures are undisclosed.\", \"deployment_scope\": \"12 Permian Basin wells; activated December 2025. Six-month evaluation against a 90-day pre-activation baseline; exact activation and publication days unknown.\", \"economics\": \"The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.\", \"practical_relevance\": \"Ask for the all-in cost and benefit bridge, the baseline calculation and the log of exceptions, overrides and rejected adjustments.\", \"assessment\": \"Shows how execution permissions and exception handling can matter as much as recommendation quality. Test the baseline and all-in economics before copying the workflow.\", \"limitations\": [\"Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case. The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion. Return definition, full initial/recurring costs, realized cash, adjustment for other operating changes and matched controls are undisclosed. The article says engineering work changed toward exceptions, not that headcount or total workload fell.\"], \"source_ids\": [\"WATCH20PM-USEDC\"], \"partners\": [\"ROAM-AI\"], \"disclosed_results\": [{\"metric\": \"Author-reported production and return\", \"value\": \"The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.\", \"period\": \"First six months; production comparator is preceding 90 days\", \"attribution\": \"Operator/vendor coauthored; not independent\", \"source_ids\": [\"WATCH20PM-USEDC\"]}], \"timeline\": [{\"date\": \"2025-12\", \"event\": \"12-well activation, per operator coauthor.\", \"source_ids\": [\"WATCH20PM-USEDC\"]}, {\"date\": \"2026-07\", \"event\": \"July cover story; exact publication day not established.\", \"source_ids\": [\"WATCH20PM-USEDC\"]}], \"sources\": [{\"id\": \"WATCH20PM-USEDC\", \"title\": \"The Bottleneck Isn’t the Model. It’s the Execution.\", \"url\": \"https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution\", \"date\": \"2026-07\", \"publisher\": \"The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors\", \"locator\": \"July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies\", \"excerpt\": \"We activated the system across 12 Permian wells beginning in December 2025.\", \"access\": \"Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.\", \"type\": \"Operator/vendor-coauthored field account\", \"review_status\": \"Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.\"}], \"field_sources\": {\"business_problem\": [\"WATCH20PM-USEDC\"], \"workflow\": [\"WATCH20PM-USEDC\"], \"technology_data\": [\"WATCH20PM-USEDC\"], \"integration_human_role\": [\"WATCH20PM-USEDC\"], \"deployment_scope\": [\"WATCH20PM-USEDC\"], \"economics\": [\"WATCH20PM-USEDC\"]}}", "Research_status": "Added20October 2026; missed July 2026 account", "Research_category": "Internal adoption", "Research_verification": "verified partner", "Research_origin": ["research/watch-2026-10-20-pm/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Vendor-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "12-well operating cohort; not whole fleet", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case. The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion. Return definition, full initial/recurring costs, realized cash, adjustment for other operating changes and matched controls are undisclosed. The article says engineering work changed toward exceptions, not that headcount or total workload fell.", "Outcome_counted": true, "Outcome_count_note": "One source-reported numerical-result profile. Reported return is undefined and not verified full-cost net cash ROI; no new cost/cash or paired investment-benefit count.", "Current_source_check": "2026-10-20", "Evidence_review": {"source_read": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.", "claim_status": "Operating cohort; operator/vendor-reported", "independent_validation": "Not established"}, "Detail_words": 76}, {"ID": "AI-297", "Company": "PTTEP", "Company_primary": "PTTEP", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "X.brain enterprise engine for specialist agents", "Extracted_finding": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data.", "Stage_at_disclosure": "Engine launch described; eight-agent rollout unverified", "Reported_metric": "No X.brain measured result disclosed. Eight specialist agents were planned; completed rollout and active users are unverified.", "Interpretation_limit": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.", "Source_IDs": "WATCH21AM-PTTEP", "Source_date": "2025-11-03", "Source_type": "Operator technology feature", "Source_locator": "Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples", "Source_URL": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "Evidence_status": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.", "PE_relationship": null, "Coverage_bucket": "21 October morning original-source review", "Time_scope": "Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.", "Context": "Make company-specific knowledge and data available to specialist assistants across operating and corporate work.", "Implementation": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live. Engine launch described; eight specialist agents planned across five domains. No active-user count, asset population or completed eight-agent rollout established.", "Assessment": "An internal engine is a useful architectural direction. Ask for the agents people use and the decisions they improve before treating a launch as a working company brain.", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Reservoir engineering & field development", "Drilling & completions", "Production & field operations", "Supply chain & procurement"], "Department_assignment_basis": "Analyst mapping from planned specialist domains; no named organizational owner disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["21 October Luna discovery; editor original-source review"], "Reference_IDs": [], "Related_IDs": ["AI-067", "AI-245"], "Disclosures": [{"ID": "WATCH21AM-AI-297", "Initiative_ID": "AI-297", "Title": "PTTEP describes an in-house engine for specialist agents", "Date": "2025-11-03", "Finding": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live.", "Metric": "No X.brain user count, adoption rate or measured return is disclosed. The same article separately reports LogSight interpretation falling from two days to one hour and DORA drilling savings described only as “millions of US dollars annually”. Those are product-specific claims, not demonstrated X.brain results.", "Source_IDs": "WATCH21AM-PTTEP", "Source_URL": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "Evidence": "Accessible original reviewed; issuer/vendor claims not independently validated", "Origin": "21 October morning Luna research; editor source and attribution review", "Assessment": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons."}], "Latest_disclosure": "2025-11-03", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch21am-ai-297"], "Research_search": "{\"packet_id\": \"watch21am-ai-297\", \"key\": \"watch21am-ai-297\", \"record_ids\": [\"AI-297\"], \"company\": \"PTTEP\", \"title\": \"PTTEP describes an in-house engine for specialist agents\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-21\", \"verification\": \"verified primary\", \"stage\": \"Engine launch described; eight-agent rollout unverified\", \"business_problem\": \"Make company-specific knowledge and data available to specialist assistants across operating and corporate work.\", \"workflow\": \"PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live.\", \"technology_data\": \"In-house Enterprise AI Engine grounded in PTTEP knowledge and internal data. Model provider, retrieval stack, permissions and runtime architecture not disclosed.\", \"integration_human_role\": \"Analysis and recommendations are described. Approval, override and execution rights are undisclosed; no autonomous executive is established.\", \"deployment_scope\": \"Engine launch described; eight specialist agents planned across five domains. No active-user count, asset population or completed eight-agent rollout established.\", \"economics\": \"No X.brain benefit or full-cost return disclosed. DORA and LogSight appear separately in the same feature; their metrics are not credited to the engine.\", \"practical_relevance\": \"Ask which agents people actually use, what data they can reach and which decisions still need a named human approver.\", \"assessment\": \"An internal engine is a useful architectural direction. Ask for the agents people use and the decisions they improve before treating a launch as a working company brain.\", \"limitations\": [\"No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.\"], \"source_ids\": [\"WATCH21AM-PTTEP\"], \"partners\": [], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2025-11-03\", \"event\": \"Operator feature describes launched engine and planned eight-agent scope; exact launch date unknown.\", \"source_ids\": [\"WATCH21AM-PTTEP\"]}], \"sources\": [{\"id\": \"WATCH21AM-PTTEP\", \"title\": \"PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy\", \"url\": \"https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy\", \"date\": \"2025-11-03\", \"publisher\": \"PTTEP\", \"locator\": \"Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples\", \"excerpt\": \"PTTEP has launched its in-house Enterprise AI Engine, “X.brain,”\", \"access\": \"Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.\", \"type\": \"Operator technology feature\", \"review_status\": \"Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.\"}], \"field_sources\": {\"business_problem\": [\"WATCH21AM-PTTEP\"], \"workflow\": [\"WATCH21AM-PTTEP\"], \"technology_data\": [\"WATCH21AM-PTTEP\"], \"integration_human_role\": [\"WATCH21AM-PTTEP\"], \"deployment_scope\": [\"WATCH21AM-PTTEP\"], \"economics\": [\"WATCH21AM-PTTEP\"]}}", "Research_status": "Added 21 October 2026; missed November 2025 feature", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/watch-2026-10-21-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Company-reported", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "Launched engine described; eight-agent scope is a plan, not verified adoption", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.", "Outcome_counted": false, "Outcome_count_note": "Launch and planned agents do not establish measured deployment outcomes. Separate DORA/LogSight comparisons are not X.brain results.", "Current_source_check": "2026-10-21", "Evidence_review": {"source_read": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.", "claim_status": "Engine launch described; operating adoption unverified", "independent_validation": "Not established"}, "Detail_words": 51}, {"ID": "AI-298", "Company": "Vista Energy", "Company_primary": "Vista Energy", "Segment": "Upstream", "Theme": "Operational AI", "PE_relationship": null, "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH22PM-AI-298", "Initiative_ID": "AI-298", "Title": "Vista uses AI monitoring to adjust completion designs", "Date": "2025-11-12", "Finding": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025.", "Metric": "Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.", "Source_IDs": "WATCH22PM-VISTA; WATCH22PM-VISTA-SEC", "Source_URL": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "Evidence": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted."}], "Research_packets": ["watch22pm-ai-298"], "Source_IDs": "WATCH22PM-VISTA; WATCH22PM-VISTA-SEC", "Provenance": ["22 October evening Luna discovery; editor source review", "22 October evening original-source review; previous evidence retained as dated history"], "Research_origin": ["research/watch-2026-10-22-pm/README.md"], "Initiative": "Proprietary AI monitoring for real-time completion design", "Extracted_finding": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta.", "Stage_at_disclosure": "Operating since March 2025, per issuer", "Reported_metric": "Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.", "Interpretation_limit": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "Source_date": "2025-11-12", "Source_type": "Company investor presentation", "Source_locator": "Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.", "Source_URL": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "Evidence_status": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "Coverage_bucket": "22 October evening targeted review", "Time_scope": "Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "Context": "Recognize frac hits while a completion is still underway and adjust the design before more stages are lost.", "Implementation": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025. Vaca Muerta, Argentina; specific block, pad and well count undisclosed.", "Assessment": "What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?", "Primary_department": "Drilling & completions", "Secondary_departments": ["Reservoir engineering & field development"], "Department_assignment_basis": "Analyst workflow mapping; company organizational ownership not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Latest_disclosure": "2025-11-12", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_search": "{\"packet_id\": \"watch22pm-ai-298\", \"key\": \"watch22pm-ai-298\", \"record_ids\": [\"AI-298\"], \"company\": \"Vista Energy\", \"title\": \"Vista uses AI monitoring to adjust completion designs\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"report_date\": \"2026-10-22\", \"verification\": \"verified primary\", \"stage\": \"Operating since March 2025, per issuer\", \"business_problem\": \"Recognize frac hits while a completion is still underway and adjust the design before more stages are lost.\", \"workflow\": \"Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025.\", \"technology_data\": \"Proprietary AI monitoring for real-time completion-design decisions; no LLM or autonomous pumping authority established.\", \"integration_human_role\": \"Engineering decision support is described. The slide does not define human approval, override or execution rights.\", \"deployment_scope\": \"Vaca Muerta, Argentina; specific block, pad and well count undisclosed. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.\", \"economics\": \"Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.\", \"practical_relevance\": \"What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?\", \"assessment\": \"What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?\", \"limitations\": [\"No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted.\", \"2025 technical measures; matched baseline and denominator undisclosed.\"], \"source_ids\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"partners\": [], \"disclosed_results\": [{\"metric\": \"Issuer-reported completion-stage measures\", \"value\": \"Less than 1% stages lost in 2025; one runaway stage per well avoided on average\", \"period\": \"2025; exact measurement period unknown\", \"attribution\": \"Issuer; not independently validated or causally isolated\", \"source_ids\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"]}], \"timeline\": [{\"date\": \"2025-11-12\", \"event\": \"Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta.\", \"source_ids\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"]}], \"sources\": [{\"id\": \"WATCH22PM-VISTA\", \"title\": \"Vista Investor Day 2025\", \"url\": \"https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf\", \"date\": \"2025-11-12\", \"publisher\": \"Vista Energy\", \"locator\": \"Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.\", \"excerpt\": \"Implemented since March 2025\", \"access\": \"Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.\", \"type\": \"Company investor presentation\", \"review_status\": \"Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.\"}, {\"id\": \"WATCH22PM-VISTA-SEC\", \"title\": \"Vista Investor Day 2025 — SEC exhibit\", \"url\": \"https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm\", \"date\": \"2025-11-12\", \"publisher\": \"Vista Energy / SEC EDGAR\", \"locator\": \"Exhibit heading: November 12, 2025; same slide 19 presentation.\", \"excerpt\": \"Vista Investor Day\", \"access\": \"Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes.\", \"type\": \"Company investor presentation\", \"review_status\": \"Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes.\"}], \"field_sources\": {\"business_problem\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"workflow\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"technology_data\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"integration_human_role\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"deployment_scope\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"], \"economics\": [\"WATCH22PM-VISTA\", \"WATCH22PM-VISTA-SEC\"]}}", "Research_status": "Added 22 October 2026; missed historical disclosure", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Company investor disclosure", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "Operating since March 2025, per issuer Vaca Muerta, Argentina; specific block, pad and well count undisclosed.", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted.", "Outcome_counted": true, "Outcome_count_note": "Source-reported technical-stage measures only; unresolved dollar claim excluded; no net cash ROI.", "Current_source_check": "2026-10-22", "Evidence_review": {"source_read": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "claim_status": "Operating since March 2025, per issuer", "independent_validation": "Not established"}, "Detail_words": 42}, {"ID": "AI-299", "Company": "Saipem", "Company_primary": "Saipem", "Segment": "Services / EPC", "Theme": "Operational AI", "PE_relationship": null, "Reference_IDs": [], "Related_IDs": ["AI-160", "AI-241", "AI-242"], "Disclosures": [{"ID": "WATCH22PM-AI-299", "Initiative_ID": "AI-299", "Title": "Saipem installs a predictive-maintenance pilot on one drillship", "Date": "2026-02-23", "Finding": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests.", "Metric": "No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.", "Source_IDs": "WATCH22PM-SAIPEM", "Source_URL": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "Evidence": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed."}], "Research_packets": ["watch22pm-ai-299"], "Source_IDs": "WATCH22PM-SAIPEM", "Provenance": ["22 October evening Luna discovery; editor source review", "22 October evening original-source review; previous evidence retained as dated history"], "Research_origin": ["research/watch-2026-10-22-pm/README.md"], "Initiative": "Predictive-maintenance pilot on Saipem 12000", "Extracted_finding": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it.", "Stage_at_disclosure": "Installed pilot on Saipem 12000; fleet extension planned", "Reported_metric": "No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.", "Interpretation_limit": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed. Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.", "Source_date": "2026-02-23", "Source_type": "Company announcement", "Source_locator": "Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.", "Source_URL": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "Evidence_status": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "Coverage_bucket": "22 October evening targeted review", "Time_scope": "Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.", "Context": "Find emerging equipment problems early enough to plan maintenance on a drillship.", "Implementation": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests. Saipem 12000 ultra-deepwater drillship. Saipem 7000 is separate future context.", "Assessment": "Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?", "Primary_department": "Maintenance & reliability", "Secondary_departments": ["Production & field operations"], "Department_assignment_basis": "Analyst workflow mapping; company organizational ownership not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Latest_disclosure": "2026-02-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_search": "{\"packet_id\": \"watch22pm-ai-299\", \"key\": \"watch22pm-ai-299\", \"record_ids\": [\"AI-299\"], \"company\": \"Saipem\", \"title\": \"Saipem installs a predictive-maintenance pilot on one drillship\", \"category\": \"Internal adoption\", \"department\": \"Maintenance & reliability\", \"report_date\": \"2026-10-22\", \"verification\": \"verified primary\", \"stage\": \"Installed pilot on Saipem 12000; fleet extension planned\", \"business_problem\": \"Find emerging equipment problems early enough to plan maintenance on a drillship.\", \"workflow\": \"Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests.\", \"technology_data\": \"Equipment-data anomaly detection and failure prediction; no LLM or autonomous maintenance execution established.\", \"integration_human_role\": \"Maintenance planning is described. Approval, execution, override and fallback boundaries are undisclosed.\", \"deployment_scope\": \"Saipem 12000 ultra-deepwater drillship. Saipem 7000 is separate future context. Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.\", \"economics\": \"No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.\", \"practical_relevance\": \"Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?\", \"assessment\": \"Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?\", \"limitations\": [\"The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed.\", \"No pre-pilot reliability comparator or evaluation period disclosed.\"], \"source_ids\": [\"WATCH22PM-SAIPEM\"], \"partners\": [\"ADC Energy\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-02-23\", \"event\": \"Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it.\", \"source_ids\": [\"WATCH22PM-SAIPEM\"]}], \"sources\": [{\"id\": \"WATCH22PM-SAIPEM\", \"title\": \"Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000\", \"url\": \"https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000\", \"date\": \"2026-02-23\", \"publisher\": \"Saipem\", \"locator\": \"Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.\", \"excerpt\": \"This pilot project\", \"access\": \"Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.\", \"type\": \"Company announcement\", \"review_status\": \"Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.\"}], \"field_sources\": {\"business_problem\": [\"WATCH22PM-SAIPEM\"], \"workflow\": [\"WATCH22PM-SAIPEM\"], \"technology_data\": [\"WATCH22PM-SAIPEM\"], \"integration_human_role\": [\"WATCH22PM-SAIPEM\"], \"deployment_scope\": [\"WATCH22PM-SAIPEM\"], \"economics\": [\"WATCH22PM-SAIPEM\"]}}", "Research_status": "Added 22 October 2026; missed historical disclosure", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Company announcement", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "Installed pilot on Saipem 12000; fleet extension planned Saipem 12000 ultra-deepwater drillship. Saipem 7000 is separate future context.", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed.", "Outcome_counted": false, "Outcome_count_note": "Pilot installation, tool names and program outputs do not establish a measured operating or cash result.", "Current_source_check": "2026-10-22", "Evidence_review": {"source_read": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "claim_status": "Installed pilot on Saipem 12000; fleet extension planned", "independent_validation": "Not established"}, "Detail_words": 52}, {"ID": "AI-300", "Company": "Worley", "Company_primary": "Worley", "Segment": "Services / EPC", "Theme": "Enterprise AI", "PE_relationship": null, "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH22PM-AI-300", "Initiative_ID": "AI-300", "Title": "Worley puts four internal tools into engineering and bid work", "Date": "2025-08-27", "Finding": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest.", "Metric": "Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.", "Source_IDs": "WATCH22PM-WORLEY", "Source_URL": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "Evidence": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established."}], "Research_packets": ["watch22pm-ai-300"], "Source_IDs": "WATCH22PM-WORLEY", "Provenance": ["22 October evening Luna discovery; editor source review", "22 October evening original-source review; previous evidence retained as dated history"], "Research_origin": ["research/watch-2026-10-22-pm/README.md"], "Initiative": "Four internal AI tools for engineering and commercial work", "Extracted_finding": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting.", "Stage_at_disclosure": "Four recently deployed internal solutions, per issuer", "Reported_metric": "Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.", "Interpretation_limit": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established. FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.", "Source_date": "2025-08-27", "Source_type": "Company investor presentation", "Source_locator": "Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.", "Source_URL": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "Evidence_status": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "Coverage_bucket": "22 October evening targeted review", "Time_scope": "FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.", "Context": "Reduce repeated work finding engineering knowledge, checking documents and preparing technical or commercial responses.", "Implementation": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest. Internal engineering and commercial work; no client asset, basin or user population disclosed.", "Assessment": "Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": ["Supply chain & logistics"], "Department_assignment_basis": "Analyst workflow mapping; company organizational ownership not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Latest_disclosure": "2025-08-27", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_search": "{\"packet_id\": \"watch22pm-ai-300\", \"key\": \"watch22pm-ai-300\", \"record_ids\": [\"AI-300\"], \"company\": \"Worley\", \"title\": \"Worley puts four internal tools into engineering and bid work\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-22\", \"verification\": \"verified primary\", \"stage\": \"Four recently deployed internal solutions, per issuer\", \"business_problem\": \"Reduce repeated work finding engineering knowledge, checking documents and preparing technical or commercial responses.\", \"workflow\": \"Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest.\", \"technology_data\": \"Internal AI-assisted document review, search, comparison and drafting. Exact model architecture and execution permissions undisclosed.\", \"integration_human_role\": \"Tools support review and drafting. Named approval roles and execution rights are undisclosed.\", \"deployment_scope\": \"Internal engineering and commercial work; no client asset, basin or user population disclosed. FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.\", \"economics\": \"Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.\", \"practical_relevance\": \"Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?\", \"assessment\": \"Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?\", \"limitations\": [\"These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established.\", \"No tool-level before/after or quality-controlled comparator disclosed.\"], \"source_ids\": [\"WATCH22PM-WORLEY\"], \"partners\": [], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2025-08-27\", \"event\": \"Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting.\", \"source_ids\": [\"WATCH22PM-WORLEY\"]}], \"sources\": [{\"id\": \"WATCH22PM-WORLEY\", \"title\": \"FY2025 full-year results presentation\", \"url\": \"https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf\", \"date\": \"2025-08-27\", \"publisher\": \"Worley\", \"locator\": \"Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.\", \"excerpt\": \"Recently deployed solutions through our Advanced Development Lab\", \"access\": \"Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.\", \"type\": \"Company investor presentation\", \"review_status\": \"Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.\"}], \"field_sources\": {\"business_problem\": [\"WATCH22PM-WORLEY\"], \"workflow\": [\"WATCH22PM-WORLEY\"], \"technology_data\": [\"WATCH22PM-WORLEY\"], \"integration_human_role\": [\"WATCH22PM-WORLEY\"], \"deployment_scope\": [\"WATCH22PM-WORLEY\"], \"economics\": [\"WATCH22PM-WORLEY\"]}}", "Research_status": "Added 22 October 2026; missed historical disclosure", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Company investor disclosure", "Attribution": "operator", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "Four recently deployed internal solutions, per issuer Internal engineering and commercial work; no client asset, basin or user population disclosed.", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established.", "Outcome_counted": false, "Outcome_count_note": "Pilot installation, tool names and program outputs do not establish a measured operating or cash result.", "Current_source_check": "2026-10-22", "Evidence_review": {"source_read": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "claim_status": "Four recently deployed internal solutions, per issuer", "independent_validation": "Not established"}, "Detail_words": 41}, {"ID": "AI-301", "Company": "Azule Energy", "Company_primary": "Azule Energy", "Segment": "Upstream", "Theme": "Operational AI", "Initiative": "Digital well planning linked to AI-enabled rig automation", "Extracted_finding": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola.", "Stage_at_disclosure": "Operating single-well example; wider rollout unverified", "Reported_metric": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "Interpretation_limit": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate. Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown.", "Source_IDs": "WATCH23AM-SLB-Q3; WATCH23AM-AZULE", "Source_date": "2026-10-23", "Source_type": "Supplier earnings release / customer account", "Source_locator": "PDF p. 16: Digital / Azule Energy paragraph; same case and key measures.", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence_status": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "PE_relationship": null, "Coverage_bucket": "23 October fresh supplier earnings disclosure", "Time_scope": "Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown.", "Context": "Carry an engineering plan into rig execution with fewer manual transfers.", "Implementation": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola. DrillPlan passes the plan into DrillOps automation. The supplier case identifies AI-driven operations; it does not isolate AI from engineering automation.", "Assessment": "Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?", "Primary_department": "Drilling & completions", "Secondary_departments": [], "Department_assignment_basis": "Analyst workflow mapping; organizational owner not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["23 October fresh original supplier release and case-study review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH23AM-AI-301", "Initiative_ID": "AI-301", "Title": "Azule connects the well plan to drilling automation", "Date": "2026-10-23", "Finding": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola.", "Metric": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "Source_IDs": "WATCH23AM-SLB-Q3; WATCH23AM-AZULE", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate."}], "Latest_disclosure": "2026-10-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch23am-ai-301"], "Research_search": "{\"packet_id\": \"watch23am-ai-301\", \"key\": \"watch23am-ai-301\", \"record_ids\": [\"AI-301\"], \"company\": \"Azule Energy\", \"title\": \"Azule connects the well plan to drilling automation\", \"category\": \"Internal adoption\", \"department\": \"Drilling & completions\", \"report_date\": \"2026-10-23\", \"verification\": \"verified vendor\", \"stage\": \"Operating single-well example; wider rollout unverified\", \"business_problem\": \"Carry an engineering plan into rig execution with fewer manual transfers.\", \"workflow\": \"SLB describes Azule using one digital workflow from well design through rig execution offshore Angola. DrillPlan passes the plan into DrillOps automation. The supplier case identifies AI-driven operations; it does not isolate AI from engineering automation.\", \"technology_data\": \"AI-enabled DrillOps execution alongside digital well planning; underlying algorithms and model type undisclosed.\", \"integration_human_role\": \"Autonomous section execution is reported. Approval, override, stop and fallback boundaries are not disclosed.\", \"deployment_scope\": \"One offshore Angola well; no named field or basin disclosed. Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown.\", \"economics\": \"Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.\", \"practical_relevance\": \"Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?\", \"assessment\": \"Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?\", \"limitations\": [\"No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate.\", \"Planning before/after and an unspecified field ROP benchmark; no matched controlled comparison.\"], \"source_ids\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"partners\": [\"SLB\"], \"disclosed_results\": [{\"metric\": \"Supplier-reported planning and section performance\", \"value\": \"Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.\", \"period\": \"Single-well example; exact dates unknown\", \"attribution\": \"SLB supplier account; not independently validated or AI-only\", \"source_ids\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"]}], \"timeline\": [{\"date\": \"2026-10-23\", \"event\": \"Q3 release describes the named example or contract; exact operating/award dates unknown.\", \"source_ids\": [\"WATCH23AM-SLB-Q3\"]}], \"sources\": [{\"id\": \"WATCH23AM-SLB-Q3\", \"title\": \"SLB third-quarter 2026 results — selected operating and digital highlights\", \"url\": \"https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f\", \"date\": \"2026-10-23\", \"publisher\": \"SLB\", \"locator\": \"PDF p. 16: Digital / Azule Energy paragraph; same case and key measures.\", \"excerpt\": \"cut planning time from 192 to 31 hours\", \"access\": \"Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.\", \"format\": \"pdf\", \"type\": \"Supplier earnings release / customer account\", \"review_status\": \"Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.\"}, {\"id\": \"WATCH23AM-AZULE\", \"title\": \"Azule Energy pioneers digital well planning and rig automation in Africa\", \"url\": \"https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa\", \"date\": \"Unknown\", \"publisher\": \"SLB\", \"locator\": \"Location; Results: Achieving operational excellence; Products Used: DrillOps AI-driven workflows.\", \"excerpt\": \"digitally ingested the well plan\", \"access\": \"Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration.\", \"type\": \"Supplier earnings release / customer account\", \"review_status\": \"Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration.\"}], \"field_sources\": {\"business_problem\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"workflow\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"technology_data\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"integration_human_role\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"deployment_scope\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"], \"economics\": [\"WATCH23AM-SLB-Q3\", \"WATCH23AM-AZULE\"]}}", "Research_status": "Added 23 October 2026; supplier customer disclosure", "Research_category": "Internal adoption", "Research_verification": "verified vendor", "Research_origin": ["research/watch-2026-10-23-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Supplier-reported", "Attribution": "vendor", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "Operating single-well example; wider rollout unverified One offshore Angola well; no named field or basin disclosed.", "Segment_norm": "Independent E&P", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate.", "Outcome_counted": true, "Outcome_count_note": "Supplier-reported technical measures for a single-well example; no full-cost cash ROI.", "Current_source_check": "2026-10-23", "Evidence_review": {"source_read": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "claim_status": "Operating single-well example; wider rollout unverified", "independent_validation": "Not established"}, "Detail_words": 36}, {"ID": "AI-302", "Company": "Brunei Shell Petroleum", "Company_primary": "Brunei Shell Petroleum", "Segment": "Upstream", "Theme": "Enterprise AI", "Initiative": "On-premise subsurface and production software contract with AI capabilities", "Extracted_finding": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work.", "Stage_at_disclosure": "Contract awarded; go-live and usage unverified", "Reported_metric": "No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.", "Interpretation_limit": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout. Published 23 October 2026. Award date, implementation timetable and active-use dates unknown.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_date": "2026-10-23", "Source_type": "Supplier earnings release / customer account", "Source_locator": "PDF p. 16: Digital / Brunei BSP software contract paragraph.", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence_status": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "PE_relationship": null, "Coverage_bucket": "23 October fresh supplier earnings disclosure", "Time_scope": "Published 23 October 2026. Award date, implementation timetable and active-use dates unknown.", "Context": "Connect subsurface and production software with a consistent implementation and support arrangement.", "Implementation": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work. Analyst takeaway: a broad contract establishes buying intent and implementation scope; it does not show which AI tasks employees complete.", "Assessment": "Which workflows will go live first, and what acceptance test separates installed software from useful work?", "Primary_department": "IT, data & enterprise transformation", "Secondary_departments": [], "Department_assignment_basis": "Analyst workflow mapping; organizational owner not disclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["23 October fresh original supplier release and case-study review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH23AM-AI-302", "Initiative_ID": "AI-302", "Title": "BSP awards an on-premise software contract", "Date": "2026-10-23", "Finding": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work.", "Metric": "No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout."}], "Latest_disclosure": "2026-10-23", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch23am-ai-302"], "Research_search": "{\"packet_id\": \"watch23am-ai-302\", \"key\": \"watch23am-ai-302\", \"record_ids\": [\"AI-302\"], \"company\": \"Brunei Shell Petroleum\", \"title\": \"BSP awards an on-premise software contract\", \"category\": \"Internal adoption\", \"department\": \"IT, data & enterprise transformation\", \"report_date\": \"2026-10-23\", \"verification\": \"verified vendor\", \"stage\": \"Contract awarded; go-live and usage unverified\", \"business_problem\": \"Connect subsurface and production software with a consistent implementation and support arrangement.\", \"workflow\": \"BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work. Analyst takeaway: a broad contract establishes buying intent and implementation scope; it does not show which AI tasks employees complete.\", \"technology_data\": \"Contract includes AI capabilities; no particular live model or LLM workflow established.\", \"integration_human_role\": \"No project-level AI approval or execution boundaries disclosed.\", \"deployment_scope\": \"BSP, Brunei; subsurface and production scope, no specific asset or user count. Published 23 October 2026. Award date, implementation timetable and active-use dates unknown.\", \"economics\": \"No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.\", \"practical_relevance\": \"Which workflows will go live first, and what acceptance test separates installed software from useful work?\", \"assessment\": \"Which workflows will go live first, and what acceptance test separates installed software from useful work?\", \"limitations\": [\"Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout.\", \"No live-use or outcome baseline established.\"], \"source_ids\": [\"WATCH23AM-SLB-Q3\"], \"partners\": [\"SLB\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-10-23\", \"event\": \"Q3 release describes the named example or contract; exact operating/award dates unknown.\", \"source_ids\": [\"WATCH23AM-SLB-Q3\"]}], \"sources\": [{\"id\": \"WATCH23AM-SLB-Q3\", \"title\": \"SLB third-quarter 2026 results — selected operating and digital highlights\", \"url\": \"https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f\", \"date\": \"2026-10-23\", \"publisher\": \"SLB\", \"locator\": \"PDF p. 16: Digital / Brunei BSP software contract paragraph.\", \"excerpt\": \"on-premises software contract\", \"access\": \"Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.\", \"format\": \"pdf\", \"type\": \"Supplier earnings release / customer account\", \"review_status\": \"Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.\"}], \"field_sources\": {\"business_problem\": [\"WATCH23AM-SLB-Q3\"], \"workflow\": [\"WATCH23AM-SLB-Q3\"], \"technology_data\": [\"WATCH23AM-SLB-Q3\"], \"integration_human_role\": [\"WATCH23AM-SLB-Q3\"], \"deployment_scope\": [\"WATCH23AM-SLB-Q3\"], \"economics\": [\"WATCH23AM-SLB-Q3\"]}}", "Research_status": "Added 23 October 2026; supplier customer disclosure", "Research_category": "Internal adoption", "Research_verification": "verified vendor", "Research_origin": ["research/watch-2026-10-23-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Supplier-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "announced", "Scale_note": "Contract awarded; go-live and usage unverified BSP, Brunei; subsurface and production scope, no specific asset or user count.", "Segment_norm": "Integrated & supermajor", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout.", "Outcome_counted": false, "Outcome_count_note": "Contract award only; no operating or quantified-result credit.", "Current_source_check": "2026-10-23", "Evidence_review": {"source_read": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "claim_status": "Contract awarded; go-live and usage unverified", "independent_validation": "Not established"}, "Detail_words": 40, "Segment_assignment_basis": "Shell-affiliated upstream operator, shown separately from Shell in company counts. Analyst sector mapping, not a Shell-wide deployment claim."}, {"ID": "AI-303", "Company": "Petrobras", "Company_primary": "Petrobras", "Segment": "Integrated / NOC", "Theme": "Enterprise AI", "Initiative": "AVAL IA reviews supplier documents, Wave 2", "Extracted_finding": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them.", "Stage_at_disclosure": "Wave 2 pilot; rollout unconfirmed", "Reported_metric": "No accuracy, workload, completed-case count, cost, saving or ROI reported.", "Interpretation_limit": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established. Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.", "Source_IDs": "WATCH23PM-PETROBRAS", "Source_date": "Unknown", "Source_type": "Company supplier-channel pilot notice", "Source_locator": "Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.", "Source_URL": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "Evidence_status": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "PE_relationship": null, "Coverage_bucket": "23 October evening targeted review", "Time_scope": "Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.", "Context": "Check supplier evidence during registration.", "Implementation": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them. The evaluation team records its decision and feeds it back to the AI. Petronect handles registration; the notice does not identify the AI model or developer.", "Assessment": "Does the first AI pass reduce total review work while preserving the quality of supplier decisions?", "Primary_department": "Supply chain & logistics", "Secondary_departments": [], "Department_assignment_basis": "Supplier Performance and Qualification management is named in the source; analyst department mapping.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["23 October evening original-source pilot review"], "Reference_IDs": [], "Related_IDs": [], "Disclosures": [{"ID": "WATCH23PM-AI-303", "Initiative_ID": "AI-303", "Title": "AI checks supplier documents. Petrobras staff make the decision.", "Date": "Unknown", "Finding": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them.", "Metric": "No accuracy, workload, completed-case count, cost, saving or ROI reported.", "Source_IDs": "WATCH23PM-PETROBRAS", "Source_URL": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "Evidence": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "Origin": "23 October evening bounded Luna research; editor original-source review", "Assessment": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established."}], "Latest_disclosure": "Unknown", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch23pm-ai-303"], "Research_search": "{\"packet_id\": \"watch23pm-ai-303\", \"key\": \"watch23pm-ai-303\", \"record_ids\": [\"AI-303\"], \"company\": \"Petrobras\", \"title\": \"AI checks supplier documents. Petrobras staff make the decision.\", \"category\": \"Internal adoption\", \"department\": \"Supply chain & logistics\", \"report_date\": \"2026-10-23\", \"verification\": \"verified primary\", \"stage\": \"Wave 2 pilot; rollout unconfirmed\", \"business_problem\": \"Check supplier evidence during registration.\", \"workflow\": \"Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them. The evaluation team records its decision and feeds it back to the AI. Petronect handles registration; the notice does not identify the AI model or developer.\", \"technology_data\": \"Preliminary document analysis; model type and developer undisclosed.\", \"integration_human_role\": \"Technical evaluators perform the final check and retain the decision.\", \"deployment_scope\": \"Petrobras corporate supplier registration; no producing asset or basin applies. Participating supplier volume is undisclosed. Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.\", \"economics\": \"No accuracy, workload, completed-case count, cost, saving or ROI reported.\", \"practical_relevance\": \"Does the first AI pass reduce total review work while preserving the quality of supplier decisions?\", \"assessment\": \"Does the first AI pass reduce total review work while preserving the quality of supplier decisions?\", \"limitations\": [\"Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.\", \"No before/after comparison or measurement period disclosed.\"], \"source_ids\": [\"WATCH23PM-PETROBRAS\"], \"partners\": [\"Petronect (registration portal, not established AI developer)\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"Unknown\", \"event\": \"Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.\", \"source_ids\": [\"WATCH23PM-PETROBRAS\"]}], \"sources\": [{\"id\": \"WATCH23PM-PETROBRAS\", \"title\": \"Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice\", \"url\": \"https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores\", \"date\": \"Unknown\", \"publisher\": \"Petrobras\", \"locator\": \"Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.\", \"excerpt\": \"A supervisão e a decisão final permanecem sempre com nossos avaliadores técnicos.\", \"access\": \"Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.\", \"type\": \"Company supplier-channel pilot notice\", \"review_status\": \"Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.\"}], \"field_sources\": {\"business_problem\": [\"WATCH23PM-PETROBRAS\"], \"workflow\": [\"WATCH23PM-PETROBRAS\"], \"technology_data\": [\"WATCH23PM-PETROBRAS\"], \"integration_human_role\": [\"WATCH23PM-PETROBRAS\"], \"deployment_scope\": [\"WATCH23PM-PETROBRAS\"], \"economics\": [\"WATCH23PM-PETROBRAS\"]}}", "Research_status": "Added 23 October 2026; undated source, schedule from September 2026", "Research_category": "Internal adoption", "Research_verification": "verified primary", "Research_origin": ["research/watch-2026-10-23-pm/README.md"], "Record_type": "company_adoption", "Evidence_grade": "B", "Evidence_label": "B · Company disclosure", "Attribution": "operator", "Outcome_type": "none", "Stage_norm": "pilot", "Scale_note": "Wave 2 supplier-document pilot; evaluation volume unreported.", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.", "Outcome_counted": false, "Outcome_count_note": "Pilot workflow and review boundary only; no measured outcome, failure or return established.", "Current_source_check": "2026-10-23", "Evidence_review": {"source_read": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "claim_status": "Wave 2 pilot; rollout unconfirmed", "independent_validation": "Not established"}, "Detail_words": 41}, {"ID": "AI-304", "Company": "Halliburton", "Company_primary": "Halliburton", "Segment": "Oilfield services", "Theme": "Operational AI", "Initiative": "LOGIX Unit Vitality monitors cementing equipment", "Extracted_finding": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs.", "Stage_at_disclosure": "Supplier reports four deployments; customers unidentified", "Reported_metric": "The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.", "Interpretation_limit": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016. Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.", "Source_IDs": "WATCH24AM-HAL-RELIABILITY; WATCH24AM-HAL-PRODUCT", "Source_date": "2026-10-22", "Source_type": "Supplier product account", "Source_locator": "A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.", "Source_URL": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "Evidence_status": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "PE_relationship": null, "Coverage_bucket": "24 October morning targeted review", "Time_scope": "Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.", "Context": "Find developing equipment problems early enough for a useful maintenance intervention.", "Implementation": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs. Sensors feed more than 400 equipment parameters into cloud models. Staff check warnings, arrange parts and schedule corrective work. The product page separately describes transferring approved job designs from VIDA and iCem into controllers; this is not proof of autonomous maintenance.", "Assessment": "Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?", "Primary_department": "Maintenance & reliability", "Secondary_departments": [], "Department_assignment_basis": "Analyst mapping of equipment health alerts and maintenance workflow; team roles described by supplier.", "Scope_type": "Commercial AI product", "Review_tier": "Reviewed source", "Provenance": ["24 October morning original-source product review"], "Reference_IDs": [], "Related_IDs": ["AI-016"], "Disclosures": [{"ID": "WATCH24AM-AI-304", "Initiative_ID": "AI-304", "Title": "A maintenance alert only helps if a crew can act on it.", "Date": "2026-10-22", "Finding": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs.", "Metric": "The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.", "Source_IDs": "WATCH24AM-HAL-RELIABILITY; WATCH24AM-HAL-PRODUCT", "Source_URL": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "Evidence": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "Origin": "24 October morning editor original-source check alongside three bounded Luna tracks", "Assessment": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016."}], "Latest_disclosure": "2026-10-22", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch24am-ai-304"], "Research_search": "{\"packet_id\": \"watch24am-ai-304\", \"key\": \"watch24am-ai-304\", \"record_ids\": [\"AI-304\"], \"company\": \"Halliburton\", \"title\": \"A maintenance alert only helps if a crew can act on it.\", \"category\": \"Commercial AI product\", \"department\": \"Maintenance & reliability\", \"report_date\": \"2026-10-24\", \"verification\": \"verified vendor\", \"stage\": \"Supplier reports four deployments; customers unidentified\", \"business_problem\": \"Find developing equipment problems early enough for a useful maintenance intervention.\", \"workflow\": \"Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs. Sensors feed more than 400 equipment parameters into cloud models. Staff check warnings, arrange parts and schedule corrective work. The product page separately describes transferring approved job designs from VIDA and iCem into controllers; this is not proof of autonomous maintenance.\", \"technology_data\": \"Statistical models and machine learning for equipment condition; no LLM identified.\", \"integration_human_role\": \"Crews investigate alerts, obtain parts and schedule repairs; authorization, override and fallback rules are not specified.\", \"deployment_scope\": \"Cementing equipment; four deployments reported, with no identified basin, asset or customer. Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.\", \"economics\": \"The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.\", \"practical_relevance\": \"Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?\", \"assessment\": \"Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?\", \"limitations\": [\"No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.\", \"No comparator, productivity definition or cost model disclosed.\"], \"source_ids\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"partners\": [\"Halliburton\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-10-22\", \"event\": \"Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.\", \"source_ids\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"]}], \"sources\": [{\"id\": \"WATCH24AM-HAL-RELIABILITY\", \"title\": \"How predictive intelligence reduces downtime and improves equipment reliability\", \"url\": \"https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability\", \"date\": \"2026-10-22\", \"publisher\": \"Halliburton\", \"locator\": \"A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.\", \"excerpt\": \"helped teams avoid an estimated $125,000 to $150,000 in downtime costs\", \"access\": \"Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.\", \"type\": \"Supplier product account\", \"review_status\": \"Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.\"}, {\"id\": \"WATCH24AM-HAL-PRODUCT\", \"title\": \"LOGIX Unit Vitality — cement equipment health monitoring\", \"url\": \"https://www.halliburton.com/en/products/logix-unit-vitality\", \"date\": \"Unknown\", \"publisher\": \"Halliburton\", \"locator\": \"Automated cement equipment health assessment; AI-driven models; Real-time monitoring closes the loop.\", \"excerpt\": \"more than 400 real-time parameters\", \"access\": \"Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments.\", \"type\": \"Supplier product account\", \"review_status\": \"Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments.\"}], \"field_sources\": {\"business_problem\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"workflow\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"technology_data\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"integration_human_role\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"deployment_scope\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"], \"economics\": [\"WATCH24AM-HAL-RELIABILITY\", \"WATCH24AM-HAL-PRODUCT\"]}}", "Research_status": "Added 24 October 2026; dated supplier account and undated product capabilities", "Research_category": "Commercial AI product", "Research_verification": "verified vendor", "Research_origin": ["research/watch-2026-10-24-am/README.md"], "Record_type": "vendor_product", "Evidence_grade": "C", "Evidence_label": "C · Supplier-reported", "Attribution": "vendor", "Outcome_type": "scale_only", "Stage_norm": "operating", "Scale_note": "Supplier reports four deployments; operator identities and equipment population unknown.", "Segment_norm": "Oilfield services", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "Outcome_counted": false, "Outcome_count_note": "Commercial product with unidentified deployments and estimated avoided costs; excluded from named operator and result totals.", "Current_source_check": "2026-10-24", "Evidence_review": {"source_read": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "claim_status": "Supplier reports four deployments; customers unidentified", "independent_validation": "Not established"}, "Detail_words": 60}, {"ID": "AI-305", "Company": "Petroleum Development Oman", "Company_primary": "Petroleum Development Oman", "Segment": "Upstream / NOC", "Theme": "Operational AI", "Initiative": "On-premise ESP failure and run-life prediction", "Extracted_finding": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework.", "Stage_at_disclosure": "Operating deployment, supplier-reported; scale unknown", "Reported_metric": "No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.", "Interpretation_limit": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case. Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.", "Source_IDs": "WATCH24PM-WFRD-Q1", "Source_date": "2026-04-21", "Source_type": "Supplier earnings release", "Source_locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "Source_URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "Evidence_status": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "PE_relationship": null, "Coverage_bucket": "24 October evening targeted review", "Time_scope": "Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.", "Context": "Help production teams anticipate electric submersible pump failures and estimate remaining run life.", "Implementation": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework. The on-premise installation provides failure forecasts, run-life estimates and real-time pump-performance insight while meeting the customer’s cybersecurity requirements. The release does not say whether forecasts trigger automatic changes or how people review alerts and approve maintenance.", "Assessment": "Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.", "Primary_department": "Production & field operations", "Secondary_departments": ["Maintenance & reliability"], "Department_assignment_basis": "Analyst mapping from pump performance and maintenance forecasting; organizational owner undisclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["24 October evening original-source named-customer review"], "Reference_IDs": [], "Related_IDs": ["AI-113"], "Disclosures": [{"ID": "WATCH24PM-AI-305", "Initiative_ID": "AI-305", "Title": "PDO has moved ESP prediction from pilot to operating use", "Date": "2026-04-21", "Finding": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework.", "Metric": "No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.", "Source_IDs": "WATCH24PM-WFRD-Q1", "Source_URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "Evidence": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "Origin": "24 October evening editor original-source check alongside three bounded Luna tracks", "Assessment": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case."}], "Latest_disclosure": "2026-04-21", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch24pm-ai-305"], "Research_search": "{\"packet_id\": \"watch24pm-ai-305\", \"key\": \"watch24pm-ai-305\", \"record_ids\": [\"AI-305\"], \"company\": \"Petroleum Development Oman\", \"title\": \"PDO has moved ESP prediction from pilot to operating use\", \"category\": \"Internal adoption\", \"department\": \"Production & field operations\", \"report_date\": \"2026-10-24\", \"verification\": \"verified vendor\", \"stage\": \"Operating deployment, supplier-reported; scale unknown\", \"business_problem\": \"Help production teams anticipate electric submersible pump failures and estimate remaining run life.\", \"workflow\": \"In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework. The on-premise installation provides failure forecasts, run-life estimates and real-time pump-performance insight while meeting the customer’s cybersecurity requirements. The release does not say whether forecasts trigger automatic changes or how people review alerts and approve maintenance.\", \"technology_data\": \"PetroVisor machine-learning framework supports pump failure/run-life prediction. No LLM or autonomous pump control established.\", \"integration_human_role\": \"Human review, override, alert triage and maintenance approval rules are undisclosed; do not infer autonomous control.\", \"deployment_scope\": \"Oman; exact field, basin, pump population and operating period undisclosed. Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.\", \"economics\": \"No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.\", \"practical_relevance\": \"Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.\", \"assessment\": \"Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.\", \"limitations\": [\"A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.\", \"No counterfactual, baseline or result denominator disclosed.\"], \"source_ids\": [\"WATCH24PM-WFRD-Q1\"], \"partners\": [\"Weatherford\", \"PetroVisor\"], \"disclosed_results\": [], \"timeline\": [{\"date\": \"2026-04-21\", \"event\": \"Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.\", \"source_ids\": [\"WATCH24PM-WFRD-Q1\"]}], \"sources\": [{\"id\": \"WATCH24PM-WFRD-Q1\", \"title\": \"Weatherford Announces First Quarter 2026 Results — PDO ESP analytics\", \"url\": \"https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491\", \"date\": \"2026-04-21\", \"publisher\": \"Weatherford\", \"locator\": \"Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.\", \"excerpt\": \"moving the technology from pilot to operational use\", \"access\": \"Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.\", \"type\": \"Supplier earnings release\", \"review_status\": \"Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.\"}], \"field_sources\": {\"business_problem\": [\"WATCH24PM-WFRD-Q1\"], \"workflow\": [\"WATCH24PM-WFRD-Q1\"], \"technology_data\": [\"WATCH24PM-WFRD-Q1\"], \"integration_human_role\": [\"WATCH24PM-WFRD-Q1\"], \"deployment_scope\": [\"WATCH24PM-WFRD-Q1\"], \"economics\": [\"WATCH24PM-WFRD-Q1\"]}}", "Research_status": "Added 24 October 2026; missed April disclosure, supplier-reported operating use", "Research_category": "Internal adoption", "Research_verification": "verified vendor", "Research_origin": ["research/watch-2026-10-24-pm/README.md"], "Record_type": "company_adoption", "Evidence_grade": "C", "Evidence_label": "C · Supplier-reported", "Attribution": "vendor", "Outcome_type": "none", "Stage_norm": "operating", "Scale_note": "Named customer operating use; pump, well and asset counts unknown.", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": true, "Review_note": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.", "Outcome_counted": false, "Outcome_count_note": "Operating stage is supplier-reported. No measured benefit or complete cash return disclosed.", "Current_source_check": "2026-10-24", "Evidence_review": {"source_read": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "claim_status": "Operating deployment, supplier-reported; scale unknown", "independent_validation": "Not established"}, "Detail_words": 73}, {"ID": "AI-306", "Company": "PTTEP", "Company_primary": "PTTEP", "Segment": "Upstream / NOC", "Theme": "Enterprise AI", "Initiative": "Procurement X.Bot purchase-order processing", "Extracted_finding": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing.", "Stage_at_disclosure": "Operating procurement workflow; scale undisclosed", "Reported_metric": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "Interpretation_limit": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot. Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.", "Source_IDs": "WATCH25AM-PTTEP-XBOT; WATCH25AM-PTTEP-OPEX", "Source_date": "Unknown", "Source_type": "Operator annual report", "Source_locator": "Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate.", "Source_URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Evidence_status": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "PE_relationship": null, "Coverage_bucket": "25 October morning targeted review", "Time_scope": "Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.", "Context": "Purchase-order processing and a backlog of pending orders; underlying task bottlenecks are not described.", "Implementation": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing. The annual report describes improved processing, decision support and compliance. It does not identify model, vendor, user population or approval permissions.", "Assessment": "Which steps changed, who approves orders, and do the time gains persist after all operating costs?", "Primary_department": "Supply chain & logistics", "Secondary_departments": ["Finance, investment & strategy"], "Department_assignment_basis": "Analyst mapping from purchase orders and pending procurement backlog; organizational owner undisclosed.", "Scope_type": "Internal adoption", "Review_tier": "Reviewed source", "Provenance": ["25 October morning original-source named-customer review"], "Reference_IDs": [], "Related_IDs": ["AI-245", "AI-297"], "Disclosures": [{"ID": "WATCH25AM-AI-306", "Initiative_ID": "AI-306", "Title": "PTTEP reports faster purchase-order processing with X.Bot", "Date": "Unknown", "Finding": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing.", "Metric": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "Source_IDs": "WATCH25AM-PTTEP-XBOT; WATCH25AM-PTTEP-OPEX", "Source_URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Evidence": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "Origin": "25 October morning editor original-source check alongside three bounded Luna tracks", "Assessment": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot."}], "Latest_disclosure": "Unknown", "Has_pending_disclosures": false, "Assessment_origin": "Analyst assessment", "Research_packets": ["watch25am-ai-306"], "Research_search": "{\"packet_id\": \"watch25am-ai-306\", \"key\": \"watch25am-ai-306\", \"record_ids\": [\"AI-306\"], \"company\": \"PTTEP\", \"title\": \"PTTEP reports faster purchase-order processing with X.Bot\", \"category\": \"Internal adoption\", \"department\": \"Supply chain & logistics\", \"report_date\": \"2026-10-25\", \"verification\": \"verified operator\", \"stage\": \"Operating procurement workflow; scale undisclosed\", \"business_problem\": \"Purchase-order processing and a backlog of pending orders; underlying task bottlenecks are not described.\", \"workflow\": \"PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing. The annual report describes improved processing, decision support and compliance. It does not identify model, vendor, user population or approval permissions.\", \"technology_data\": \"AI plus automation; no LLM, autonomous purchasing or X.brain integration established.\", \"integration_human_role\": \"Who approves or releases purchase orders, reviews the output or overrides it is not disclosed.\", \"deployment_scope\": \"Procurement process; geography, business-unit and supplier coverage unknown. Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.\", \"economics\": \"Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.\", \"practical_relevance\": \"Which steps changed, who approves orders, and do the time gains persist after all operating costs?\", \"assessment\": \"Which steps changed, who approves orders, and do the time gains persist after all operating costs?\", \"limitations\": [\"No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.\", \"Average per-PO time and six-month backlog change; baseline values and cohort unknown.\"], \"source_ids\": [\"WATCH25AM-PTTEP-XBOT\", \"WATCH25AM-PTTEP-OPEX\"], \"partners\": [], \"disclosed_results\": [{\"metric\": \"Purchase-order process measures\", \"value\": \"Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.\", \"period\": \"Evaluation dates undisclosed; backlog change within six months\", \"attribution\": \"Operator-reported; AI and conventional automation not isolated\", \"source_ids\": [\"WATCH25AM-PTTEP-XBOT\"]}], \"timeline\": [{\"date\": \"Unknown\", \"event\": \"Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.\", \"source_ids\": [\"WATCH25AM-PTTEP-XBOT\", \"WATCH25AM-PTTEP-OPEX\"]}], \"sources\": [{\"id\": \"WATCH25AM-PTTEP-XBOT\", \"title\": \"PTTEP 56-1 One Report 2025 — Procurement X.Bot\", \"url\": \"https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025\", \"date\": \"Unknown\", \"publisher\": \"PTTEP\", \"locator\": \"Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate.\", \"excerpt\": \"average processing time by five hours per purchase order\", \"access\": \"Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.\", \"format\": \"pdf\", \"type\": \"Operator annual report\", \"review_status\": \"Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.\"}, {\"id\": \"WATCH25AM-PTTEP-OPEX\", \"title\": \"PTTEP organizes OpEx Day 2025, advancing operational excellence\", \"url\": \"https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence\", \"date\": \"2025-07-09\", \"publisher\": \"PTTEP\", \"locator\": \"Press release, procurement AI sentence in operational-excellence examples.\", \"excerpt\": \"application of AI to streamline procurement processes\", \"access\": \"Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.\", \"type\": \"Operator press release\", \"review_status\": \"Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.\"}], \"field_sources\": {\"business_problem\": [\"WATCH25AM-PTTEP-XBOT\"], \"workflow\": [\"WATCH25AM-PTTEP-XBOT\"], \"technology_data\": [\"WATCH25AM-PTTEP-XBOT\"], \"integration_human_role\": [\"WATCH25AM-PTTEP-XBOT\"], \"deployment_scope\": [\"WATCH25AM-PTTEP-XBOT\"], \"economics\": [\"WATCH25AM-PTTEP-XBOT\"]}}", "Research_status": "Added 25 October 2026; missed 2025 annual-report evidence; operator-reported process results", "Research_category": "Internal adoption", "Research_verification": "verified operator", "Research_origin": ["research/watch-2026-10-25-am/README.md"], "Record_type": "company_adoption", "Evidence_grade": "A", "Evidence_label": "A · Operator filing", "Attribution": "operator", "Outcome_type": "realized", "Stage_norm": "operating", "Scale_note": "Purchase-order process in use; PO volume, users, geography and eligible population undisclosed.", "Segment_norm": "NOC", "Cost_cash_disclosed": false, "Financial_benefits_disclosed": false, "Single_source_vendor": false, "Review_note": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "Outcome_counted": true, "Outcome_count_note": "Reported process time/backlog measures from a combined AI-and-automation system; no isolated AI effect, cash benefit or net ROI.", "Current_source_check": "2026-10-25", "Evidence_review": {"source_read": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "claim_status": "Operating procurement workflow; scale undisclosed", "independent_validation": "Not established"}, "Detail_words": 32}], "departments": ["Exploration & geoscience", "Reservoir engineering & field development", "Drilling & completions", "Production & field operations", "Maintenance & reliability", "Refining, LNG & process operations", "Supply chain & logistics", "Land & legal", "Finance, investment & strategy", "IT, data & enterprise transformation", "Research & innovation", "Commercial, power & infrastructure", "Workforce & learning", "HSE & emissions", "Trading & marketing"], "shortlist": [{"category": "Operational AI ambition", "ids": ["AI-040"], "name": "ADNOC / AIQ — ENERGYai", "rationale": "Domain agents intended to connect seismic interpretation, reservoir monitoring and production insights, with a plan to scale to thousands of wells.", "evidence_limit": "January 2025 proof of concept followed by a March 2025 $340m, three-year contract for ENERGYai and related solutions. Planned rollout is not proof of completion.", "source": "S23"}, {"category": "Industry software platform", "ids": ["AI-008", "AI-009", "AI-047"], "name": "SLB — Tela / Lumi / marketplace", "rationale": "An agent layer embedded across existing upstream applications, backed by shared data infrastructure and extensible customer and partner agents.", "evidence_limit": "Related platform layers grouped here. Launch, engagement and platform customer counts do not establish autonomous performance or AI-only returns.", "source": "S45"}, {"category": "Upstream workflow integration", "ids": ["AI-010"], "name": "Shell / SLB — agentic collaboration", "rationale": "A proposed common data and AI environment spanning subsurface, well construction and production, with broader industry commercialization potential.", "evidence_limit": "December 2025 collaboration agreement establishes intent and architecture, not completed deployment.", "source": "S46"}, {"category": "Enterprise deployment scale", "ids": ["AI-038"], "name": "Aramco — industrial AI portfolio", "rationale": "442 identified use cases, over 200 deployed solutions and over 100 in development at the late-2025 disclosure.", "evidence_limit": "A portfolio snapshot rather than a single project. Counts are company reported and not a like-for-like measure of complexity, adoption or returns.", "source": "S22"}, {"category": "Organizational operating model", "ids": ["AI-031"], "name": "Repsol / Accenture — multi-agent platform", "rationale": "Coordinated teams of specialized AI agents with dedicated knowledge bases and human supervision change how work is assigned and executed.", "evidence_limit": "Pilot in Digitalization and Services. The separate 22 scaled GenAI use cases are not the same as the pilot agents.", "source": "S19"}, {"category": "Private-equity operating platform", "ids": ["AI-052", "AI-053", "AI-054", "AI-055"], "name": "Quantum — Energy Cloud / QIQ / investment and operations", "rationale": "A shared information base connects institutional knowledge, acquisition analysis, capital allocation and portfolio operating workflows.", "evidence_limit": "Firm-described and undated. Public material does not identify adoption by portfolio company or independently verified economics; outside strict dated-announcement counts.", "source": "S31"}, {"category": "PE-backed commercial workflow platform", "ids": ["AI-065"], "name": "Enverus ONE", "rationale": "Four launch workflows connect AFE evaluation, production valuation, project siting and internal-document knowledge with a governed energy AI platform.", "evidence_limit": "Product launch in April 2026. Platform breadth and advertised time savings do not establish realized customer returns.", "source": "S48"}, {"category": "Asset-wide data and AI integration", "ids": ["AI-078"], "name": "TotalEnergies / Cognite", "rationale": "A three-year program spanning all operated upstream assets worldwide, connecting drilling and production through a shared industrial data foundation.", "evidence_limit": "September 2025 announcement establishes intended scope. Completion and isolated AI returns remain undisclosed.", "source": "R024"}, {"category": "Enterprise adoption with named workflows", "ids": ["AI-084"], "name": "OMV / AI Hub", "rationale": "A common employee entry point accompanies a portfolio of deployed AI projects and named subsurface, drilling and maintenance workflows.", "evidence_limit": "Company-reported adoption and portfolio counts. Individual workflow economics remain limited.", "source": "R037"}], "sources": [{"Source_ID": "S01", "Source_date": "2025-11-12", "Company": "Chevron", "Source_type": "Investor day transcript", "Title": "2025 Chevron Investor Day transcript", "URL": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "Locator": "PDF pp. 21–23, 27–28, 33–34; slides 38–39; Jeff Gustavson / Mark Nelson", "Short_excerpt": "APOLO uses multivariate machine learning to analyze data from more than 50 thousand wells", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-11-12", "status": "full text reviewed", "locator": "Slides 39–40; transcript pp. 21–22"}], "Audit_reviews": [{"url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "title": "APOLO prospectivity and development optimization", "publication_date": "2025-11-12", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}], "Current_review": {"source_read": "Earlier full-text review recorded; not reverified in the later audit", "current_access": "Inaccessible in the 27 September audit and 5 October recheck", "last_checked": "2026-10-05", "independent_validation": "Not established", "note": "Retain the earlier review as provenance. A failed reopen is an access limitation, not evidence the claim is false."}}, {"Source_ID": "S02", "Source_date": "2026-02", "Company": "Devon Energy", "Source_type": "Earnings call transcript", "Title": "Q4 2025 commentary transcript", "URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Locator": "PDF p. 5; Clay Gaspar, prepared remarks", "Short_excerpt": "accelerating the implementation of AI-enabled artificial lift optimization and advanced analytics", "Verification": "inaccessible", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-02", "status": "inaccessible", "locator": "Existing register cites second-half trials and 2026 rollout; PDF not independently opened in this pass"}], "Audit_reviews": [{"url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "title": "Devon Q4 2025 Earnings Q&A Webcast Commentary Transcript", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing profile claims second-half 2025 lift trials and Q1–Q2 2026 broader rollout; those details remain unverified in this review."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S03", "Source_date": "2026-02", "Company": "Devon Energy", "Source_type": "Earnings presentation", "Title": "Q4 2025 earnings presentation", "URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf", "Locator": "Slides 7 and 12", "Short_excerpt": "Utilizing AI models/agents to flag unproductive time and best practices", "Verification": "inaccessible", "Notes": "Deck date recorded at month precision; associated call was February 18.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-02", "status": "inaccessible", "locator": "Existing profile cites D&C monitoring / AI agents; exact slide not reopened"}], "Audit_reviews": [{"url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf", "title": "AI agents for drilling and completions monitoring", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing profile says models and agents identify unproductive time and repeat better drilling/completion practices; claim is not reverified in this pass."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S04", "Source_date": "2026-01-23", "Company": "SLB", "Source_type": "Earnings call transcript", "Title": "Q4 2025 earnings call", "URL": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f", "Locator": "PDF pp. 3, 10; Olivier Le Peuch", "Short_excerpt": "launching Tela, an agentic AI assistant purpose-built to transform the upstream energy sector", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f", "access": "full text reviewed", "title": "Q4 2025 earnings call", "publication_date": "2026-01-23", "claims_checked": ["AI-008 source claim(s) checked against cited profile", "AI-009 source claim(s) checked against cited profile"], "notes": "SLB Q4 2025 earnings call transcript opened; partner/product references and broad digital ARR checked.", "partition": 2}]}, {"Source_ID": "S05", "Source_date": "2026-01-23", "Company": "SLB and customers", "Source_type": "Earnings release", "Title": "Fourth-quarter and full-year 2025 results — Digital highlights", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "Locator": "Digital and New Horizons of Growth sections", "Short_excerpt": "engineers to diagnose issues and optimize wells in minutes instead of days", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2026-01-23", "status": "full text reviewed", "locator": "Digital segment customer collaborations section"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "access": "full text reviewed", "title": "Fourth-quarter and full-year 2025 results — Digital highlights", "publication_date": "2026-01-23", "claims_checked": ["engineers to diagnose issues and optimize wells in minutes instead of days"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "access": "full text reviewed", "title": "Fourth-quarter and full-year 2025 results — Digital highlights", "publication_date": "2026-01-23", "claims_checked": ["AI-013 source claim(s) checked against cited profile", "AI-015 source claim(s) checked against cited profile"], "notes": "SLB Q4 results page opened; broad digital business results distinguishable from AI-specific economics.", "partition": 2}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "title": "SLB Q4 and full-year 2025 results", "publication_date": "2026-01-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Q4 report repeats Shell collaboration aims and scope; this is an in-window disclosure/update, not evidence the collaboration was launched in January or deployed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S06", "Source_date": "2025-11-03", "Company": "ADNOC / SLB", "Source_type": "Company announcement", "Title": "ADNOC and SLB launch AI-powered production optimization", "URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "Locator": "Opening paragraphs and deployment roadmap", "Short_excerpt": "initial deployment across eight fields", "Verification": "full text reviewed", "Notes": "All-field deployment target is 2027; a target, not evidence of completion.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-11-03", "status": "full text reviewed", "locator": "AiPSO deployment paragraphs"}], "Audit_reviews": [{"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "access": "full text reviewed", "title": "ADNOC and SLB launch AI-powered production optimization", "publication_date": "2025-11-03", "claims_checked": ["Production System Optimization platform launched with initial deployment across eight fields.", "Connects office and field workflows so engineers diagnose issues and optimize wells in minutes rather than days; planned deployment across all onshore/offshore fields by 2027.", "ADNOC says complex tasks can be completed up to ten times faster; this is a company claim, not independently validated."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-03; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "S07", "Source_date": "2025-Q1", "Company": "Halliburton", "Source_type": "Investor presentation", "Title": "Q1 2025 investor presentation", "URL": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "Locator": "Slides 11 and 19", "Short_excerpt": "Combines physics-based models and machine learning to drill wells autonomously, consistently, and on-target", "Verification": "full text reviewed", "Notes": "Quarter label used; precise publication day not established.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-Q1", "status": "full text reviewed", "locator": "Slides 11 and 19"}, {"cohort": "services", "date": "2025-Q1", "status": "full text reviewed", "locator": "Slides 11 and 19"}, {"cohort": "services", "date": "2025-Q1", "status": "full text reviewed", "locator": "Slides 11 and 19"}], "Audit_reviews": [{"url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "access": "full text reviewed", "title": "Q1 2025 investor presentation", "publication_date": "2025-Q1", "claims_checked": ["AI-016 source claim(s) checked against cited profile", "AI-017 source claim(s) checked against cited profile"], "notes": "Halliburton Q1 2025 investor presentation opened; LOGIX and DecisionSpace product portfolio statements reviewed.", "partition": 2}]}, {"Source_ID": "S08", "Source_date": "2026-07-21", "Company": "Halliburton", "Source_type": "Earnings release", "Title": "Second-quarter 2026 results", "URL": "https://www.halliburton.com/en/about-us/press-release/halliburton-announces-second-quarter-2026-results", "Locator": "Selective Technology & Highlights", "Short_excerpt": "applied AI to support predictive, asset-level decision-making over the full production lifecycle", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.halliburton.com/en/about-us/press-release/halliburton-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results", "publication_date": "2026-07-21", "claims_checked": ["AI-018 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "S09", "Source_date": "2026-01-26", "Company": "Baker Hughes", "Source_type": "Earnings presentation", "Title": "4Q 2025 results", "URL": "https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf", "Locator": "Slide 6", "Short_excerpt": "increasing 3-year data center order target to $3B from $1.5B", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf", "title": "Data-center orders", "publication_date": "2026-01-26", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S10", "Source_date": "2025-04-15", "Company": "Baker Hughes / EPAM", "Source_type": "Partner presentation announcement", "Title": "Speed to value in asset performance with Azure AI", "URL": "https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes", "Locator": "Webinar description and key learnings", "Short_excerpt": "Implement Generative AI tools to enhance asset strategy decisions and operations.", "Verification": "excerpt reviewed", "Notes": "Partner-authored webinar page; full webinar was not viewed.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-04-15", "status": "excerpt reviewed", "locator": "Description and key learnings"}], "Audit_reviews": [{"url": "https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes", "title": "Cordant Asset Strategy AI with EPAM", "publication_date": "2025-04-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["EPAM describes generative AI and unstructured text processing for Cordant Asset Strategy, including engineering document analysis and Azure integration."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S11", "Source_date": "2025-06", "Company": "Baker Hughes / C3 AI", "Source_type": "Regulatory filing", "Title": "C3 AI FY2025 Form 10-K", "URL": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm", "Locator": "Strategic partnerships; April 2025 renewal", "Short_excerpt": "co-selling efforts, co-investment in AI solutions, and scaling deployments of joint solutions", "Verification": "Primary filing indexed passage reviewed", "Notes": "Renewal occurred April 2025; filing month June. C3 AI later states Baker Hughes is no longer exclusive reseller.", "Accessed": "2026-09-26", "Audit_reviews": [{"source_id": "S11", "url": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm", "access": "full text reviewed", "title": "C3 AI FY2025 Form 10-K", "publisher": "C3 AI", "publication_date": "2025-06-26", "source_type": "SEC regulatory filing", "locator": "FY2025 Form 10-K, strategic partnership discussion", "claims_checked": ["C3 AI says Baker Hughes renewed and expanded the agreement in April 2025; the filing was submitted in June.", "The agreement continues co-selling, co-investment, and scaling joint solutions; C3 AI states Baker Hughes is no longer its exclusive oil-and-gas reseller."], "notes": "Direct SEC filing opened and relevant strategic-partnership and exclusivity text searched. Renewal date is the operating/corporate event; filing publication month is June 2025.", "accessed": "2026-09-27"}]}, {"Source_ID": "S12", "Source_date": "2026-06-24", "Company": "C3 AI", "Source_type": "Regulatory filing", "Title": "FY2026 Form 10-K", "URL": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497", "Locator": "p. 35, channel partnerships", "Short_excerpt": "Baker Hughes is no longer C3 AI’s exclusive reseller.", "Verification": "Primary filing indexed passage reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497", "access": "full text reviewed", "title": "FY2026 Form 10-K", "publication_date": "2026-06-24 (Form 10-K; FY ended 2026-04-30)", "claims_checked": ["Page 35 says Baker Hughes ceased being C3 AI’s exclusive oil-and-gas channel/reseller after amendments in January 2023 and April 2025.", "Filing still describes a strategic relationship; it does not say the partnership ended."], "notes": "Primary C3 AI filing directly opened and claim passage located."}]}, {"Source_ID": "S13", "Source_date": "2025-10-31", "Company": "ExxonMobil", "Source_type": "Earnings call transcript", "Title": "Q3 2025 earnings call", "URL": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "Locator": "Darren Woods response to Paul Sankey", "Short_excerpt": "We’re deploying it downhole in the work that we’re doing in the Permian.", "Verification": "Third-party-hosted transcript reviewed", "Notes": "Official transcript not obtained for this call; quotation apostrophes normalized.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "access": "full text reviewed", "title": "Q3 2025 earnings call", "publication_date": "2025-10-31", "claims_checked": ["We’re deploying it downhole in the work that we’re doing in the Permian."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S14", "Source_date": "2026-01-30", "Company": "ExxonMobil", "Source_type": "Earnings call transcript", "Title": "Q4 2025 earnings call", "URL": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "Locator": "Kathryn Mikells response on SAP / ERP transformation", "Short_excerpt": "the ability to apply AI at scale much more easily", "Verification": "Third-party-hosted transcript reviewed", "Notes": "Official earnings page located; transcript download returned tool error.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "access": "full text reviewed", "title": "Q4 2025 earnings call", "publication_date": "2026-01-30", "claims_checked": ["the ability to apply AI at scale much more easily"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S15", "Source_date": "2025-03", "Company": "Shell", "Source_type": "Investor presentation", "Title": "Capital Markets Day 2025 — accessible slides", "URL": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "Locator": "Slides 15 and 46", "Short_excerpt": "Cost savings from technology & AI implementation", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "title": "Technology and AI cost reduction", "publication_date": "2025-03", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S16", "Source_date": "2026", "Company": "Eni", "Source_type": "Investor presentation", "Title": "2026 Capital Markets Update", "URL": "https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "Locator": "Slide 59", "Short_excerpt": "ENI AI | use cases: around 300", "Verification": "Primary source text reviewed", "Notes": "Year confirmed; exact event date not separately checked.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "access": "full text reviewed", "title": "2026 Capital Markets Update", "publication_date": "2026", "claims_checked": ["ENI AI | use cases: around 300"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S17", "Source_date": "2025-07-11", "Company": "Eni / Khazna", "Source_type": "Company announcement", "Title": "500 MW AI Data Center Campus near Milan", "URL": "https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html", "Locator": "Opening paragraphs", "Short_excerpt": "total IT capacity of 500 MW", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html", "access": "full text reviewed", "title": "500 MW AI Data Center Campus near Milan", "publication_date": "2025-07-11", "claims_checked": ["total IT capacity of 500 MW"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S18", "Source_date": "2025-06-12", "Company": "TotalEnergies / Mistral AI", "Source_type": "Company announcement", "Title": "AI collaboration and innovation lab", "URL": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "Locator": "Shared AI innovation lab and technological ambition sections", "Short_excerpt": "design an assistant for its 1,000 researchers", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2025-06-12", "status": "full text reviewed", "locator": "A shared AI innovation lab"}], "Audit_reviews": [{"url": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "title": "Joint AI innovation lab for researcher and industrial decision support", "publication_date": "2025-06-12", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Joint lab use cases include an assistant for 1,000 researchers, decision-support for industrial performance and CO2 emissions, and customer energy-saving support.", "TotalEnergies reported Digital Factory had developed over 100 solutions, 60 using ML through generative AI; these are aggregate portfolio counts, not outcome metrics."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S19", "Source_date": "2025-10-15", "Company": "Repsol / Accenture", "Source_type": "Company announcement", "Title": "Intelligent multi-agent platform launch", "URL": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "Locator": "Platform announcement and generative AI competence center", "Short_excerpt": "more than 60 use cases in production, 22 of which have already been scaled up", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2025-10-15", "status": "full text reviewed", "locator": "Pilot project / GenAI competence center"}, {"cohort": "majors_backfill", "date": "2025-10-15", "status": "full text reviewed", "locator": "Competence Center, platform pilot and user adoption"}], "Audit_reviews": [{"url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "title": "Multi-agent workflow platform for employee productivity", "publication_date": "2026-02-19", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Accenture collaboration produced an agent-construction platform and pilot in Digitalization and Services department.", "Three teams with 22 agents supported processes for over four months; over 50 people supervised the agent teams.", "Repsol reports 60 production GenAI use cases and 22 scaled as of Oct 2025; these are company-reported portfolio measures.", "Repsol says its GenAI Competence Center had produced 60+ use cases in production and 22 scaled; 5,000+ employees trained.", "Platform was a pilot in Digitalization and Services with 3 teams and 22 agents supporting processes for more than four months; 50+ people supervised agents.", "A separate 2025 consolidated management report later says 34 agents across three workflows; dates/scope differ, so counts should not be blended."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S20", "Source_date": "2026-01-07", "Company": "Equinor", "Source_type": "Company announcement", "Title": "AI value creation and savings in 2025", "URL": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "Locator": "Three implemented contributors and Hege Skryseth comments", "Short_excerpt": "We primarily use “traditional” machine learning on our operational data.", "Verification": "full text reviewed", "Notes": "Quotation uses normalized typographic quotation marks.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2026-01-07", "status": "full text reviewed", "locator": "Implemented applications and value disclosure"}], "Audit_reviews": [{"url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "title": "AI-enabled maintenance, field planning and seismic portfolio", "publication_date": "2026-01-07", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Company says AI generated USD 130m value for Equinor and partners in calendar 2025, with >USD 330m cumulative since 2020.", "Three applications: 700+ rotating machines/24,000 sensors predictive maintenance (USD 120m since 2020); well and field planning, including USD12m Johan Sverdrup phase-3 saving; seismic interpretation of 2m km2 in 2025 with tenfold capacity."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S21", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Results media-call remarks", "Title": "CEO opening remarks: FY2025 results", "URL": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "Locator": "Amin Nasser; technology performance paragraph", "Short_excerpt": "$5.3 billion in Technology Realized Value in 2025, with over half enabled by AI", "Verification": "Full primary text reviewed", "Notes": "Media results call; not classified as investor Q&A transcript. Date aligns with FY2025 release.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2026-03-10", "status": "Full primary text reviewed", "locator": "Technology Realized Value paragraph"}], "Audit_reviews": [{"url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "access": "full text reviewed", "title": "CEO opening remarks: FY2025 results", "publication_date": "2026-03-10", "claims_checked": ["AI-037: $5.3bn total 2025 TRV; CEO says over half enabled by AI; mixed-technology components are not an AI-only breakdown."], "notes": "Primary Aramco FY2025 remarks/presentation reviewed. CEO remarks explicitly state total TRV was $5.3bn, with over half enabled by AI; this is high-level company attribution, not independently validated project-level value.", "partition": 2}]}, {"Source_ID": "S22", "Source_date": "2025-11-24", "Company": "Aramco", "Source_type": "Company announcement", "Title": "AI and supercomputing digital transformation", "URL": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "Locator": "AI adoption, HUMAIN and NVIDIA sections", "Short_excerpt": "$1.8 billion of AI-driven Technology Realized Value in 2024", "Verification": "Full primary text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2025-11-24", "status": "Full primary text reviewed", "locator": "AI deployment and NVIDIA sections"}], "Audit_reviews": [{"url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "access": "full text reviewed", "title": "AI and supercomputing digital transformation", "publication_date": "2025-11-24", "claims_checked": ["AI-038 source claim(s) checked against cited profile", "AI-039 source claim(s) checked against cited profile"], "notes": "Aramco AI/supercomputing page opened; broad AI portfolio claims reviewed; no single workflow results certified from it.", "partition": 2}]}, {"Source_ID": "S23", "Source_date": "2025-01-16", "Company": "ADNOC / AIQ", "Source_type": "Company announcement", "Title": "ENERGYai proof of concept", "URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution/", "Locator": "Proof-of-concept results and next-stage roadmap", "Short_excerpt": "70% improvement in accuracy in major seismic interpretation aspects", "Verification": "Primary source text reviewed", "Notes": "URL contains 2024 but article explicitly dated January 16, 2025.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution/", "access": "full text reviewed", "title": "ENERGYai proof of concept", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}]}, {"Source_ID": "S24", "Source_date": "updated 2026-05-27; page undated", "Company": "Woodside", "Source_type": "Operator publication", "Title": "Artificial Intelligence at Woodside", "URL": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "Locator": "Maintenance section; last-updated stamp at foot of page", "Short_excerpt": "a potential 15% reduction of maintenance hours", "Verification": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "Notes": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.", "Accessed": "2026-10-18", "Research_reviews": [{"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Startup Advisor"}, {"cohort": "majors_backfill", "date": "updated 2026-05-27; page undated", "status": "full text reviewed", "locator": "Lumina, Copilot, ChatGPT Enterprise and governance"}], "Audit_reviews": [{"url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "access": "full text reviewed", "title": "Artificial Intelligence at Woodside", "publication_date": "updated 2026-05-27; page undated", "claims_checked": ["AI-041 source claim(s) checked against cited profile", "AI-042 source claim(s) checked against cited profile", "AI-043 source claim(s) checked against cited profile"], "notes": "Woodside current AI page opened; Lumina and Startup Advisor claims reviewed, but date/first-deployment absent.", "partition": 2}], "Publisher": "Woodside", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original text read; targeted workflow/metric check", "independent_validation": "Not established", "note": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified."}}, {"Source_ID": "S25", "Source_date": "2026-05-06", "Company": "Occidental", "Source_type": "Earnings call transcript", "Title": "Q1 2026 earnings call", "URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "Locator": "PDF p. 3, prepared remarks; PDF p. 9, workforce-efficiency strategy; PDF p. 13, AI across disciplines", "Short_excerpt": "advanced analytics and AI to further optimize development and performance", "Verification": "excerpt reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-05-06", "status": "excerpt reviewed", "locator": "CEO remarks on AI across disciplines"}, {"cohort": "north_america", "date": "2026-05-06", "status": "excerpt reviewed", "locator": "COO answer on workforce efficiency"}], "Audit_reviews": [{"url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "title": "AI-enhanced subsurface development", "publication_date": "2026-05-06", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing register cites AI-enhanced subsurface development; excerpt reviewed from same transcript only supports a broader increase in AI use and does not name this workflow.", "COO said AI is taking a larger role across company disciplines as part of workforce efficiency; no workflow, deployed scale or AI-only savings quantified."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}], "Current_review": {"source_read": "Relevant official transcript passages reviewed", "current_access": "Official PDF reopened on 9 October; workforce-efficiency and broader-AI passages read", "last_checked": "2026-10-09", "independent_validation": "Not established", "note": "Review covers the management statements. It does not establish deployment or realized AI returns."}}, {"Source_ID": "S26", "Source_date": "2026-08-10", "Company": "Diamondback Energy", "Source_type": "Earnings call transcript", "Title": "Q2 2026 earnings call transcript", "URL": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "Locator": "Chad McAllaster / Kaes Van’t Hof response to Geoff Jay", "Short_excerpt": "We’re tackling it first on artificial lift", "Verification": "Third-party-hosted transcript reviewed", "Notes": "Date is transcript publication date; not independently verified as call date. Apostrophe normalized.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "Q2 2026 earnings call transcript", "publication_date": "2026-08-10", "claims_checked": ["We’re tackling it first on artificial lift"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S27", "Source_date": "2026-07-24", "Company": "SLB", "Source_type": "Earnings release", "Title": "Second-quarter 2026 results — Digital highlights", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Locator": "Digital Marketplace, Qualcomm and technology highlights", "Short_excerpt": "enable edge AI solutions for the energy industry", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-07-24", "status": "full text reviewed", "locator": "Quarterly highlights, contract awards and technology highlights"}, {"cohort": "services", "date": "2026-07-24", "status": "full text reviewed", "locator": "Quarterly highlights, contract awards and technology highlights"}, {"cohort": "services", "date": "2026-07-24", "status": "full text reviewed", "locator": "Quarterly highlights, contract awards and technology highlights"}, {"cohort": "services", "date": "2026-07-24", "status": "full text reviewed", "locator": "Quarterly highlights, contract awards and technology highlights"}, {"cohort": "majors_backfill", "date": "2026-07-24", "status": "full text reviewed", "locator": "Digital/technology projects section"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results — Digital highlights", "publication_date": "2026-07-24", "claims_checked": ["Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy).", "Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs.", "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-24; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results — Digital highlights", "publication_date": "2026-07-24", "claims_checked": ["AI-046 source claim(s) checked against cited profile", "AI-047 source claim(s) checked against cited profile", "AI-048 source claim(s) checked against cited profile", "AI-067 source claim(s) checked against cited profile"], "notes": "SLB Q2 2026 results opened; PTTEP/Cognite/OptiFlow/Tela offshore pilot, ONGC DrillOps and CNOOC terminal automation statements checked as partner claims.", "partition": 2}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "title": "Closed-loop process optimization at CNOOC Bohai offshore terminal", "publication_date": "2026-07-24", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy).", "Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs.", "Q2 disclosure: PTTEP one-asset offshore pilot using Cognite, OptiFlow and Tela agents.", "Q2 disclosure: CNOOC subsidiary terminal connects Symmetry twins to APC/DCS.", "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S28", "Source_date": "2025-09-16", "Company": "Aker BP / Cognite", "Source_type": "Partner announcement", "Title": "Aker BP expands Cognite Atlas AI adoption", "URL": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "Locator": "Partnership announcement", "Short_excerpt": "expanded adoption of Cognite Atlas AI", "Verification": "full text reviewed", "Notes": "Technology partner announcement.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-09-16", "status": "full text reviewed", "locator": "RCA agents and additional workflows"}], "Audit_reviews": [{"url": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "title": "Atlas AI agents for root-cause analysis and E&P workflows", "publication_date": "2025-09-16", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Aker BP expanded Atlas AI adoption. Described workflows include document processing for supply chain, barrier management, automatic equipment registration, data foundation for predictive maintenance and production chemistry.", "Root-cause agents identify relevant info, build cause maps and suggest steps; Cognite/Aker BP report >70% reduction in engineer RCA time."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S29", "Source_date": "2025-11-03", "Company": "PETRONAS", "Source_type": "Company announcement", "Title": "TIRAI myPROdata chatbot launch", "URL": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "Locator": "Subscriber announcement; feature release October 18, 2025", "Short_excerpt": "the TIRAI Generative AI (GenAI) Chatbot", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-11-03", "status": "full text reviewed", "locator": "TIRAI release"}], "Audit_reviews": [{"url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "access": "full text reviewed", "title": "TIRAI myPROdata chatbot launch", "publication_date": "2025-11-03", "claims_checked": ["AI-050 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "S30", "Source_date": "2025-08-06", "Company": "Blackstone / Enverus", "Source_type": "PE transaction announcement", "Title": "Agreement to acquire Enverus", "URL": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "Locator": "Transaction announcement and management statements", "Short_excerpt": "pioneering GenAI-powered solutions", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-08-06", "status": "full text reviewed", "locator": "Transaction announcement"}, {"cohort": "software_capital", "date": "2025-08-06", "status": "full text reviewed", "locator": "Transaction announcement"}], "Audit_reviews": [{"url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "access": "full text reviewed", "title": "Agreement to acquire Enverus", "publication_date": "2025-08-06", "claims_checked": ["AI-051 source claim(s) checked against cited profile"], "notes": "Blackstone agreement announcement opened; Aug 2025 definitive agreement and expected transaction timing checked.", "partition": 2}]}, {"Source_ID": "S31", "Source_date": "undated", "Company": "Quantum Capital Group", "Source_type": "PE operating-strategy disclosure", "Title": "Digital at Quantum", "URL": "https://www.quantumcap.com/digital-mission/", "Locator": "Quantum Energy Cloud; Quantum IQ; AI at Work", "Short_excerpt": "Our proprietary generative AI platform brings the full context of the firm’s knowledge into everyday work.", "Verification": "full text reviewed", "Notes": "Current undated page; cannot date initiatives to the 24-month window.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "undated", "status": "full text reviewed", "locator": "QEC, QIQ and AI at Work"}], "Audit_reviews": [{"url": "https://www.quantumcap.com/digital-mission/", "title": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "publication_date": "2025-05-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Quantum Energy Cloud unifies portfolio financial, operational and market data using Microsoft and Databricks.", "QIQ is a proprietary generative AI interface that answers questions, automates analyses and initiates workflows over portfolio data.", "Named workflows include single-well economics, geospatial capital allocation, completion design, artificial lift, anomaly detection and downtime prediction."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S32", "Source_date": "2026-06-25", "Company": "Quorum / Streamba", "Source_type": "PE-backed portfolio announcement", "Title": "Quorum acquires Streamba", "URL": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "Locator": "Acquisition announcement; VOR platform and roadmap", "Short_excerpt": "an AI-native supply chain platform purpose-built for energy", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-06-25", "status": "full text reviewed", "locator": "Streamba/VOR and DaWinci"}], "Audit_reviews": [{"url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "access": "full text reviewed", "title": "Quorum acquires Streamba", "publication_date": "2026-06-25", "claims_checked": ["Streamba AI-native supply chain platform and Visibility, Orchestration and Response system unify logistics data, workflows and teams.", "Quorum says agentic capabilities can identify disruption and propose or execute plan changes with human oversight."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-06-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "S33", "Source_date": "Undated; accessed 2026-09-26", "Company": "Francisco Partners", "Source_type": "Ownership corroboration", "Title": "Quorum Software investment", "URL": "https://www.franciscopartners.com/investments/quorum-software", "Locator": "Status and Detail", "Short_excerpt": "Private Buy Out", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.franciscopartners.com/investments/quorum-software", "access": "full text reviewed", "title": "Quorum Software investment", "publication_date": "Undated; accessed 2026-09-26", "claims_checked": ["Private Buy Out"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S34", "Source_date": "2025-09-25", "Company": "PakEnergy", "Source_type": "PE-backed portfolio announcement", "Title": "AI-powered lease provision launch", "URL": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "Locator": "Product release and platform AI roadmap", "Short_excerpt": "reduces processing time from hours to mere seconds", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-09-25", "status": "full text reviewed", "locator": "Land Management lease provision"}], "Audit_reviews": [{"url": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "title": "Embedded AI lease provision processing", "publication_date": "2025-09-25", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Embedded AI processes lease provision research and data inside PakEnergy Land Management.", "Company says work previously taking hours can be done in seconds; no accuracy or independent outcome data disclosed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S35", "Source_date": "2026-05-20", "Company": "PakEnergy", "Source_type": "PE-backed portfolio announcement", "Title": "PakCAPTURE launch", "URL": "https://www.prnewswire.com/news-releases/pakenergy-launches-pakcapture-delivering-ai-powered-ticket--receipt-management-to-pakenergy-transportation-302777729.html", "Locator": "Company-issued release carried by PR Newswire", "Short_excerpt": "PakCAPTURE eliminates manual data entry for drivers and back-office teams", "Verification": "Company-issued release text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"baseline_id": "AI-058", "source_id": "S35", "url": "https://www.prnewswire.com/news-releases/pakenergy-launches-pakcapture-delivering-ai-powered-ticket--receipt-management-to-pakenergy-transportation-302777729.html", "title": "PakCAPTURE launch", "publication_date": "2026-05-20", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["PakEnergy-provided release says drivers photograph tickets/receipts at pickup or delivery; PakAI extracts data, populates forms, and associates the record with a load for office processing.", "The announcement calls PakCAPTURE immediately available as part of PakEnergy Transportation; it reports no measured accuracy, time savings, or cash-flow results."], "notes": "Issuer release distributed by PR Newswire. Consistent with AI-058; supports launch capability, not realized outcomes. The baseline also links PakEnergy's own copy and another source, so this does not require a new profile."}]}, {"Source_ID": "S36", "Source_date": "2025; day not checked", "Company": "PakEnergy / Charlesbank", "Source_type": "Ownership corroboration", "Title": "Petrosight acquisition announcement", "URL": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition", "Locator": "Charlesbank support paragraph", "Short_excerpt": "Charlesbank Capital Partners (Charlesbank), PakEnergy’s private equity partner.", "Verification": "Company-issued indexed passage reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"source_id": "S36", "url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition", "access": "full text reviewed", "title": "PakEnergy Expands Solutions for Energy Customers With Petrosight Acquisition", "publisher": "PakEnergy", "publication_date": "2025-01-07", "source_type": "Company announcement", "locator": "Acquisition and Charlesbank ownership/support paragraph", "claims_checked": ["PakEnergy announced the Petrosight acquisition on Jan. 7, 2025.", "The source identifies Charlesbank as PakEnergy’s private-equity partner; Petrosight’s wellsite software organizes lifecycle data across drilling, completions, facilities/pipelines, and environmental/reclamation."], "notes": "Full official PakEnergy announcement opened. This supports ownership and deal context only; it is not evidence of AI.", "accessed": "2026-09-27"}]}, {"Source_ID": "S37", "Source_date": "2025-07-30", "Company": "ECP / KKR / CyrusOne / Calpine", "Source_type": "PE infrastructure announcement", "Title": "Texas hyperscale data-center campus", "URL": "https://media.kkr.com/news-details?news_id=05131f1d-9cac-4cc4-afb2-19a83bd4cee3", "Locator": "Campus scope, power supply and construction timing", "Short_excerpt": "Calpine has secured a long-term contract to supply dedicated power to the data center campus.", "Verification": "Primary source text reviewed", "Notes": null, "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://media.kkr.com/news-details?news_id=05131f1d-9cac-4cc4-afb2-19a83bd4cee3", "access": "full text reviewed", "title": "Texas hyperscale data-center campus", "publication_date": "2025-07-30", "claims_checked": ["AI-059 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "S38", "Source_date": "2024-10-29", "Company": "bp", "Source_type": "Earnings call transcript", "Title": "Q3 2024 earnings call", "URL": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "Locator": "Murray Auchincloss response to Lydia Rainforth", "Short_excerpt": "digital now with Palantir and Infosys that will help the refining system over time", "Verification": "Third-party-hosted transcript reviewed", "Notes": "Digital program reference; AI relationship corroborated by supplier announcement outside window. Official deck fetch blocked.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "access": "full text reviewed", "title": "Q3 2024 earnings call", "publication_date": "2024-10-29", "claims_checked": ["digital now with Palantir and Infosys that will help the refining system over time"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S39", "Source_date": "2024-09-09", "Company": "bp / Palantir", "Source_type": "Historical corroboration", "Title": "Five-year strategic relationship with new AI capabilities", "URL": "https://palantir2020ipo.q4web.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/default.aspx", "Locator": "Announcement", "Short_excerpt": "New AI Capabilities", "Verification": "Primary indexed announcement reviewed", "Notes": "17 days before confirmed scope; background only, not counted as a new in-window announcement.", "Accessed": "2026-09-26", "Audit_reviews": [{"url": "https://palantir2020ipo.q4web.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/default.aspx", "access": "full text reviewed", "title": "Five-year strategic relationship with new AI capabilities", "publication_date": "2024-09-09", "claims_checked": ["New AI Capabilities"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}]}, {"Source_ID": "S40", "Source_date": "2026-02-18", "Company": "Cenovus", "Source_type": "Operator annual filing", "Title": "2025 Management’s Discussion and Analysis", "URL": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "Locator": "Printed MD&A page 47: Technology, Information Systems and Data Privacy; page 2: document date", "Short_excerpt": "beginning with pilot-phase use cases that focus on improving productivity and enhancing decision-support capabilities", "Verification": "Original source read by editor 17 October 2026; claims remain issuer-reported.", "Notes": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.", "Accessed": "2026-10-17", "Research_reviews": [{"cohort": "north_america", "date": "2026", "status": "inaccessible", "locator": "Existing register cites MD&A AI disclosure; not reopened"}], "Audit_reviews": [{"url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "access": "full text reviewed", "title": "2025 Management’s Discussion and Analysis", "publication_date": "2026", "claims_checked": ["AI-061 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}], "Publisher": "Cenovus Energy / SEC", "Current_review": {"last_checked": "2026-10-17", "source_read": "Original source read", "current_access": "Original PDF or SEC/issuer release available", "independent_validation": "Not established", "note": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established."}}, {"Source_ID": "S41", "Source_date": "2025-02-10", "Company": "Energy Transfer / CloudBurst", "Source_type": "Company announcement", "Title": "Natural gas supply for AI-focused Texas data center", "URL": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "Locator": "Agreement scope and condition precedent", "Short_excerpt": "subject to CloudBurst reaching a final investment decision (FID) with its customer", "Verification": "full text reviewed", "Notes": null, "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-02-10", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "title": "CloudBurst San Marcos natural-gas supply for AI campus", "publication_date": "2025-02-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Up to 450,000 MMBtu/day firm gas supply for campus near San Marcos, supporting up to ~1.2 GW behind-the-meter power, contingent on customer FID; AI-demand infrastructure."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S42", "Source_date": "2025-07", "Company": "EQT / Homer City Redevelopment", "Source_type": "Partner announcement", "Title": "Agreement in principle for Homer City gas supply", "URL": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "Locator": "Opening agreement announcement", "Short_excerpt": "an agreement in principle", "Verification": "excerpt reviewed", "Notes": "Published July 15, updated July 21, 2025; project-partner announcement.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-07", "status": "excerpt reviewed", "locator": "Agreement in principle"}], "Audit_reviews": [{"url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "access": "full text reviewed", "title": "Agreement in principle for Homer City gas supply", "publication_date": "2025-07", "claims_checked": ["Project announcement describes EQT agreement in principle for gas supply to redevelopment targeting AI/HPC campus; not an executed definitive supply contract."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-07; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "S43", "Source_date": "2026-08-04", "Company": "Williams", "Source_type": "Earnings presentation", "Title": "Second-quarter 2026 earnings presentation", "URL": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "Locator": "Slide 30 (PDF p. 30); Socrates project", "Short_excerpt": "556 MW ISO capacity", "Verification": "full text reviewed", "Notes": "Date from presentation footer; project update reflects status at disclosure.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026-08-04", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "access": "full text reviewed", "title": "Second-quarter 2026 earnings presentation", "publication_date": "2026-08-04", "claims_checked": ["Project details include gas supply, pipeline and generation; presented as power infrastructure for data-center load."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-08-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "S44", "Company": "Chevron", "Source_date": "2026-06-22", "Title": "Microsoft power agreement for Project Kilby", "URL": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center", "Locator": "Project terms and forward-looking milestones", "Source_type": "Primary announcement / investor material", "Notes": "Added during contextual expansion. Accessed 2026-09-26.", "Audit_reviews": [{"url": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center", "access": "inaccessible", "title": "Microsoft power agreement for Project Kilby", "publication_date": "2026-06-22", "claims_checked": ["Attempted Chevron’s direct announcement URL; web fetch returned HTTP 403."], "notes": "Not reverified from the cited page; this does not disprove the agreement. No claim promoted to verified from access metadata."}]}, {"Source_ID": "S45", "Company": "SLB", "Source_date": "2025-11-03", "Title": "Tela agentic AI launch", "URL": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "Locator": "Product architecture and examples", "Source_type": "Primary announcement / investor material", "Notes": "Added during contextual expansion. Accessed 2026-09-26.", "Research_reviews": [{"cohort": "services", "date": "2025-11-03", "status": "full text reviewed", "locator": "Tela architecture and deployment availability"}], "Verification": "full text reviewed", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "access": "full text reviewed", "title": "Tela agentic AI launch", "publication_date": "2025-11-03", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "SLB Tela launch release opened; launch date and product claims checked.", "partition": 2}]}, {"Source_ID": "S46", "Company": "SLB / Shell", "Source_date": "2025-12-11", "Title": "Collaboration agreement for digital and AI solutions", "URL": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "Locator": "Lumi infrastructure and upstream scope", "Source_type": "Primary announcement / investor material", "Notes": "Added during contextual expansion. Accessed 2026-09-26.", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-12-11", "status": "full text reviewed", "locator": "Shell collaboration agreement"}], "Verification": "full text reviewed", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "title": "Upstream data/AI infrastructure and agentic-solutions collaboration", "publication_date": "2026-01-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["SLB and Shell agreed to develop upstream digital and AI solutions; roadmap aims for open data/AI infrastructure across subsurface, well construction and production on SLB Lumi.", "Announcement states ambition/aim, not a deployed customer workflow or realized results."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "S47", "Company": "Blackstone", "Source_date": "2026-04-23", "Title": "First-quarter 2026 earnings results", "URL": "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf", "Locator": "Printed slide 11, PDF page 15, private-equity capital deployed", "Source_type": "Primary announcement / investor material", "Notes": "Added during contextual expansion. Accessed 2026-09-26.", "Research_reviews": [{"cohort": "software_capital", "date": "2026-04-23", "status": "Primary PDF section reviewed", "locator": "PDF page 15 / printed page 11, Private Equity capital deployed"}], "Verification": "Primary PDF section reviewed", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf", "access": "full text reviewed", "title": "First-quarter 2026 earnings results", "publication_date": "2026-04-23", "claims_checked": ["AI-051 source claim(s) checked against cited profile"], "notes": "Blackstone Q1 2026 earnings release opened; transaction and reported capital context checked; no Enverus-specific fund deployment inferred.", "partition": 2}]}, {"Source_ID": "S48", "Company": "Enverus", "Source_date": "2026-04-07", "Title": "Enverus ONE launch", "URL": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "Locator": "Astra, four launch workflows and governance", "Source_type": "Primary announcement / investor material", "Notes": "Added during contextual expansion. Accessed 2026-09-26.", "Research_reviews": [{"cohort": "software_capital", "date": "2026-04-07", "status": "full text reviewed", "locator": "Launch, platform design and four Flows"}], "Verification": "full text reviewed", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "access": "full text reviewed", "title": "Enverus ONE launch", "publication_date": "2026-04-07", "claims_checked": ["AI-065 source claim(s) checked against cited profile"], "notes": "Enverus ONE launch announcement opened; product launch, capabilities and rollout wording checked, not customer use evidence.", "partition": 2}]}, {"Source_ID": "R001", "Source_date": "2024-03", "Company": "Saudi Aramco", "Source_type": "Middle East AI News", "Title": "Aramco launches Metabrain 250B-parameter industrial LLM", "URL": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm", "access": "excerpt reviewed", "title": "Aramco launches Metabrain 250B-parameter industrial LLM", "publication_date": "2024-03", "claims_checked": ["Historical March 4, 2024 secondary account; model size and intended tasks are reported claims. The planned larger model is not proof of completion."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R002", "Source_date": "2025-10", "Company": "Saudi Aramco", "Source_type": "Aramco", "Title": "Aramco takes significant minority stake in HUMAIN", "URL": "https://www.aramco.com/en/news-media/news/2025/pif-and-aramco-agree-for-aramco-to-acquire-a-significant-minority-stake-in-humain", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.aramco.com/en/news-media/news/2025/pif-and-aramco-agree-for-aramco-to-acquire-a-significant-minority-stake-in-humain", "access": "full text reviewed", "title": "Aramco takes significant minority stake in HUMAIN", "publication_date": "2025-10", "claims_checked": ["AI-039 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R003", "Source_date": "2026-02-12", "Company": "Saudi Aramco", "Source_type": "Aramco", "Title": "Aramco-Microsoft MoU on industrial AI and digital talent", "URL": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Full primary text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2026-02-12", "status": "Full primary text reviewed", "locator": "MoU scope and four focus areas"}], "Audit_reviews": [{"url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "access": "full text reviewed", "title": "Aramco-Microsoft MoU on industrial AI and digital talent", "publication_date": "2026-02-12", "claims_checked": ["AI-070 source claim(s) checked against cited profile"], "notes": "Aramco/Microsoft release opened; non-binding MoU and exploration language confirmed, not deployment.", "partition": 2}]}, {"Source_ID": "R004", "Source_date": "2025-06", "Company": "Saudi Aramco", "Source_type": "Aramco Europe", "Title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "URL": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai", "access": "full text reviewed", "title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "publication_date": "2025-06", "claims_checked": ["AI-038 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R005", "Source_date": "2025-11", "Company": "ADNOC", "Source_type": "The National", "Title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "URL": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/", "access": "full text reviewed", "title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "publication_date": "2025-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R006", "Source_date": "2025-03-12", "Company": "ADNOC", "Source_type": "PR Newswire / AIQ", "Title": "AIQ wins $340M contract to deploy ENERGYai agents", "URL": "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "Primary source text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "access": "full text reviewed", "title": "AIQ wins $340M contract to deploy ENERGYai agents", "publication_date": "2025-03-12", "claims_checked": ["Reviewed during consolidation; metrics remain company or vendor reported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R007", "Source_date": "2025-01", "Company": "ADNOC", "Source_type": "ADNOC", "Title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "access": "full text reviewed", "title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}]}, {"Source_ID": "R008", "Source_date": "2026-08", "Company": "ADNOC", "Source_type": "Computer Weekly", "Title": "ADNOC moves AI from pilots to enterprise-wide operations", "URL": "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "access": "full text reviewed", "title": "ADNOC moves AI from pilots to enterprise-wide operations", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R009", "Source_date": "2025-05", "Company": "ADNOC", "Source_type": "SLB", "Title": "AIQ and SLB co-develop ENERGYai for subsurface", "URL": "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr", "access": "full text reviewed", "title": "AIQ and SLB co-develop ENERGYai for subsurface", "publication_date": "2025-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R010", "Source_date": "2026-03", "Company": "ExxonMobil", "Source_type": "Offshore Magazine", "Title": "First fully autonomous AI-drilled well offshore Guyana", "URL": "https://www.offshore-mag.com/drilling-completion/news/55380289/exxonmobil-achieves-milestone-automated-well-placement-offshore-guyana", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.offshore-mag.com/drilling-completion/news/55380289/exxonmobil-achieves-milestone-automated-well-placement-offshore-guyana", "access": "inaccessible", "title": "First fully autonomous AI-drilled well offshore Guyana", "publication_date": "2026-03", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R011", "Source_date": "2026-08", "Company": "ExxonMobil", "Source_type": "World Oil", "Title": "AI identifies four new Guyana exploration opportunities", "URL": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "access": "full text reviewed", "title": "AI identifies four new Guyana exploration opportunities", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R012", "Source_date": "2024-11", "Company": "ExxonMobil", "Source_type": "Constellation Research", "Title": "AI plus process mining behind $15B cost-cut push", "URL": "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts", "access": "full text reviewed", "title": "AI plus process mining behind $15B cost-cut push", "publication_date": "2024-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R013", "Source_date": "2025-11-12", "Company": "Chevron", "Source_type": "Chevron via MarketScreener", "Title": "ApEX multi-agent GenAI searches 1M+ exploration files", "URL": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-074", "source_id": "R013", "url": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c", "title": "Chevron: A smarter way to prospect for oil and gas", "publication_date": "2025-11-12", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Chevron's article says ApEX launched in August 2024 and was first used in Gulf of America deepwater exploration; users query a chat interface and specialized agents search more than one million files.", "The article reports a 4 TB corpus and describes seconds-fast retrieval. It does not quantify a reduction from weeks of pre-drill studies to minutes."], "notes": "The supplied MarketScreener page is Chevron-authored copy redistributed via Publicnow. The same article was located and reviewed on Chevron's own newsroom page, which is the preferred primary citation. The supplied reference's December month is not the publication date; the original article is dated November 12, 2025."}]}, {"Source_ID": "R014", "Source_date": "2025-02", "Company": "Chevron", "Source_type": "Inspectioneering", "Title": "Up to 20% workforce cut, technology and hubs cited", "URL": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce", "access": "inaccessible", "title": "Up to 20% workforce cut, technology and hubs cited", "publication_date": "2025-02", "claims_checked": ["Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R015", "Source_date": "2026-06-02", "Company": "Chevron", "Source_type": "DataRobot", "Title": "DataRobot agentic AI for autonomous robotic inspections", "URL": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2026-06-02", "status": "full text reviewed", "locator": "Safe Start agent, robot missions and edge systems"}], "Audit_reviews": [{"url": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "title": "Edge agentic assessment for robotic inspections", "publication_date": "2026-06-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Chevron and DataRobot collaboration applies edge agents to support autonomous inspection at Chevron facilities.", "Safe Start assesses conditions before/during robotic missions using gas sensors, vision and locally running models; integrates with existing operating systems and NVIDIA NIM.", "Vendor and Chevron program manager describe reduced manual point-in-time checks and human oversight/control; no measured outcomes."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R016", "Source_date": "2026-06", "Company": "Shell", "Source_type": "Business Wire / C3 AI", "Title": "Shell extends C3 AI with agent-based root-cause analysis", "URL": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"source_id": "R016", "url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations", "access": "full text reviewed", "title": "C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations", "publisher": "C3 AI / Business Wire", "publication_date": "2026-06-04", "source_type": "Company announcement distributed by Business Wire", "locator": "Release body, deployment and platform description", "claims_checked": ["C3 AI says Shell’s existing predictive-maintenance program monitors more than 13,000 pieces of equipment.", "The new multi-year agreement extends Reliability and plans to introduce agent-based root-cause analysis and remediation; current predictive maintenance runs on C3 AI Reliability/Agentic AI Platform deployed on Microsoft Azure.", "Hundreds of millions of dollars in value is a C3 AI executive’s attributed statement, not independently quantified in this release."], "notes": "Full Business Wire release opened and key passage reviewed. Distinguish existing program from new planned agentic capabilities; no independent Shell validation located in this source.", "accessed": "2026-09-27"}]}, {"Source_ID": "R017", "Source_date": "2025-12", "Company": "Shell", "Source_type": "JPT (SPE)", "Title": "SLB and Shell co-develop agentic AI for upstream", "URL": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts", "access": "full text reviewed", "title": "SLB and Shell co-develop agentic AI for upstream", "publication_date": "2025-12-18", "claims_checked": ["SLB and Shell signed a strategic collaboration to develop AI/agentic AI for technical experts; planned objective is to unify data/workflows across subsurface, well construction and production, using Lumi."], "notes": "JPT/SPE report is secondary editorial confirmation; it says aim/develop/deploy, not that a production system is already deployed."}]}, {"Source_ID": "R018", "Source_date": "2024-09-09", "Company": "Shell", "Source_type": "NVIDIA Technical Blog", "Title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "URL": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/", "access": "excerpt reviewed", "title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "publication_date": "2024-09-09", "claims_checked": ["NVIDIA dates this technical article September 9, 2024, before the reporting window. Computational speedup and accuracy describe model tests, not field-wide economic returns."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R019", "Source_date": "2024-09", "Company": "BP", "Source_type": "Palantir IR", "Title": "bp renews Palantir for 5 years with AIP LLMs", "URL": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/", "access": "inaccessible", "title": "bp renews Palantir for 5 years with AIP LLMs", "publication_date": "2024-09", "claims_checked": ["Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R020", "Source_date": "2025-08", "Company": "BP", "Source_type": "Yahoo Finance UK", "Title": "bp expands job cuts as AI drives efficiency", "URL": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html", "access": "inaccessible", "title": "bp expands job cuts as AI drives efficiency", "publication_date": "2025-08", "claims_checked": ["Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R021", "Source_date": "2026-09-15", "Company": "TotalEnergies", "Source_type": "TotalEnergies", "Title": "TotalEnergies and Mistral launch €100M frontier subsurface AI program", "URL": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "Locator": "Release dated 2026-09-15; >€100m / 3-year program; “petaflops of data” wording flagged", "Short_excerpt": "", "Verification": "Relevant original primary source reviewed on 2026-10-06.", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-10-06", "Research_reviews": [{"cohort": "majors", "date": "2026-09-15", "status": "full text reviewed", "locator": "Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data"}], "Audit_reviews": [{"url": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "title": "Frontier AI models for reservoir exploration and engineering", "publication_date": "2026-09-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three-year joint program represents over €100 million investment to develop frontier AI models supporting exploration, characterization and development of oil and gas reservoirs.", "Program will use agentic AI and combine almost 10 petaflops of data with geoscience/reservoir expertise; develop multiple scenarios for exploration opportunities and optimizing existing projects.", "Scientific lab will combine TotalEnergies subsurface experts with Mistral AI science and technology."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}], "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Original issuer release independently reviewed", "publisher": "TotalEnergies", "locator": "Release dated 2026-09-15; >€100m / 3-year program; “petaflops of data” wording flagged"}, "Publisher": "TotalEnergies"}, {"Source_ID": "R022", "Source_date": "2025-06", "Company": "TotalEnergies", "Source_type": "TotalEnergies", "Title": "Mistral joint lab: assistant for 1,000 researchers", "URL": "https://totalenergies.com/news/press-releases/totalenergies-collaborate-mistral-ai-increase-application-artificial", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-029", "source_id": "R022", "url": "https://totalenergies.com/news/press-releases/totalenergies-collaborate-mistral-ai-increase-application-artificial", "title": "TotalEnergies to Collaborate with Mistral AI to Increase the Application of Artificial Intelligence", "publication_date": "2025-06-12", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["TotalEnergies announced a shared AI lab and named an assistant for its 1,000 researchers, industrial-asset decision support, customer energy support, and renewable-production use cases.", "The release describes work to design/develop use cases; it does not report production rollout or measured performance."], "notes": "Official TotalEnergies press release. Supports the existing AI-029 planned-use description and the stated researcher cohort; it does not substantiate realized impacts."}]}, {"Source_ID": "R023", "Source_date": "2025-01-24", "Company": "TotalEnergies", "Source_type": "TotalEnergies", "Title": "Early adopter of Microsoft Copilot across the workforce", "URL": "https://totalenergies.com/news/news/ia-expediting-energy-transition", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://totalenergies.com/news/news/ia-expediting-energy-transition", "access": "excerpt reviewed", "title": "Early adopter of Microsoft Copilot across the workforce", "publication_date": "2025-01-24", "claims_checked": ["The referenced TotalEnergies article is dated January 24, 2025, not February 2024. It discusses a prior Copilot adoption and continuing supervised AI use; it is context alongside existing TotalEnergies dossiers."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R024", "Source_date": "2025-09-26", "Company": "TotalEnergies", "Source_type": "TotalEnergies", "Title": "Cognite industrial AI across all operated upstream assets", "URL": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "Primary source text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Audit_reviews": [{"source_id": "R024", "url": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale", "access": "full text reviewed", "title": "Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI", "publisher": "TotalEnergies", "publication_date": "2025-09-26", "source_type": "Company announcement", "locator": "Partnership deployment scope and objectives", "claims_checked": ["The dated release says TotalEnergies will scale Cognite’s industrial data and AI platform over three years across all operated upstream assets worldwide, drilling through production.", "The intended foundation makes industrial data AI-ready and supports dynamic asset visualization, decision support, critical-equipment monitoring, and later AI application adoption."], "notes": "Full official TotalEnergies release opened after redirect to the newsroom page. It describes a three-year scale-up plan, not completed global deployment at announcement.", "accessed": "2026-09-27"}]}, {"Source_ID": "R026", "Source_date": "2025-06", "Company": "Eni", "Source_type": "Eni", "Title": "HPC6 supercomputer and company-wide responsible-AI training", "URL": "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html", "access": "full text reviewed", "title": "HPC6 supercomputer and company-wide responsible-AI training", "publication_date": "2025-06", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R027", "Source_date": "2026-02", "Company": "Repsol", "Source_type": "Repsol", "Title": "Copilot to 25,000 staff; pilot saved 121 min/week", "URL": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml", "access": "full text reviewed", "title": "Copilot to 25,000 staff; pilot saved 121 min/week", "publication_date": "Undated current Repsol webpage; includes historic 2023 launch and 4-month pilot results", "claims_checked": ["Copilot for Web extended to around 25,000 employees; Microsoft 365 Copilot licenses are being progressively scaled.", "Four-month M365 Copilot pilot involved 550 professionals and reported average 121 minutes saved weekly per participant, 16% higher work quality and >60% preference to continue.", "More than 60 use cases are “in process”; this page does not support interpreting that count as 60+ in-production cases or 22 scaled cases."], "notes": "Primary Repsol webpage opened. Metrics are company-reported pilot results; time savings are not a realized labor-cost reduction."}]}, {"Source_ID": "R029", "Source_date": "2025-02", "Company": "Repsol", "Source_type": "Accenture Newsroom", "Title": "Accenture and NVIDIA to scale AI agents", "URL": "https://newsroom.accenture.com/news/2025/repsol-boosts-its-digital-strategy-by-collaborating-with-accenture-to-adopt-ai-agents", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-031", "source_id": "R029", "url": "https://newsroom.accenture.com/news/2025/repsol-boosts-its-digital-strategy-by-collaborating-with-accenture-to-adopt-ai-agents", "title": "Repsol Boosts its Digital Strategy by Collaborating with Accenture to Adopt AI Agents", "publication_date": "2025-02-06", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Accenture's release describes Repsol's planned co-innovation and deployment of custom AI agents using Accenture and NVIDIA platforms, across processes including planning/forecasting, application maintenance, and incident resolution.", "It also says Repsol would explore agents with digital twins and robotics for industrial/logistics maintenance and expand staff training; these are forward-looking plans, not measured operating outcomes."], "notes": "Partner-issued primary announcement; Repsol also published its own release. This is an earlier collaboration/update and is consistent with AI-031's later, separately sourced deployment counts."}]}, {"Source_ID": "R030", "Source_date": "2023-12", "Company": "Petrobras", "Source_type": "Microsoft News Center Latam", "Title": "ChatPetrobras GenAI rolled out to 110,000 workers", "URL": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br", "access": "excerpt reviewed", "title": "ChatPetrobras GenAI rolled out to 110,000 workers", "publication_date": "2023-12", "claims_checked": ["Microsoft publication is December 5, 2023. It describes ChatPetrobras using Azure OpenAI; accessibility to workers is not a measured active-user count."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R031", "Source_date": "2025-07-31", "Company": "Petronas", "Source_type": "Petronas", "Title": "Petronas-Microsoft MoU on agentic AI and Copilot", "URL": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "Primary source text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "access": "full text reviewed", "title": "Petronas-Microsoft MoU on agentic AI and Copilot", "publication_date": "2025-07-31", "claims_checked": ["AI-079 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R032", "Source_date": "2026-09", "Company": "Petronas", "Source_type": "Offshore Engineer", "Title": "Agentic AI added to myPROdata upstream platform", "URL": "https://www.oedigital.com/news/542771-petronas-turns-to-ai-to-accelerate-upstream-investment", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.oedigital.com/news/542771-petronas-turns-to-ai-to-accelerate-upstream-investment", "access": "full text reviewed", "title": "Agentic AI added to myPROdata upstream platform", "publication_date": "2026-09", "claims_checked": ["AI-080 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R033", "Source_date": "2025-05", "Company": "PetroChina / CNPC", "Source_type": "CNPC", "Title": "CNPC releases 300B-parameter Kunlun energy model", "URL": "https://www.cnpc.com.cn/en/nr2025/202506/e9b44b36040f452796711916fbf49135.shtml", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.cnpc.com.cn/en/nr2025/202506/e9b44b36040f452796711916fbf49135.shtml", "access": "inaccessible", "title": "CNPC releases 300B-parameter Kunlun energy model", "publication_date": "2025-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R034", "Source_date": "2026-04", "Company": "CNOOC", "Source_type": "China Daily", "Title": "Offshore predictive maintenance cuts failures ~20%", "URL": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "title": "Offshore predictive maintenance cuts failures ~20%", "publication_date": "2026-04", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Specific claim in the assigned record was checked against the opened secondary/editorial/transcript page, not an issuer-original page."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R036", "Source_date": "2025-07", "Company": "Santos", "Source_type": "Energy News Bulletin", "Title": "Xecta AI digital twin expanded to East Australia and PNG", "URL": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta", "access": "excerpt reviewed", "title": "Xecta AI digital twin expanded to East Australia and PNG", "publication_date": "2025-07", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R037", "Source_date": "Undated; see individual source review", "Company": "OMV", "Source_type": "OMV Annual Report 2025", "Title": "OMV AI Hub: 2,500 monthly GPT users, 10,000 trained", "URL": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Digitalization; AI Across the Value Chain"}, {"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Digitalization; AI Across the Value Chain"}, {"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Digitalization; AI Across the Value Chain"}, {"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Digitalization; AI Across the Value Chain"}, {"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Digitalization; AI Across the Value Chain"}], "Audit_reviews": [{"url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "access": "full text reviewed", "title": "OMV AI Hub: 2,500 monthly GPT users, 10,000 trained", "publication_date": "Undated; see individual source review", "claims_checked": ["AI-084 source claim(s) checked against cited profile", "AI-132 source claim(s) checked against cited profile", "AI-133 source claim(s) checked against cited profile", "AI-134 source claim(s) checked against cited profile", "AI-135 source claim(s) checked against cited profile"], "notes": "OMV 2025 annual-report Digitalization section reviewed in full; AI Hub counts, Norwegian Companion, field-agent test, iDEA wording, and Romanian pump-ML pilot checked.", "partition": 2}]}, {"Source_ID": "R038", "Source_date": "2025-09", "Company": "Kuwait Oil Company", "Source_type": "Kuwait Times", "Title": "AI Innovation Center delivers G Agent drilling win", "URL": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/", "access": "full text reviewed", "title": "AI Innovation Center delivers G Agent drilling win", "publication_date": "2025-09", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R039", "Source_date": "2025-07", "Company": "Devon Energy", "Source_type": "American Oil & Gas Reporter", "Title": "ChatDVN internal AI platform plus AI-101 training", "URL": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "Locator": "Proprietary AI Platform section, productivity, compressor, simulfrac and video-development paragraphs", "Short_excerpt": "", "Verification": "Full trade-press interview rechecked 6 October 2026", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-10-06", "Research_reviews": [{"cohort": "north_america", "date": "2025-07", "status": "full text reviewed", "locator": "Interview with CTO Trey Lowe"}], "Audit_reviews": [{"url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "title": "ChatDVN and AI-101 workforce enablement", "publication_date": "2025-07", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["ChatDVN internal AI platform; leaders trained via AI-101; engineers, software developers and geoscientists report 15%-30% productivity improvement; humans review/approve AI recommendations; video analytics plans for 2026."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R040", "Source_date": "2026-02", "Company": "Devon Energy", "Source_type": "The Motley Fool (Q4 2025 call transcript)", "Title": "AI-enabled lift and maintenance drive $1B optimization plan", "URL": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"source_id": "R040", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "Devon Q4 2025 Earnings Call Transcript", "publisher": "Devon Energy / The Motley Fool transcription", "publication_date": "2026-02-18", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "Management Q&A, production and AI investment remarks", "claims_checked": ["Management says Devon trialed production-space gas-lift optimization and other projects, while investments in AI and platforms made during the year were beginning to come to fruition in production; projects trialed in 2025 were expected to scale during Q1/Q2 2026.", "The source does not expressly establish that the gas-lift trials themselves were AI-driven, and does not attribute the company’s broader optimization value targets exclusively to AI."], "notes": "Full transcript opened and direct AI passages checked. Secondary transcript; use to narrow language, not to merge gas-lift trials with AI without an explicit linkage.", "accessed": "2026-09-27"}]}, {"Source_ID": "R041", "Source_date": "2026-05", "Company": "Devon Energy", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "publication_date": "2026-05-06", "claims_checked": ["Transcript says Smart Gas Lift uses AI models and physics-based calculations to optimize gas injection on a closed loop directly to wells.", "Management says pilot in 2025 yielded about 2–3% uplift; by Q1 2026 full implementation in Delaware Basin covered >850 wells; current uplift was said to exceed pilot but not quantified; path to 1,500 wells described."], "notes": "Motley Fool earnings transcript is a secondary transcription of management remarks, not an issuer-published transcript. All operational metrics remain management-reported and secondary-transcribed."}]}, {"Source_ID": "R042", "Source_date": "2026-08", "Company": "Devon Energy", "Source_type": "The Motley Fool (Q2 2026 call transcript)", "Title": "AI speeds Coterra integration; in-house plus vendor model", "URL": "https://www.fool.com/earnings/call-transcripts/2026/08/11/devon-energy-dvn-q2-2026-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/08/11/devon-energy-dvn-q2-2026-earnings-call-transcript/", "access": "excerpt reviewed", "title": "AI speeds Coterra integration; in-house plus vendor model", "publication_date": "2026-08", "claims_checked": ["AI-087 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R043", "Source_date": "2026-05", "Company": "Devon Energy", "Source_type": "Drilling Contractor (IADC)", "Title": "AI agents for well design and completions sequencing", "URL": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-007", "source_id": "R043", "url": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873", "title": "Post-merger Devon sees big role for AI as it tackles integration, seeks $1 billion in synergies", "publication_date": "2026-07/08 issue (article page does not expose a day in the archive listing)", "access": "excerpt reviewed", "accessed": "2026-09-27", "claims_checked": ["The IADC Drilling Contractor July/August 2026 archive lists the article title and author Jessica Whiteside.", "The specific article URL returned a fetch/cache error; the listing does not expose the article body or substantiate the separate imported description 'AI agents for well design and completions sequencing'."], "notes": "Treat the cited source as not reverified for the detailed workflow, not as disproving the earlier Q4 earnings-presentation account. The archive article title concerns post-merger integration/synergies, so the source title/summary mismatch merits reconciliation before using it as evidence for well-design agents."}]}, {"Source_ID": "R045", "Source_date": "2025-11-27", "Company": "EOG Resources", "Source_type": "The Motley Fool (Q3 2025 call transcript)", "Title": "Generative AI across geology, drilling and safety", "URL": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-11-27", "status": "excerpt reviewed", "locator": "CEO/management remarks"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "title": "Generative AI for geology, drilling and safety", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Third-party transcript reproduces EOG comments on generative AI, geology, drilling, safety, predictive maintenance and in-house applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R046", "Source_date": "2026-02-25", "Company": "EOG Resources", "Source_type": "The Motley Fool (Q4 2025 call transcript)", "Title": "ML 'production optimizers' and 24-hour control room", "URL": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-02-25", "status": "excerpt reviewed", "locator": "Management remarks on production optimizers"}, {"cohort": "north_america", "date": "2026-02-25", "status": "excerpt reviewed", "locator": "Management remarks on production optimizers"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "title": "ML production optimizers and remote Utica operations", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Third-party transcript reproduces EOG claims on ML production optimizers and remote Utica operations."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R047", "Source_date": "2026-05", "Company": "Diamondback Energy", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "ML and AI in field ops cut downtime post-Endeavor", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "ML and AI in field ops cut downtime post-Endeavor", "publication_date": "2026-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R049", "Source_date": "2026", "Company": "ConocoPhillips", "Source_type": "ConocoPhillips (2026 Annual Meeting transcript)", "Title": "CEO: embedding AI into core processes at scale", "URL": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026", "status": "excerpt reviewed", "locator": "CEO answer on AI"}], "Audit_reviews": [{"url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "access": "full text reviewed", "title": "CEO: embedding AI into core processes at scale", "publication_date": "2026", "claims_checked": ["CEO described focus on embedding AI into core processes to reduce manual work, accelerate decisions and improve consistency under enterprise governance/security/responsible-use framework; no named application or outcome disclosed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R050", "Source_date": "2025-02-19", "Company": "Occidental Petroleum", "Source_type": "The Motley Fool (Q4 2024 call transcript)", "Title": "AI center of excellence and Gulf of America AI", "URL": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-02-19", "status": "full text reviewed", "locator": "CEO remarks on Gulf of America and AI center of excellence"}], "Audit_reviews": [{"source_id": "R050", "url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "access": "full text reviewed", "title": "Occidental Petroleum Q4 2024 Earnings Call Transcript", "publisher": "Occidental Petroleum / The Motley Fool transcription", "publication_date": "2025-02-19", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "CEO remarks on Gulf of America and AI center of excellence", "claims_checked": ["CEO says Gulf of America operations are using AI for supply-chain management, asset integrity, and reservoir characterization, and that Oxy created an AI center of excellence to align intercompany initiatives.", "Oxy says its large Algeria seismic acquisition would provide a rich dataset for future AI applications; this is future use, not a reported current AI result."], "notes": "Full transcript opened; specific AI paragraph reviewed. The transcript supports named use areas but provides no implementation design or AI-attributed outcome.", "accessed": "2026-09-27"}]}, {"Source_ID": "R051", "Source_date": "2025-08-07", "Company": "Occidental Petroleum", "Source_type": "The Motley Fool (Q2 2025 call transcript)", "Title": "AI subsurface project targets complex Gulf reservoirs", "URL": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-08-07", "status": "full text reviewed", "locator": "CEO remarks on subsurface AI"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "AI subsurface project targets complex Gulf reservoirs", "publication_date": "2025-08-07", "claims_checked": ["Oxy CEO said focused AI teams were working in Gulf of America, operations, and logistics/supply chain.", "Management said the complex Gulf subsurface project was still being developed and expected to be ready to start execution in the following one to two years; AI was already helping subsurface work in the Permian and other onshore areas."], "notes": "Motley Fool transcript is secondary. Keep Gulf project future/planning stage distinct from existing Permian/onshore assistance."}]}, {"Source_ID": "R052", "Source_date": "2026-05-06", "Company": "Occidental Petroleum", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "AI now part of workforce efficiency agenda", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-091", "source_id": "R052", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/", "title": "Occidental (OXY) Q1 2026 Earnings Transcript", "publication_date": "2026-05-06", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["The transcript attributes to COO Richard Jackson a statement that Oxy was reviewing workforce processes and priorities, including technologies such as AI.", "The transcript attributes to CEO Vicki Hollub the statement that AI was taking a larger role across company disciplines as part of workforce efficiency.", "The Q1 transcript does not substantiate the profile's $4 billion incremental cash-flow target by 2030; no AI-specific productivity or savings result is quantified."], "notes": "Motley Fool is a third-party earnings-call transcript, not an issuer-authored source. The corresponding Occidental-hosted transcript PDF was also located and reviewed at https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf (excerpt/full document text available); the management statements are consistent. Keep the source type as transcript and distinguish the speaker's remarks from independently measured outcomes."}]}, {"Source_ID": "R053", "Source_date": "2025-11-06", "Company": "Ovintiv", "Source_type": "The Motley Fool (Q3 2025 call transcript)", "Title": "In-house AI across production, drilling and completions", "URL": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-11-06", "status": "excerpt reviewed", "locator": "Management remarks"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "In-house AI across production, drilling and completions", "publication_date": "2025-11-06", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-06; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R054", "Source_date": "2026-02-24", "Company": "Ovintiv", "Source_type": "The Motley Fool (Q4 2025 call transcript)", "Title": "DRIVE Center AI optimizes rigs from day one", "URL": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-02-24", "status": "excerpt reviewed", "locator": "Management remarks"}, {"cohort": "north_america", "date": "2026-02-24", "status": "excerpt reviewed", "locator": "Management remarks"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "DRIVE Center AI optimizes rigs from day one", "publication_date": "2026-02-24", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed.", "Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-24; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R055", "Source_date": "2026-05-12", "Company": "Ovintiv", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "AI tunes 70-odd frac design inputs in real time", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-05-12", "status": "excerpt reviewed", "locator": "Management remarks"}, {"cohort": "north_america", "date": "2026-05-12", "status": "excerpt reviewed", "locator": "Management remarks"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "AI tunes 70-odd frac design inputs in real time", "publication_date": "2026-05-12", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed.", "Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-12; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R056", "Source_date": "2026-02-18", "Company": "EQT", "Source_type": "The Motley Fool (Q4 2025 call transcript)", "Title": "AI tools to target service rebids and procurement", "URL": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-02-18", "status": "excerpt reviewed", "locator": "management remarks on service rebids/procurement"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "AI tools to target service rebids and procurement", "publication_date": "2026-02-18", "claims_checked": ["Management said AI tools would help target service rebids and procurement, expecting low-single-digit procurement cost reductions; not isolated to AI and no method, baseline or achieved value disclosed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-18; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R057", "Source_date": "2026-02", "Company": "Matador Resources", "Source_type": "The Motley Fool (Q4 2025 call transcript)", "Title": "Vendor-led AI: 'fast follower' approach", "URL": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "title": "Matador Q4 2025 Earnings Call", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["CEO describes vendor-supported, cautious “fast follower” approach, without naming vendors or applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R058", "Source_date": "2026-05-07", "Company": "Matador Resources", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "AI committee; AI used for lateral targeting", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-05-07", "status": "full text reviewed", "locator": "Q&A, management response on AI adoption and operations"}], "Audit_reviews": [{"source_id": "R058", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "Matador Q1 2026 Earnings Call Transcript", "publisher": "Matador Resources / The Motley Fool transcription", "publication_date": "2026-05-07", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "Management Q&A on AI operations, production control room, completions, and lateral placement", "claims_checked": ["Management describes AI-driven analytics across operations; a control room and field staff monitor more than 40 million daily data points in real time to identify downtime and inefficiencies.", "Management says real-time frac monitoring covers water/sand pressures, volumes, and logistics; drilling AI helps target the preferred portion of the lateral target.", "The same answer describes a committee-led cautious adoption approach. Operational gains cited elsewhere in the call are not assigned exclusively to AI."], "notes": "Full transcript opened; Q&A detail reviewed. This materially expands generic “cautious integration” wording; no model vendor, model architecture, or AI-only economic result is disclosed.", "accessed": "2026-09-27"}]}, {"Source_ID": "R059", "Source_date": "2026-05-06", "Company": "Chord Energy", "Source_type": "The Motley Fool (Q1 2026 call transcript)", "Title": "AI rod-pump optimization arrests base decline", "URL": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-05-06", "status": "excerpt reviewed", "locator": "Management remarks on rod pumps"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "AI rod-pump optimization arrests base decline", "publication_date": "2026-05-06", "claims_checked": ["Transcript attributes improved legacy-well base production to AI pump tuning along with deeper pump placement and workovers; no AI-only effect is isolated."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-06; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R060", "Source_date": "2026-08", "Company": "Chord Energy", "Source_type": "The Motley Fool (Q2 2026 call transcript)", "Title": "AI rod-pump program deployed field-wide", "URL": "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "AI rod-pump program deployed field-wide", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R061", "Source_date": "2026-03", "Company": "Canadian Natural Resources", "Source_type": "SEC (CNRL Form 40-F / 2025 AIF)", "Title": "Formal AI policy: no AI in production-critical processes", "URL": "https://www.sec.gov/Archives/edgar/data/1017413/000101741326000018/cnq-20251231.htm", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.sec.gov/Archives/edgar/data/1017413/000101741326000018/cnq-20251231.htm", "access": "inaccessible", "title": "Formal AI policy: no AI in production-critical processes", "publication_date": "2026-03", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R062", "Source_date": "2025-07", "Company": "Suncor Energy", "Source_type": "International Mining", "Title": "Autonomous haul fleet nears 140 trucks", "URL": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/", "access": "full text reviewed", "title": "Autonomous haul fleet nears 140 trucks", "publication_date": "2025-07-04", "claims_checked": ["Article says nearly 120 autonomous haul trucks at Suncor Base Plant, with all ore moved by driverless autonomous-haulage systems; operation began in 2018.", "Describes 15 trucks added at Millennium in May 2024, reaching 91 by year-end; further 2025 deployment, control-room oversight, obstacle/collision detection and workflow benefits.", "The article does not establish that AI/ML powers the autonomous haulage system and does not support a 140-truck planned target."], "notes": "International Mining is secondary. Autonomy is confirmed; AI basis and any 140-truck target require separate primary evidence."}]}, {"Source_ID": "R063", "Source_date": "2026-02-04", "Company": "Phillips 66", "Source_type": "Phillips 66 IR (LSEG Q4 2025 transcript)", "Title": "Machine learning shortens refinery turnarounds", "URL": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026-02-04", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "title": "Machine learning in refinery turnaround planning", "publication_date": "2026-02-04", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes machine learning tools in turnaround planning, moving work outside turnaround to shorten duration/financial impact; company says forecast/spend improved, does not isolate AI causation or savings."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R064", "Source_date": "2026-08", "Company": "Phillips 66", "Source_type": "Phillips 66 IR (S&P Global Q2 2026 transcript)", "Title": "Company AI program 'normalizing' AI for front line", "URL": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026-08", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "title": "Company AI program and frontline solution replication", "publication_date": "2026-08", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes broad AI program, copying use cases, frontline engineers extracting data and deploying solutions more quickly; leadership says ML long used for maintenance/turnarounds."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R065", "Source_date": "2025-10-23", "Company": "Valero Energy", "Source_type": "The Motley Fool (Q3 2025 call transcript)", "Title": "'Cautiously optimistic' on AI; data quality first", "URL": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-10-23", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "access": "excerpt reviewed", "title": "'Cautiously optimistic' on AI; data quality first", "publication_date": "2025-10-23", "claims_checked": ["AI-102 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R066", "Source_date": "Undated; refers to 2024 AFPM Summit", "Company": "Marathon Petroleum", "Source_type": "Imubit", "Title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "URL": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes", "access": "excerpt reviewed", "title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "publication_date": "Undated; refers to 2024 AFPM Summit", "claims_checked": ["The undated Imubit page attributes the 25% drum-cycle result to Marathon at the 2024 AFPM Summit. It does not support the separate Flyscan investment claim; removed that claim."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R067", "Source_date": "2025-08", "Company": "Williams", "Source_type": "The Motley Fool (Q2 2025 call transcript)", "Title": "New CIO; AI model out-trades 10-year gas trader", "URL": "https://www.fool.com/earnings/call-transcripts/2025/08/05/williams-wmb-q2-2025-earnings-call-transcript/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2025/08/05/williams-wmb-q2-2025-earnings-call-transcript/", "access": "inaccessible", "title": "New CIO; AI model out-trades 10-year gas trader", "publication_date": "2025-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R068", "Source_date": "2025-01-14", "Company": "Kinder Morgan", "Source_type": "PR Newswire", "Title": "Strategic investment in Flyscan AI leak detection", "URL": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-104", "source_id": "R068", "url": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html", "title": "Flyscan Systems Announces that Kinder Morgan Joins its Strategic Investor Group alongside Marathon Petroleum, Enbridge and Hatch", "publication_date": "2025-01-14", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Flyscan says Kinder Morgan joined its strategic investor group; the funding was to accelerate commercialization and entry into natural-gas detection/quantification.", "Flyscan describes remote sensing, AI, liquid-hydrocarbon detection, and automated right-of-way visual inspection from routine patrol aircraft. KMI's COO endorses the technology and says the parties will partner.", "The announcement does not establish a KMI operational rollout or quantified leak-detection result."], "notes": "Issuer Flyscan announcement distributed by PR Newswire; Flyscan's official newsroom copy was also reviewed. The date in the baseline initiative's Source_date (2025-04-16) is the date of KMI's Q1 earnings release, not the cited Flyscan announcement; correct the initiative source date to 2025-01-14."}]}, {"Source_ID": "R069", "Source_date": "2025-06", "Company": "SLB", "Source_type": "SLB", "Title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "URL": "https://www.slb.com/news-and-insights/newsroom/updates/2025/slb%E2%80%99s-flagship-digital-platform-to-be-available-on-new-mistral-compute-ai-infrastructure", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.slb.com/news-and-insights/newsroom/updates/2025/slb%E2%80%99s-flagship-digital-platform-to-be-available-on-new-mistral-compute-ai-infrastructure", "access": "full text reviewed", "title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "publication_date": "2025-06", "claims_checked": ["AI-009 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R070", "Source_date": "2025-11", "Company": "SLB", "Source_type": "Business Wire", "Title": "Tela agentic AI assistant launched for upstream", "URL": "https://www.businesswire.com/news/home/20251103326704/en/SLB-Unveils-Groundbreaking-New-Agentic-AI-Technology-for-the-Energy-Industry", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.businesswire.com/news/home/20251103326704/en/SLB-Unveils-Groundbreaking-New-Agentic-AI-Technology-for-the-Energy-Industry", "access": "full text reviewed", "title": "Tela agentic AI assistant launched for upstream", "publication_date": "2025-11", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R071", "Source_date": "2026-03", "Company": "SLB", "Source_type": "energy-pedia", "Title": "SLB and NVIDIA build 'AI Factory for Energy'", "URL": "https://www.energy-pedia.com/news/general/slb-industrializes-ai-for-the-energy-industry-with-nvidia-203316", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.energy-pedia.com/news/general/slb-industrializes-ai-for-the-energy-industry-with-nvidia-203316", "access": "full text reviewed", "title": "SLB and NVIDIA build 'AI Factory for Energy'", "publication_date": "2026-03", "claims_checked": ["AI-105 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R072", "Source_date": "2026-06", "Company": "SLB", "Source_type": "Business Wire", "Title": "SLB Digital Marketplace lists ~200 AI and digital products", "URL": "https://www.businesswire.com/news/home/20260615078195/en/SLB-Launches-Digital-Marketplace-to-Scale-AI-and-Digital-Innovation-Across-Energy", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.businesswire.com/news/home/20260615078195/en/SLB-Launches-Digital-Marketplace-to-Scale-AI-and-Digital-Innovation-Across-Energy", "access": "full text reviewed", "title": "SLB Digital Marketplace lists ~200 AI and digital products", "publication_date": "2026-06", "claims_checked": ["AI-047 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R073", "Source_date": "2026-05", "Company": "SLB / Tachyus", "Source_type": "JPT (SPE)", "Title": "SLB acquires Tachyus physics-ML reservoir startup", "URL": "https://jpt.spe.org/slb-expands-digital-portfolio-with-acquisition-of-ai-firm-tachyus", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://jpt.spe.org/slb-expands-digital-portfolio-with-acquisition-of-ai-firm-tachyus", "access": "full text reviewed", "title": "SLB acquires Tachyus physics-ML reservoir startup", "publication_date": "2026-05", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R074", "Source_date": "2026-04", "Company": "SLB / S&P Global Energy", "Source_type": "Orphan & Idle Wells (SLB release summary)", "Title": "SLB buys Kingdom, Petra upstream software for agentic AI", "URL": "https://www.orphan-idle-wells.com/news/slb-acquires-the-software-running-us-oil-fields", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.orphan-idle-wells.com/news/slb-acquires-the-software-running-us-oil-fields", "access": "full text reviewed", "title": "SLB buys Kingdom, Petra upstream software for agentic AI", "publication_date": "2026-04", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R075", "Source_date": "2025-11", "Company": "Microsoft", "Source_type": "ADNOC", "Title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "access": "full text reviewed", "title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "publication_date": "2025-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R076", "Source_date": "2025-01", "Company": "Halliburton", "Source_type": "World Oil", "Title": "Octiv Auto Frac: first fully automated frac with Coterra", "URL": "https://worldoil.com/news/2025/1/6/halliburton-coterra-launch-first-fully-automated-hydraulic-fracturing-program/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://worldoil.com/news/2025/1/6/halliburton-coterra-launch-first-fully-automated-hydraulic-fracturing-program/", "access": "full text reviewed", "title": "Octiv Auto Frac: first fully automated frac with Coterra", "publication_date": "2025-01", "claims_checked": ["AI-109 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R077", "Source_date": "2025-06", "Company": "Halliburton", "Source_type": "Chevron", "Title": "ZEUS IQ closed-loop intelligent frac with Chevron", "URL": "https://www.chevron.com/newsroom/2025/q2/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.chevron.com/newsroom/2025/q2/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "access": "full text reviewed", "title": "ZEUS IQ closed-loop intelligent frac with Chevron", "publication_date": "2025-06", "claims_checked": ["AI-110 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R078", "Source_date": "2026-04", "Company": "Halliburton / Sekal", "Source_type": "Yahoo Finance / Zacks", "Title": "Halliburton acquires Sekal drilling-automation software", "URL": "https://finance.yahoo.com/sectors/energy/articles/halliburton-strengthens-automation-strategic-sekal-133400448.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://finance.yahoo.com/sectors/energy/articles/halliburton-strengthens-automation-strategic-sekal-133400448.html", "access": "full text reviewed", "title": "Halliburton acquires Sekal drilling-automation software", "publication_date": "2026-04", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R079", "Source_date": "2025-05", "Company": "Baker Hughes / C3 AI", "Source_type": "C3 AI", "Title": "BakerHughesC3.ai JV renewed through 2028", "URL": "https://c3.ai/c3-ai-and-baker-hughes-renew-and-expand-joint-venture-agreement/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"source_id": "R079", "url": "https://c3.ai/c3-ai-and-baker-hughes-renew-and-expand-joint-venture-agreement/", "access": "full text reviewed", "title": "C3 AI and Baker Hughes Renew and Expand Joint Venture Agreement", "publisher": "C3 AI", "publication_date": "2025-05-28", "source_type": "Company announcement", "locator": "Renewal term, deployed customers, Baker Hughes internal-use applications", "claims_checked": ["C3 AI and Baker Hughes renewed and expanded the JV through June 2028 to develop, deliver, and market AI solutions to oil/gas and chemicals.", "C3 AI lists deployed solutions at Shell, Eni, QatarEnergy LNG, Petronas, ExxonMobil, LyondellBasell, and Flint Hills Resources; it describes customer benefits generally, without customer-specific metrics.", "Baker Hughes plans to expand internal use of sourcing optimization, inventory optimization, and sustainability-suite applications."], "notes": "Full issuer announcement opened. Deployment/customer and economic claims are vendor-reported and aggregate; do not overstate them as independently confirmed site-level outcomes.", "accessed": "2026-09-27"}]}, {"Source_ID": "R080", "Source_date": "2025-02", "Company": "Baker Hughes", "Source_type": "Baker Hughes", "Title": "Cordant built on Azure AI Foundry with Microsoft", "URL": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial", "access": "inaccessible", "title": "Cordant built on Azure AI Foundry with Microsoft", "publication_date": "2025-02 (per source metadata)", "claims_checked": ["Attempted Baker Hughes announcement URL; web fetch returned HTTP 403."], "notes": "Direct vendor source not accessible during this pass; claims remain unverified here, not disproved."}]}, {"Source_ID": "R081", "Source_date": "2025-06-10", "Company": "Baker Hughes", "Source_type": "Baker Hughes", "Title": "Leucipa adds genAI assistant, AI lift optimizer for Repsol", "URL": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-06-10", "status": "full text reviewed", "locator": "Leucipa assistant and lift optimizer"}], "Audit_reviews": [{"url": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "title": "Leucipa GenAI assistant and lift optimizer for Repsol", "publication_date": "2025-06-10", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R082", "Source_date": "2025-10", "Company": "Weatherford", "Source_type": "GlobeNewswire", "Title": "Weatherford Industrial Intelligence portfolio launched", "URL": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-113", "source_id": "R082", "url": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html", "title": "Weatherford Launches Industrial Intelligence at FWRD 2025", "publication_date": "2025-10-01", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["The supplied GlobeNewswire URL could not be fetched in this pass, but Weatherford's investor-relations copy of the same release was opened and reviewed.", "The issuer release describes a product portfolio and names AI/physics-based MetaFlow flow estimation, AI tools for document digitization/transcription/image recognition/translation in Engineer's Briefcase, and other data/edge components.", "The launch announcement describes offered products, not a customer's deployment or measured results."], "notes": "A broken syndication URL was replaced with the issuer-hosted copy at https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126. Do not treat the release date as a deployment date."}]}, {"Source_ID": "R083", "Source_date": "2026-01-12", "Company": "NOV", "Source_type": "NOV", "Title": "NOV and AIQ pilot AI drilling on Max platform", "URL": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-01-12", "status": "full text reviewed", "locator": "Collaboration scope"}], "Audit_reviews": [{"url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "access": "full text reviewed", "title": "NOV and AIQ pilot AI drilling on Max platform", "publication_date": "2026-01-12", "claims_checked": ["AI-114 source claim(s) checked against cited profile"], "notes": "NOV/AIQ AI-enabled drilling pilot release opened; pilot/partner scope checked, no commercial deployment claim inferred.", "partition": 2}]}, {"Source_ID": "R084", "Source_date": "2025-04", "Company": "Nabors / Corva", "Source_type": "World Oil", "Title": "RigCLOUD Powered by Corva drilling intelligence", "URL": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"source_id": "R084", "url": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/", "access": "full text reviewed", "title": "Nabors, Corva AI Join to Advance Digital Transformation in Drilling Industry", "publisher": "World Oil (Nabors/Corva announcement coverage)", "publication_date": "2025-04-11", "source_type": "Industry trade publication", "locator": "RigCLOUD Powered by Corva product and integration description", "claims_checked": ["The announcement says RigCLOUD Powered by Corva would combine Nabors edge/cloud computing and rig instrumentation/apps with Corva AI analytics for real-time data processing, predictive insight, and drilling performance optimization.", "The release language describes a solution set to be integrated and capabilities intended for customers; this story does not establish fleetwide deployment or measured field results."], "notes": "Full article opened. Trade source republishes/relays vendor announcements; treat deployment and expected-benefit wording as partner-reported and planned.", "accessed": "2026-09-27"}]}, {"Source_ID": "R085", "Source_date": "2026-02", "Company": "Helmerich & Payne", "Source_type": "Business Wire", "Title": "FlexRobotics rig-floor robotics deployed for supermajor", "URL": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results", "access": "full text reviewed", "title": "FlexRobotics rig-floor robotics deployed for supermajor", "publication_date": "2026-02-04 (quarter ended 2025-12-31)", "claims_checked": ["H&P reported FlexRobotics deployed on a Permian rig for a supermajor and supporting multiple pads; described reliable rig-floor performance and customer interest in additional units.", "The earnings release does not say that the deployed robotic equipment itself uses AI. H&P’s product page separately describes a camera-based AI zone-monitoring add-on; the cited earnings release does not establish that this add-on was deployed with the customer."], "notes": "Business Wire is the issuer-distributed earnings release. Separately reopened H&P product page: it calls FlexRobotics rig-floor automation and lists AI-enabled zone monitoring under an add-on camera-support feature."}]}, {"Source_ID": "R086", "Source_date": "2025-07-30", "Company": "Patterson-UTI", "Source_type": "SEC (Patterson-UTI 8-K)", "Title": "Cortex ML autodriller and REX monitoring drive rig margins", "URL": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-07-30", "status": "full text reviewed", "locator": "Drilling services remarks"}], "Audit_reviews": [{"url": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "access": "full text reviewed", "title": "Cortex ML autodriller and REX monitoring drive rig margins", "publication_date": "2025-07-30", "claims_checked": ["Cortex platform enables a machine-learning autodriller; REX is a cloud-based early-alert field monitoring system.", "Company cites increased demand and revenue from drilling automation technologies, but no separate AI return metric."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-07-30; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R087", "Source_date": "2025-10-16", "Company": "Liberty Energy", "Source_type": "Liberty Energy", "Title": "Liberty launches Forge LLM for frac asset orchestration", "URL": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-10-16", "status": "full text reviewed", "locator": "Forge and StimCommander paragraphs"}], "Audit_reviews": [{"url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "access": "full text reviewed", "title": "Liberty launches Forge LLM for frac asset orchestration", "publication_date": "2025-10-16", "claims_checked": ["AI-118 source claim(s) checked against cited profile"], "notes": "Liberty Q3 2025 release opened; company-wide StimCommander outcomes are not isolated to LLM contribution.", "partition": 2}]}, {"Source_ID": "R088", "Source_date": "2024-09-25", "Company": "Palantir", "Source_type": "StockTitan (Business Wire release)", "Title": "APA extends multi-year Palantir deal, adds AIP", "URL": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html", "access": "excerpt reviewed", "title": "APA extends multi-year Palantir deal, adds AIP", "publication_date": "2024-09-25", "claims_checked": ["Issuer release reproduced by StockTitan is September 25, 2024, one day before the window. Retained as historical context, not a current-period new initiative."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R089", "Source_date": "2024-08-07", "Company": "AWS", "Source_type": "AWS for Industries blog", "Title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "URL": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/", "access": "excerpt reviewed", "title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "publication_date": "2024-08-07", "claims_checked": ["AWS article is August 7, 2024. The 30% bpx production ambition is not an Xpert result; removed it from the metric field."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R090", "Source_date": "2026-01-27", "Company": "Snowflake", "Source_type": "Snowflake", "Title": "Snowflake launches Energy Solutions with 30+ partner apps", "URL": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-01-27", "status": "full text reviewed", "locator": "Energy Solutions, partner ecosystem, SAP and spatial data"}], "Audit_reviews": [{"url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "access": "full text reviewed", "title": "Snowflake launches Energy Solutions with 30+ partner apps", "publication_date": "2026-01-27", "claims_checked": ["AI-119 source claim(s) checked against cited profile"], "notes": "Snowflake Energy Solutions announcement opened; ecosystem product/partner offering, not operator adoption.", "partition": 2}]}, {"Source_ID": "R091", "Source_date": "2024-04-09", "Company": "Databricks", "Source_type": "PR Newswire", "Title": "Data Intelligence Platform for Energy with genAI", "URL": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html", "access": "excerpt reviewed", "title": "Data Intelligence Platform for Energy with genAI", "publication_date": "2024-04-09", "claims_checked": ["Databricks issuer release is April 9, 2024. Its organization count is company-wide, not energy adoption; excluded from initiative performance metrics."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R092", "Source_date": "2025-09", "Company": "Cognite", "Source_type": "World Oil", "Title": "Aker BP expands AI-first strategy with Cognite Atlas AI", "URL": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"baseline_id": "AI-049", "source_id": "R092", "url": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/", "title": "Aker BP expands AI-first strategy with Cognite Atlas AI to transform E&P efficiency", "publication_date": "2025-09-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["World Oil describes expansion of Aker BP's Cognite Atlas AI use, prebuilt agents/low-code tooling, document processing, barrier management and equipment registration.", "It attributes to Aker BP a claim that AI-agent root-cause analysis reduced engineering time by more than 70%; the article supplies no sample/baseline, and that result remains company/partner reported.", "The baseline also links Cognite's partner announcement, which is the more direct source for the initiative."], "notes": "Secondary trade-publication coverage, not issuer-primary. Keep the 70% result explicitly attributed and unbenchmarked; the added source does not independently validate it."}]}, {"Source_ID": "R093", "Source_date": "2025-09", "Company": "Cognite", "Source_type": "Business Wire", "Title": "Cognite Atlas AI major release, low-code agent workbench", "URL": "https://www.businesswire.com/news/home/20250909488860/en/Cognite-Atlas-AI-Drives-Customer-Momentum-with-New-Major-Release-Accelerates-Adoption-of-Agentic-AI-to-Unlock-More-Industrial-Value", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.businesswire.com/news/home/20250909488860/en/Cognite-Atlas-AI-Drives-Customer-Momentum-with-New-Major-Release-Accelerates-Adoption-of-Agentic-AI-to-Unlock-More-Industrial-Value", "access": "full text reviewed", "title": "Cognite Atlas AI major release, low-code agent workbench", "publication_date": "2025-09", "claims_checked": ["AI-120 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R094", "Source_date": "2025-10-14", "Company": "Cognite", "Source_type": "Cognite", "Title": "Cognite integrates NVIDIA accelerated computing for agents", "URL": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-10-14", "status": "full text reviewed", "locator": "Celanese time-series anomaly detection"}], "Audit_reviews": [{"url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "access": "full text reviewed", "title": "Cognite integrates NVIDIA accelerated computing for agents", "publication_date": "2025-10-14", "claims_checked": ["AI-121 source claim(s) checked against cited profile"], "notes": "Cognite/NVIDIA Celanese release opened; chemical-plant anomaly detection is prospective/product collaboration; not an O&G deployment.", "partition": 2}]}, {"Source_ID": "R095", "Source_date": "2025-06-25", "Company": "Novi Labs", "Source_type": "Novi Labs", "Title": "Novi Labs raises $35M for AI-first energy analytics", "URL": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "Locator": "Linked announcement / report text", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Reviewed during consolidation; metrics remain company or vendor reported.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-06-25", "status": "full text reviewed", "locator": "Funding and platform overview"}], "Audit_reviews": [{"url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "access": "full text reviewed", "title": "Novi Labs raises $35M for AI-first energy analytics", "publication_date": "2025-06-25", "claims_checked": ["Invictus Growth Partners invested $35 million to support Novi product innovation, expansion and acquisitions.", "Novi describes platform combining machine learning, proprietary datasets and reservoir science for upstream capital decisions."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-06-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R096", "Source_date": "2026-05-28", "Company": "Novi Labs", "Source_type": "Novi Labs", "Title": "Novi acquires Turing Analytics and RBN Energy", "URL": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-05-28", "status": "full text reviewed", "locator": "Combination and market workflows"}], "Audit_reviews": [{"url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "access": "full text reviewed", "title": "Novi acquires Turing Analytics and RBN Energy", "publication_date": "2026-05-28", "claims_checked": ["RBN acquisition combines Novi upstream AI forecasts with commodity, midstream and demand intelligence.", "Decision context includes basis, transport capacity, commodity flows and end-market pricing."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "R097", "Source_date": "2024-02-20", "Company": "Enverus", "Source_type": "Enverus", "Title": "Instant Analyst genAI with cited answers", "URL": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/", "access": "excerpt reviewed", "title": "Instant Analyst genAI with cited answers", "publication_date": "2024-02-20", "claims_checked": ["Enverus dates the launch February 20, 2024. Later product extensions require their own disclosure evidence."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R099", "Source_date": "2025-12", "Company": "Enverus", "Source_type": "Enverus", "Title": "Enverus AI launched as 'AI operating system for energy'", "URL": "https://www.enverus.com/newsroom/enverus-ai-powering-the-next-era-of-energy-intelligence/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.enverus.com/newsroom/enverus-ai-powering-the-next-era-of-energy-intelligence/", "access": "full text reviewed", "title": "Enverus AI launched as 'AI operating system for energy'", "publication_date": "2025-12", "claims_checked": ["AI-065 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R101", "Source_date": "2025-04-15", "Company": "Collide", "Source_type": "Business Wire", "Title": "Collide raises $5M seed for O&G genAI platform", "URL": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-04-15", "status": "full text reviewed", "locator": "Seed funding and energy knowledge platform"}], "Audit_reviews": [{"url": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "title": "Collide $5m seed and energy-specific GenAI knowledge platform", "publication_date": "2025-04-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Collide announced a $5m seed led by Mercury Fund; proceeds support platform and engineering expansion.", "Platform uses retrieval augmented generation and LLMs over curated technical papers, internal processes and Q&A to provide cited, domain-specific energy answers.", "Collide reported 6,000+ community users and estimated 25%-33% time savings for petroleum engineers; these are company estimates and community users are not equivalent to deployed enterprise users."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R102", "Source_date": "2026-05", "Company": "Collide", "Source_type": "Energy Capital HTX", "Title": "RIGGS energy LLM automates RRC filings", "URL": "https://energycapitalhtx.com/collide-riggs-ai-platform", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"source_id": "R102", "url": "https://energycapitalhtx.com/collide-riggs-ai-platform", "access": "full text reviewed", "title": "Houston AI Startup Rolls Out Platform to Reshape Oil and Gas Workflows", "publisher": "Energy Capital HTX", "publication_date": "2026-05-07", "source_type": "Secondary interview/article", "locator": "RIGGS product, Winn Resources RRC filings workflow", "claims_checked": ["Article reports that Collide and Winn Resources collaborated on monthly Texas Railroad Commission W-10/G-10 filings; Collide reported processing over 95% faster, with an example of 50 wells completed in 20 minutes.", "Winn’s operations director is quoted describing a human-in-the-loop workflow; article says reservoir analysis, production optimization, and workover report generation were future/near-term targets, not yet documented deployments."], "notes": "Full article opened. Implementation/result is attributed to Collide and Winn through the article; no primary Winn release was part of this 10-source supplemental assignment.", "accessed": "2026-09-27"}]}, {"Source_ID": "R103", "Source_date": "2026-09-08", "Company": "Collide", "Source_type": "GlobeNewswire", "Title": "Battalion Oil makes strategic investment in Collide", "URL": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-09-08", "status": "full text reviewed", "locator": "Strategic investment, intended deployment and Riggs"}], "Audit_reviews": [{"url": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "title": "Battalion strategic investment and planned Collide Operations System rollout", "publication_date": "2026-09-08", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Battalion made strategic equity investment and will deploy Collide Operations System under subscription.", "Initial plan unites more than 100 TB of records, contracts and production history; planned use cases include gas reconciliation, field invoices, power, and failure diagnosis.", "Riggs is expected to be available to production engineers around day 90; human review/control retained."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R104", "Source_date": "2026-07", "Company": "Applied Computing", "Source_type": "TechCrunch", "Title": "$20M Series A for Orbital O&G foundation model", "URL": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/", "access": "full text reviewed", "title": "$20M Series A for Orbital O&G foundation model", "publication_date": "2026-07", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}]}, {"Source_ID": "R105", "Source_date": "2026-03", "Company": "Seeq", "Source_type": "PR Newswire", "Title": "Seeq Intelligence with Agent Q and custom agents", "URL": "https://www.prnewswire.com/news-releases/introducing-seeq-intelligence-bridging-industrial-ai-and-human-expertise-for-smarter-operational-decisions-302702464.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.prnewswire.com/news-releases/introducing-seeq-intelligence-bridging-industrial-ai-and-human-expertise-for-smarter-operational-decisions-302702464.html", "access": "full text reviewed", "title": "Seeq Intelligence with Agent Q and custom agents", "publication_date": "2026-03", "claims_checked": ["AI-128 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R106", "Source_date": "2026-05-20", "Company": "AVEVA", "Source_type": "AVEVA", "Title": "AVEVA embeds agentic AI, MCP across CONNECT", "URL": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-05-20", "status": "full text reviewed", "locator": "CONNECT, Flows and AI roadmap"}], "Audit_reviews": [{"url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "access": "full text reviewed", "title": "AVEVA embeds agentic AI, MCP across CONNECT", "publication_date": "2026-05-20", "claims_checked": ["AI-129 source claim(s) checked against cited profile"], "notes": "AVEVA AI roadmap announcement opened; future Q1 2027 capabilities identified as planned, not shipped at source date.", "partition": 2}]}, {"Source_ID": "R107", "Source_date": "2025-03", "Company": "Emerson / AspenTech", "Source_type": "Emerson", "Title": "Emerson completes AspenTech buyout for industrial AI", "URL": "https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "title": "Emerson completes AspenTech buyout for industrial AI", "publication_date": "2025-03", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R108", "Source_date": "2026-08", "Company": "McKinsey & Company", "Source_type": "Oil & Gas Journal", "Title": "AI could unlock $230B a year in upstream", "URL": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value", "access": "excerpt reviewed", "title": "AI could unlock $230B a year in upstream", "publication_date": "2026-08", "claims_checked": ["Oil & Gas Journal reports McKinsey scenarios on August 26, 2026. They are modeled industry opportunities, not realized AI returns; underlying McKinsey model was not independently audited."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R109", "Source_date": "Undated source; 2026–2030 forecast period", "Company": "Rystad Energy", "Source_type": "Rystad Energy", "Title": "Digital and AI: a $500B upstream opportunity", "URL": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas", "access": "excerpt reviewed", "title": "Digital and AI: a $500B upstream opportunity", "publication_date": "Undated source; 2026–2030 forecast period", "claims_checked": ["Rystad describes four workflow categories for the $500 billion estimate. The earlier five-area wording came from a separate digitalization article and was removed; publication month is not visible in the opened page."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R110", "Source_date": "2025-12", "Company": "Rystad Energy", "Source_type": "World Oil", "Title": "Digitalization could save $320B by 2030", "URL": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/", "access": "excerpt reviewed", "title": "Digitalization could save $320B by 2030", "publication_date": "2025-12", "claims_checked": ["World Oil attributes the December 5, 2025 estimate to Rystad. It is a digitalization opportunity including AI, not an AI-only savings result."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R111", "Source_date": "2025-08", "Company": "Boston Consulting Group", "Source_type": "BCG", "Title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "URL": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf", "access": "excerpt reviewed", "title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "publication_date": "2025-08", "claims_checked": ["BCG PDF executive summary presents 30–70% potential EBIT improvement. It is a scenario, not achieved industry profit; other chart statistics require a separate detailed methodology review."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R112", "Source_date": "2026-03-19", "Company": "Bain & Company", "Source_type": "Bain & Company", "Title": "Energy execs: fewer than 20% scaling AI with results", "URL": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/", "access": "excerpt reviewed", "title": "Energy execs: fewer than 20% scaling AI with results", "publication_date": "2026-03-19", "claims_checked": ["Bain publication is March 19, 2026, not February. Its headline says fewer than 20% scaling with results while body says a quarter scaling applications/reimagining functions; preserve the source distinction and mixed-sector survey population."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R113", "Source_date": "2025-03", "Company": "EY", "Source_type": "EY", "Title": "O&G execs: 85% say reskilling decides success; 29% do it", "URL": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey", "access": "excerpt reviewed", "title": "O&G execs: 85% say reskilling decides success; 29% do it", "publication_date": "2025-03", "claims_checked": ["EY groups oil and gas AND chemicals executives together for the 85%/29% findings. Earlier text incorrectly treated these as oil-and-gas-only statistics."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R114", "Source_date": "2026-04-13", "Company": "EY", "Source_type": "EY", "Title": "Energy leaders struggle to link AI to productivity", "URL": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption", "access": "excerpt reviewed", "title": "Energy leaders struggle to link AI to productivity", "publication_date": "2026-04-13", "claims_checked": ["EY article is April 13, 2026; December 2025 is the survey wave. The statistics describe energy respondents and should not be presented as an oil-and-gas-only sample."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R115", "Source_date": "2025-01", "Company": "EY", "Source_type": "EY", "Title": "96% of O&G execs cite data and vendor challenges", "URL": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company", "access": "excerpt reviewed", "title": "96% of O&G execs cite data and vendor challenges", "publication_date": "2025-01", "claims_checked": ["EY January 6, 2025 article supports the 96% data/access/silo challenge figures. Survey methodology was not fully reviewed; this is context, not a deployment."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R116", "Source_date": "2025-10", "Company": "Deloitte", "Source_type": "Deloitte Insights", "Title": "AI heading past 50% of O&G IT spend by 2029", "URL": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html", "access": "excerpt reviewed", "title": "AI heading past 50% of O&G IT spend by 2029", "publication_date": "2025-10", "claims_checked": ["Deloitte presents IT-spending forecasts and selected maintenance examples. Spending share is a projection; isolated vendor-derived maintenance results are not typical industry outcomes."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R117", "Source_date": "2025-05", "Company": "Deloitte", "Source_type": "Deloitte via PR Newswire", "Title": "1.2M US energy and chemicals workers need upskilling", "URL": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html", "access": "excerpt reviewed", "title": "1.2M US energy and chemicals workers need upskilling", "publication_date": "2025-05", "claims_checked": ["Deloitte describes 36% of task hours as human-in-the-loop: 26% mechanical/electronic automation plus 10% generative/agentic AI. It is incorrect to imply one-third of all tasks are agentic-AI-led."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R118", "Source_date": "2025-11", "Company": "KPMG", "Source_type": "KPMG", "Title": "Energy CEOs: only 40% actively reskilling for AI", "URL": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html", "access": "excerpt reviewed", "title": "Energy CEOs: only 40% actively reskilling for AI", "publication_date": "2025-11", "claims_checked": ["KPMG landing-page figures are 40%, 80% and 82% for the energy, natural resources and chemicals CEO cohort, not exclusively oil and gas."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R119", "Source_date": "2024-10", "Company": "DNV", "Source_type": "DNV", "Title": "Only 12% of O&G firms use AI in live operations", "URL": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "access": "excerpt reviewed", "title": "Only 12% of O&G firms use AI in live operations", "publication_date": "2024-10", "claims_checked": ["DNV table percentages are NET agreement (agree minus disagree). Its prose also distinguishes 12% live use and 3% advanced use. Removed simple adoption-percentage interpretation of the net-agreement table."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R120", "Source_date": "2025-09", "Company": "DNV", "Source_type": "Oilfield Technology", "Title": "60% plan AI deployment; human-AI interaction a key risk", "URL": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/", "access": "excerpt reviewed", "title": "60% plan AI deployment; human-AI interaction a key risk", "publication_date": "2025-09", "claims_checked": ["Oilfield Technology reports DNV intentions for the year ahead, not observed completed deployment. The surveyed population spans energy companies, not solely oil and gas."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R121", "Source_date": "2025-01", "Company": "DNV Cyber", "Source_type": "DNV", "Title": "65% of energy firms see cyber as top risk", "URL": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/", "access": "excerpt reviewed", "title": "65% of energy firms see cyber as top risk", "publication_date": "2025-01", "claims_checked": ["DNV January 22, 2025 release supports the retained 375-respondent energy-sector cyber figures; these are respondent views rather than measured incident rates."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R122", "Source_date": "2025-04", "Company": "International Energy Agency", "Source_type": "IEA", "Title": "IEA: oil and gas an early AI adopter", "URL": "https://www.iea.org/reports/energy-and-ai", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.iea.org/reports/energy-and-ai", "access": "excerpt reviewed", "title": "IEA: oil and gas an early AI adopter", "publication_date": "2025-04", "claims_checked": ["IEA report landing page dated April 10, 2025 was reopened. The detailed supercomputer statistics were not checked in the underlying chapter during this context audit."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R123", "Source_date": "2026-04", "Company": "PwC", "Source_type": "PwC", "Title": "20% of companies capture 74% of AI value", "URL": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html", "access": "excerpt reviewed", "title": "20% of companies capture 74% of AI value", "publication_date": "2026-04", "claims_checked": ["PwC April 13, 2026 release supports 74%/20% and the 1,217-executive, 25-sector population. This is a cross-industry self-report study, not oil-and-gas operating evidence."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R124", "Source_date": "Undated article; compares 2024–2025 survey waves", "Company": "Novi Labs", "Source_type": "Novi Labs", "Title": "Reservoir engineer AI use rises to 56%", "URL": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/", "access": "excerpt reviewed", "title": "Reservoir engineer AI use rises to 56%", "publication_date": "Undated article; compares 2024–2025 survey waves", "claims_checked": ["Novi vendor survey supports headline adoption and well-forecasting figures among more than 150 US reservoir engineers. Self-selected/vendor survey is not a census of operators; publication month not confirmed."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R125", "Source_date": "2023-11-07", "Company": "Shell", "Source_type": "Microsoft Customer Stories", "Title": "Shell's DIY citizen developer program passes 4,000 builders", "URL": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk", "access": "excerpt reviewed", "title": "Shell's DIY citizen developer program passes 4,000 builders", "publication_date": "2023-11-07", "claims_checked": ["Microsoft case is November 7, 2023, not June. DIY is low-code/citizen development; the disclosed developer count does not establish AI adoption."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R126", "Source_date": "2025-06-23", "Company": "Aker BP", "Source_type": "Microsoft Customer Stories", "Title": "Aker BP: Copilot for all staff plus champion network", "URL": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-06-23", "status": "full text reviewed", "locator": "Employee Copilot rollout, governance, Knowledge Agents"}], "Audit_reviews": [{"url": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "title": "Copilot rollout, employee agents and Knowledge Agents", "publication_date": "2025-06-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Aker BP started with 300 Copilot licenses and planned all-staff licensing in 2025; trained employees, set up help desk and champion/accountability network.", "Employee/small-team-built agents, including Knowledge Agents, answer repetitive documentation questions and route harder queries; several agents live.", "Microsoft case quotes 96% active users in prior four weeks; vendor-published customer metric, not independent audit."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "R127", "Source_date": "2026-04", "Company": "Society of Petroleum Engineers", "Source_type": "SPE", "Title": "SPE Forum: AI in Upstream E&P 2030", "URL": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond", "access": "excerpt reviewed", "title": "SPE Forum: AI in Upstream E&P 2030", "publication_date": "2026-04", "claims_checked": ["Event page establishes April 7–10, 2026 forum and topics. Agenda is not evidence of completed operator deployments or measured benefits."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R128", "Source_date": "2025-12", "Company": "The Open Group OSDU Forum", "Source_type": "The Open Group", "Title": "OSDU pushes an AI-ready open data platform", "URL": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/", "access": "excerpt reviewed", "title": "OSDU pushes an AI-ready open data platform", "publication_date": "2025-12", "claims_checked": ["The Open Group post describes work and forward priorities. Community implementation and certification should not be represented as completed deliverables based on forward-looking testimonials."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R129", "Source_date": "Historical context; exact article date not re-established", "Company": "Rice University / Chevron", "Source_type": "Rice University D2K Lab", "Title": "Chevron Digital Scholars earn Rice data-science master's", "URL": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation", "access": "excerpt reviewed", "title": "Chevron Digital Scholars earn Rice data-science master's", "publication_date": "Historical context; exact article date not re-established", "claims_checked": ["Rice interview supports Chevron Digital Scholars and a reservoir-decline capstone. Exact publication date and complete cohort metrics were not re-established from the reviewed excerpt."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R130", "Source_date": "2026-02-12", "Company": "US Department of Energy", "Source_type": "US Department of Energy", "Title": "Genesis Mission targets AI for subsurface energy", "URL": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and", "access": "excerpt reviewed", "title": "Genesis Mission targets AI for subsurface energy", "publication_date": "2026-02-12", "claims_checked": ["DOE 26-challenge announcement is February 12, 2026. Removed unsupported $5 billion funding and 17-lab numeric claims from this single-source record."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R131", "Source_date": "2025-10", "Company": "North Sea Transition Authority", "Source_type": "NSTA", "Title": "NSTA funds AI chatbot for North Sea data", "URL": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/", "access": "excerpt reviewed", "title": "NSTA funds AI chatbot for North Sea data", "publication_date": "2025-10", "claims_checked": ["NSTA source supports a £107,000 grant and planned chatbot/data-viewer work. A grant and planned assessment are not proof of production deployment."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R132", "Source_date": "2024-03", "Company": "Havtil (Norwegian Ocean Industry Authority)", "Source_type": "Havtil", "Title": "Norway regulator flags AI black-box and data risks", "URL": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/", "access": "excerpt reviewed", "title": "Norway regulator flags AI black-box and data risks", "publication_date": "2024-03", "claims_checked": ["Havtil article supports risk considerations and supervised use. Removed the separate alleged 2025 dropped-object pilot, which the linked 2024 article does not establish."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R133", "Source_date": "2025-01", "Company": "Saudi Aramco", "Source_type": "Middle East AI News", "Title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "URL": "https://www.middleeastainews.com/p/aramco-ai--drives-4-billion-value", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.middleeastainews.com/p/aramco-ai--drives-4-billion-value", "access": "full text reviewed", "title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "publication_date": "2025-01", "claims_checked": ["AI-037 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}]}, {"Source_ID": "R134", "Source_date": "2026-08", "Company": "Journal of Petroleum Technology (SPE)", "Source_type": "JPT", "Title": "Why O&G AI pilots fail to scale", "URL": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas", "access": "excerpt reviewed", "title": "Why O&G AI pilots fail to scale", "publication_date": "2026-08", "claims_checked": ["SPE commentary cites a cross-industry generative-AI study for 95% without measurable P&L impact. It must not be generalized to an observed 95% failure rate for all oil-and-gas AI."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R135", "Source_date": "2026-08", "Company": "Enabled Emissions Campaign / Nature study", "Source_type": "Inside Climate News", "Title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "URL": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/", "access": "inaccessible", "title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "publication_date": "2026-08", "claims_checked": ["Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R136", "Source_date": "2025-12", "Company": "Microsoft (shareholders and ex-employees)", "Source_type": "As You Sow", "Title": "Investors and ex-staff challenge Microsoft's O&G AI work", "URL": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products", "access": "excerpt reviewed", "title": "Investors and ex-staff challenge Microsoft's O&G AI work", "publication_date": "2025-12", "claims_checked": ["As You Sow December 4, 2025 statement establishes an investor proposal and advocacy claims. Removed unsupported cloud-market-share and production-uplift figures from this reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "R137", "Source_date": "2025-09", "Company": "ConocoPhillips", "Source_type": "Offshore Technology", "Title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "URL": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/", "Locator": "Locator not independently checked", "Short_excerpt": "", "Verification": "Imported reference; source verification pending", "Notes": "User-supplied reference preserved. Claims below require source review.", "Accessed": "Not independently accessed", "Audit_reviews": [{"url": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/", "access": "inaccessible", "title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "publication_date": "2025-09", "claims_checked": ["Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}]}, {"Source_ID": "RS001", "Source_date": "2025-11-12", "Company": "Chevron", "Source_type": "company announcement", "Title": "A smarter way to prospect for oil and gas — Chevron ApEX", "URL": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "Locator": "How does ApEX work? / Advantages in exploration", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2025-11-12", "status": "full text reviewed", "locator": "How does ApEX work? / Advantages in exploration"}], "Audit_reviews": [{"url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "title": "A smarter way to prospect for oil and gas", "publication_date": "2025-11-12", "notes": "Official Chevron newsroom article reopened successfully; replaces the earlier stale inaccessible status.", "claims_checked": ["Chevron says ApEX launched Aug. 2024 and uses specialized agents to search more than one million files; the issuer page reports a 4 TB corpus and seconds-fast retrieval.", "The article does not state that pre-drill studies fell from weeks to minutes."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS002", "Source_date": "Undated; see individual source review", "Company": "bp", "Source_type": "annual report", "Title": "bp Sustainability Report 2024", "URL": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "Locator": "Advancing safety through AI, bpx energy", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": null, "status": "excerpt reviewed", "locator": "Advancing safety through AI, bpx energy"}], "Audit_reviews": [{"url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "access": "excerpt reviewed", "title": "bp Sustainability Report 2024", "publication_date": "Undated; bp Sustainability Report 2024 reporting period", "claims_checked": ["Official indexed report text says bpx implemented AI driving-risk alerts across Permian operations in 2024 and attributes 89% reduction in distraction events and 75% reduction in tailgating between Dec 2023 and Dec 2024."], "notes": "Direct open attempted twice this audit and returned 403. Official bp PDF is indexed by the web search tool and returned the substantive “Advancing safety through AI” passage; this is excerpt-level evidence, not direct PDF full-text review. Metrics are company-reported and period is Dec 2023–Dec 2024.", "partition": 1}]}, {"Source_ID": "RS003", "Source_date": "Undated; see individual source review", "Company": "ExxonMobil", "Source_type": "company announcement", "Title": "Digital technologies — ExxonMobil", "URL": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "Locator": "Autonomous drilling", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Autonomous drilling"}], "Audit_reviews": [{"url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "access": "full text reviewed", "title": "Digital technologies — ExxonMobil", "publication_date": "Undated; see individual source review", "claims_checked": ["ExxonMobil describes a proprietary AI drilling advisory system in use in Guyana that determines drilling parameters and enables closed-loop control without human intervention."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: ; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS004", "Source_date": "2025-09-26", "Company": "TotalEnergies / Cognite", "Source_type": "company announcement", "Title": "Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI", "URL": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "Locator": "PDF pp. 1–2", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2025-09-26", "status": "full text reviewed", "locator": "PDF pp. 1–2"}], "Audit_reviews": [{"url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "title": "Industrial data and AI foundation across operated upstream", "publication_date": "2026-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three-year rollout of Cognite industrial data and AI platform across all TotalEnergies-operated upstream assets worldwide, drilling through production.", "The initiative contextualizes industrial data, supports dynamic asset visualization and monitoring of critical equipment, and aims to accelerate applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS005", "Source_date": "2026-02", "Company": "TotalEnergies", "Source_type": "investor presentation", "Title": "2025 Results and 2026 Objectives Presentation", "URL": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf", "Locator": "PDF slides 28–29", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-02", "status": "full text reviewed", "locator": "PDF slides 28–29"}], "Audit_reviews": [{"url": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf", "title": "2025 Results and 2026 Objectives Presentation", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Reported data platform footprint: 40 upstream, 16 refining and chemicals, and 500 integrated power sites; platform captures 10x real-time data points and contextualizes 35 million points.", "The deck names three AI strategic programs and a new global competency center in India, with target production and availability uplift but no quantified realized uplift."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS006", "Source_date": "2026-01-05", "Company": "Woodside", "Source_type": "Operator publication", "Title": "Right fix at right time – AI delivers for Woodside", "URL": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "Locator": "Angel pilot / North West Shelf launch, AWS collaboration and model processing-time paragraphs", "Short_excerpt": "five days to under two hours", "Verification": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "Notes": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.", "Accessed": "2026-10-18", "Research_reviews": [{"cohort": "majors", "date": "2026-01-05", "status": "full text reviewed", "locator": "Maint Intel"}], "Audit_reviews": [{"url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "access": "full text reviewed", "title": "Right fix at right time – AI delivers for Woodside", "publication_date": "2026-01-05", "claims_checked": ["AI-042 source claim(s) checked against cited profile"], "notes": "Woodside page reviewed in full: Maint Intel workflow, Angel testing, North West Shelf launch, planned 2026 rollout and processing-time metric; the packet 15% annual hours estimate does not appear in the reopened page.", "partition": 2}], "Publisher": "Woodside", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original text read; targeted workflow/metric check", "independent_validation": "Not established", "note": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified."}}, {"Source_ID": "RS007", "Source_date": "2026-02", "Company": "Repsol", "Source_type": "annual report", "Title": "Consolidated Management Report 2025 — Repsol", "URL": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "Locator": "5.5 Technology and digitalization", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-02-19", "status": "excerpt reviewed", "locator": "5.5 Technology and digitalization"}, {"cohort": "majors_backfill", "date": "2026-02", "status": "excerpt reviewed", "locator": "Section 5.5 Technology and digitalization"}], "Audit_reviews": [{"url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "title": "Consolidated Management Report 2025 — Repsol", "publication_date": "2026-02-19", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["Repsol reports Automated Well Data Ingestion automates daily drilling and geology reports and integrates data into corporate systems to speed exploration decisions.", "Annual report later lists 34 agents in three workflows, distinct from the Oct 2025 pilot snapshot of 22 agents.", "2025 report says Competence Center promoted >75 initiatives after two years, and the multi-agent platform had 34 agents in three workflows.", "Report names automated well-data ingestion for daily drilling/geology reports, integration into corporate systems and faster exploration decision-making."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS008", "Source_date": "2026-07-28", "Company": "Eni", "Source_type": "company announcement", "Title": "Eni opens its supercomputing capabilities to the innovation ecosystem", "URL": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "Locator": "Research and Development", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-07-28", "status": "full text reviewed", "locator": "Research and Development"}], "Audit_reviews": [{"url": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "access": "full text reviewed", "title": "Eni opens its supercomputing capabilities to the innovation ecosystem", "publication_date": "2026-07-28", "claims_checked": ["Eni opened HPC infrastructure, expertise and tools to companies, startups and research centres to develop AI and digital innovation; HPC6 and HPC7 together exceed exascale-class performance."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS009", "Source_date": "2026-04", "Company": "Eni", "Source_type": "annual report", "Title": "Eni for 2025 — Just Transition", "URL": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "Locator": "Data and artificial intelligence", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-04", "status": "excerpt reviewed", "locator": "Data and artificial intelligence"}], "Audit_reviews": [{"url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "access": "inaccessible", "title": "Eni for 2025 — Just Transition", "publication_date": "2026-04", "claims_checked": ["Eni reports around 300 AI use cases (+40% vs. 2024), around 20 AI agents; digital plant target ranges of 2–3% production enhancement and emissions reduction and drilling automation target of 35% less drilling time."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS010", "Source_date": "Undated; see individual source review", "Company": "Aker BP", "Source_type": "company announcement", "Title": "Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation", "URL": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "Locator": "Yggdrasil remote operations", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Yggdrasil remote operations"}], "Audit_reviews": [{"url": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "title": "AI agents for Yggdrasil remote operations", "publication_date": "2026-02-11", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Yggdrasil operations from onshore centre planned from 2027; AI agents planned from day one for condition-based maintenance and planning, with scaling to other workflows as mature.", "Aker BP says industrial data platform Cognite Data Fusion structures thousands of sensor streams and data is to be shared with people and AI agents."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS011", "Source_date": "2026-02-11", "Company": "Aker BP", "Source_type": "earnings call", "Title": "Aker BP Q4 2025 transcript", "URL": "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf", "Locator": "PDF pp. 5–6", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-02-11", "status": "full text reviewed", "locator": "PDF pp. 5–6"}], "Audit_reviews": [{"url": "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf", "title": "Aker BP Q4 2025 transcript", "publication_date": "2026-02-11", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes AI-enabled earlier exploration decisions and AI agents reducing operational troubleshooting time and improving uptime.", "The CEO describes high-quality structured real-time data connection across the company; 10 years of digital investment is contextual foundation.", "Microsoft speaker (partner attribution) says Copilot adoption nearly 100%; not independently audited."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS012", "Source_date": "2026-05-26", "Company": "Santos", "Source_type": "investor presentation", "Title": "Santos 2026 Investor Briefing Day", "URL": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "Locator": "Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross", "Short_excerpt": "", "Verification": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "Notes": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology.", "Accessed": "2026-10-14", "Research_reviews": [{"cohort": "majors", "date": "2026-05", "status": "excerpt reviewed", "locator": "Technology program"}], "Audit_reviews": [{"url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "access": "inaccessible", "title": "Santos 2026 Investor Briefing Day", "publication_date": "2026-05", "claims_checked": ["Santos says technology program applies digital and AI capability across integrated value chain, including cost reduction and production uplift across drilling, completions, projects and facilities; specific quantified AI-only impacts not isolated."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}], "Current_review": {"last_checked": "2026-10-14", "source_read": "Original 84-page issuer PDF and rendered slide 72 reviewed", "current_access": "Original PDF retrieved with standard browser request header after web-reader access failure", "independent_validation": "Not established", "note": "Release cover confirms 26 May 2026. Earlier excerpt/access notes remain historical; current column attribution is reviewed. The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology."}}, {"Source_ID": "RS013", "Source_date": "2026-06-04", "Company": "C3 AI", "Source_type": "partner case study", "Title": "C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations", "URL": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "Locator": "News Details, lines 285–297", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": "2026-06-04", "status": "full text reviewed", "locator": "News Details, lines 285–297"}], "Audit_reviews": [{"url": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "title": "Agentic root-cause analysis in global reliability program", "publication_date": "2026-06-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["C3 AI says it has worked with Shell since 2018 on predictive maintenance monitoring over 13,000 equipment items.", "A new multi-year agreement announced June 2026 extends reliability AI and introduces agent-based root cause analysis and remediation; the agreement is prospective.", "Program uses C3 AI Reliability and C3 Agentic AI Platform on Microsoft Azure.", "Hundreds of millions of dollars value is a C3 AI executive claim, not independent or Shell-attributed validation."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS014", "Source_date": "Undated; see individual source review", "Company": "Shell", "Source_type": "company announcement", "Title": "Safety at Pearl: Every second counts, every role matters", "URL": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "Locator": "Being smart about safety", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors", "date": null, "status": "full text reviewed", "locator": "Being smart about safety"}], "Audit_reviews": [{"url": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "title": "Real-time machine vision and equipment health AI at Pearl GTL", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Shell says Pearl GTL uses machine vision to flag unsafe behaviors and hazardous conditions in real time and alert supervisors.", "Pearl uses AI for near-real-time equipment health insights to reduce manual inspections and support earlier intervention.", "The plant was recognized by the World Economic Forum Global Lighthouse Network in September 2025."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS015", "Source_date": "Undated; accessed 2026-09-27", "Company": "Aramco", "Source_type": "company announcement", "Title": "Aramco AI and Big Data", "URL": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "Locator": "Fadhili Gas Plant section", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Fadhili Gas Plant section"}], "Audit_reviews": [{"url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "access": "full text reviewed", "title": "Aramco AI and Big Data", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-141 source claim(s) checked against cited profile"], "notes": "Aramco AI and Big Data page directly opened; current page is undated, so accessed date is not publication date. Fadhili claim must be scoped to pilot and use separate dates from any source evidence.", "partition": 2}]}, {"Source_ID": "RS016", "Source_date": "2025-01-16", "Company": "ADNOC", "Source_type": "company announcement", "Title": "ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution", "URL": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "Locator": "Trial findings and next-phase paragraphs", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-01-16", "status": "full text reviewed", "locator": "Trial findings and next-phase paragraphs"}], "Audit_reviews": [{"url": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "access": "full text reviewed", "title": "ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}]}, {"Source_ID": "RS017", "Source_date": "2026-02-27", "Company": "PETRONAS", "Source_type": "annual report", "Title": "PETRONAS Integrated Report 2025", "URL": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "Locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI); 21 October additional locator: PDF p.29 / printed p.55; Turning Data into Production Upside", "Short_excerpt": "", "Verification": "Original 10,849,024-byte PDF retrieved HTTP200; editor extracted and visually inspected spread54–55 (PDF29). Publication day retained from existing source register; exact posting day not re-audited in this pass. Earlier claim reviews retain their original scope.", "Notes": "Research expansion. Company and vendor claims are not independently validated. Follow-up source review on 29 September 2026; performance remains source-attributed.", "Accessed": "2026-10-21", "Research_reviews": [{"cohort": "nocs", "date": "2026-02-27", "status": "excerpt reviewed", "locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)"}, {"cohort": "nocs", "date": "2026-02-27", "status": "excerpt reviewed", "locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)"}], "Audit_reviews": [{"url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "access": "inaccessible", "title": "PETRONAS Integrated Report 2025", "publication_date": "2026-02-27", "claims_checked": ["AI-142 source claim(s) checked against cited profile", "AI-143 source claim(s) checked against cited profile"], "notes": "PETRONAS 2025 Integrated Report PDF directly attempted twice; web viewer returned payload/size error. Search excerpts only; not full-text verified.", "partition": 2}], "Publisher": "PETRONAS", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original 10,849,024-byte PDF retrieved HTTP200; editor extracted and visually inspected spread54–55 (PDF29). Publication day retained from existing source register; exact posting day not re-audited in this pass.", "independent_validation": "Not established", "note": "No field names, user population, human approval or automatic control permissions are disclosed. The chapter’s broad AI heading cannot establish IAM’s method. Do not transfer PROTEAN’s equipment footprint or ERMAI’s time comparison to IAM."}}, {"Source_ID": "RS018", "Source_date": "2025-06-18", "Company": "PETRONAS", "Source_type": "company announcement", "Title": "PETRONAS Carigali Unveils Pioneering AI Venture - TriCipta AI", "URL": "https://www.petronas.com/bm/node/1688", "Locator": "Media release", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-06-18", "status": "full text reviewed", "locator": "Media release"}], "Audit_reviews": [{"url": "https://www.petronas.com/bm/node/1688", "access": "full text reviewed", "title": "PETRONAS Carigali Unveils Pioneering AI Venture - TriCipta AI", "publication_date": "2025-06-18", "claims_checked": ["AI-143 source claim(s) checked against cited profile"], "notes": "PETRONAS TriCipta company release opened; product/venture scope checked, but no operator deployment or performance result established.", "partition": 2}]}, {"Source_ID": "RS019", "Source_date": "2025-08-01", "Company": "CNPC", "Source_type": "annual report", "Title": "2024 Annual Report", "URL": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "Locator": "Launch of 70-billion-parameter Kunlun Large Model", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-08-01", "status": "excerpt reviewed", "locator": "Launch of 70-billion-parameter Kunlun Large Model"}], "Audit_reviews": [{"url": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "access": "inaccessible", "title": "2024 Annual Report", "publication_date": "2025-08-01", "claims_checked": ["CNPC reports launching a 70-billion-parameter Kunlun model in 2024 with 43 professional/general scenarios for petroleum.", "Report names a joint development group of CNPC, China Mobile, Huawei and iFLYTEK; platform included eight trained large models, 18 application scenarios, specialized datasets, AI middleware and computing environment."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-08-01; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS020", "Source_date": "2026-05-29", "Company": "CNPC News Center", "Source_type": "company announcement", "Title": "中国石油昆仑大模型迈入主动智能新阶段", "URL": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml", "Locator": "Release overview and application examples", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2026-05-29", "status": "full text reviewed", "locator": "Release overview and application examples"}], "Audit_reviews": [{"url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml", "access": "excerpt reviewed", "title": "中国石油昆仑大模型迈入主动智能新阶段", "publication_date": "2026-05-29", "claims_checked": ["CNPC says Kunlun reached 152 deployed scenarios across the energy and chemical value chain in May 2026.", "Specific example: crude purchase evaluation with processing plans and risk advice in five minutes versus hours or days of manual analysis, as reported by a refinery manager."], "notes": "Exact direct URL failed due encoding/internal error. Official CNPC portal search-index result exposed the article text and was reviewed as an excerpt, not direct full-text. Use official 2025 annual report as additional accessible CNPC source for scale/context; preserve this URL’s direct-fetch limitation.", "partition": 1}]}, {"Source_ID": "RS021", "Source_date": "2025-05-28", "Company": "Sinopec / Jianghan Oilfield", "Source_type": "company announcement", "Title": "〖中国石化报〗多场景推进人工智能技术应用", "URL": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "Locator": "Jianghan Oilfield AI deployment section", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-05-28", "status": "full text reviewed", "locator": "Jianghan Oilfield AI deployment section"}], "Audit_reviews": [{"url": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "access": "inaccessible", "title": "〖中国石化报〗多场景推进人工智能技术应用", "publication_date": "2025-05-28", "claims_checked": ["Jianghan Oilfield describes AI for shale-gas digital-twin reservoir updates, drilling risk alerts, fracture optimization, video safety monitoring, field inspection and work assistance.", "Reported examples include fracture simulation accuracy above 80%, video safety inspection efficiency up over 50% and missed-violation detection below 10%.", "A liquid-loading warning use case had more than 800 well instances, parsed 150,000 records/day, and company-reported warning accuracy above 95%; treatment efficiency above 90%."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS022", "Source_date": "2025-08-27", "Company": "CNOOC", "Source_type": "company announcement", "Title": "2025 Interim Results", "URL": "https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html", "Locator": "AI+ key scenarios and intelligent drilling paragraphs", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-08-27", "status": "full text reviewed", "locator": "AI+ key scenarios and intelligent drilling paragraphs"}], "Audit_reviews": [{"url": "https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html", "title": "AI+ smart field, intelligent drilling and typhoon response", "publication_date": "2025-08-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["CNOOC reported implementing priority AI+ scenarios, scaling intelligent injection-production technology, and advancing intelligent drilling/completion construction.", "Company reported demonstration project drilling speed improvement of 26%; the release does not allocate this result exclusively to AI.", "Deep Sea No.1 smart gas field named as a national excellent-level smart factory; satellite remote sensing, unmanned equipment and AI algorithms were combined for typhoon response."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS023", "Source_date": "Undated; accessed 2026-09-27", "Company": "Petrobras", "Source_type": "company announcement", "Title": "Inovação e tecnologia: nossa fonte de energia", "URL": "https://petrobras.com.br/inovacao-e-tecnologia", "Locator": "Smart Tocha entry", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Smart Tocha entry"}], "Audit_reviews": [{"url": "https://petrobras.com.br/inovacao-e-tecnologia", "access": "full text reviewed", "title": "Inovação e tecnologia: nossa fonte de energia", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-146 source claim(s) checked against cited profile"], "notes": "Petrobras innovation and technology page opened; umbrella page does not itself prove the specific Smart Tocha workflow in sufficient detail.", "partition": 2}]}, {"Source_ID": "RS024", "Source_date": "2026-03-04", "Company": "Ecopetrol", "Source_type": "annual report", "Title": "Informe de gestión y resultados 2025", "URL": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "Locator": "Technology and innovation section; AI ecosystem and upstream planner", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2026-03-31", "status": "excerpt reviewed", "locator": "Technology and innovation section; AI ecosystem and upstream planner"}], "Audit_reviews": [{"url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "access": "full text reviewed", "title": "Informe de gestión y resultados 2025", "publication_date": "2026-03-31", "claims_checked": ["AI-147 source claim(s) checked against cited profile"], "notes": "Ecopetrol 2025 management report PDF opened; report text checked for AI/analytics distinctions. Group-science benefits are not automatically AI results.", "partition": 2}]}, {"Source_ID": "RS025", "Source_date": "2026-05-04", "Company": "Ecopetrol", "Source_type": "company announcement", "Title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "URL": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/", "Locator": "CADE paragraphs", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2026-05-04", "status": "full text reviewed", "locator": "CADE paragraphs"}], "Audit_reviews": [{"url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/", "access": "inaccessible", "title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "publication_date": "2026-05-04", "claims_checked": ["AI-148 source claim(s) checked against cited profile"], "notes": "Ecopetrol CADE announcement URL directly attempted; web cache miss. No claim-level verification from this URL.", "partition": 2}]}, {"Source_ID": "RS026", "Source_date": "2025; publication day not confirmed", "Company": "YPF", "Source_type": "annual report", "Title": "Reporte de Sustentabilidad 2024", "URL": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "Locator": "Innovation and Technology / Use of artificial intelligence", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025; publication day not confirmed", "status": "excerpt reviewed", "locator": "Innovation and Technology / Use of artificial intelligence"}], "Audit_reviews": [{"url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "title": "Generative AI for contract cataloging, customer service and refinery inspection", "publication_date": "Undated source; see chronology", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS027", "Source_date": "2025-09-01", "Company": "Pertamina Hulu Rokan", "Source_type": "company announcement", "Title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "URL": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "Locator": "PINTAR description", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-09-01", "status": "excerpt reviewed", "locator": "PINTAR description"}], "Audit_reviews": [{"url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "access": "inaccessible", "title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "publication_date": "2025-09-01", "claims_checked": ["AI-150 source claim(s) checked against cited profile"], "notes": "PHE PINTAR PDF URL directly attempted; request rejected. No full-text verification.", "partition": 2}]}, {"Source_ID": "RS028", "Source_date": "2025-10-27", "Company": "Pertamina Geothermal Energy", "Source_type": "company announcement", "Title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "URL": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "Locator": "G-Bionic value case", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-10-27", "status": "full text reviewed", "locator": "G-Bionic value case"}], "Audit_reviews": [{"url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "access": "full text reviewed", "title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "publication_date": "2025-10-27", "claims_checked": ["AI-151 source claim(s) checked against cited profile"], "notes": "PGE release opened; its digital/geothermal scope is adjacent energy, not oil-and-gas operations. Bundled efficiency results cannot be isolated as AI outcomes.", "partition": 2}]}, {"Source_ID": "RS029", "Source_date": "2025-11-03", "Company": "ONGC", "Source_type": "company announcement", "Title": "ONGC tender for automated gas-lift system", "URL": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "Locator": "AI-driven optimization and predictive maintenance requirements", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-11-03", "status": "excerpt reviewed", "locator": "AI-driven optimization and predictive maintenance requirements"}], "Audit_reviews": [{"url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "access": "full text reviewed", "title": "ONGC tender for automated gas-lift system", "publication_date": "2025-11-03", "claims_checked": ["Tender technical scope requires AI/ML algorithms to analyze well data and optimize gas-lift injection in real time, with predictive maintenance based on valve response and pressure trends.", "SCADA and existing wellhead control panel integration are specified. Procurement requirements show intended scope, not completed deployment."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-03; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS030", "Source_date": "2025-11-02", "Company": "ADNOC", "Source_type": "company announcement", "Title": "ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI", "URL": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "Locator": "Copilot history, training and productivity aggregate", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-11-02", "status": "full text reviewed", "locator": "Copilot history, training and productivity aggregate"}, {"cohort": "nocs", "date": "2025-11-02", "status": "full text reviewed", "locator": "Operational AI agents and separate energy-for-AI component"}], "Audit_reviews": [{"url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "access": "full text reviewed", "title": "ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI", "publication_date": "2025-11-02", "claims_checked": ["ADNOC reports enterprise Copilot rollout in 2023, >40,000 employees completing AI training, utilization above 90%, and >70,000 monthly productivity hours; no training causal attribution stated.", "ADNOC and Microsoft announced plans to co-develop and deploy AI agents across ADNOC’s value chain; no particular agent or deployment was named.", "The wider release separately includes Masdar/XRG energy and infrastructure solutions for Microsoft AI/data-center growth."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-02; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS031", "Source_date": "2026", "Company": "ADNOC", "Source_type": "annual report", "Title": "ADNOC Sustainability Report 2025", "URL": "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "Locator": "AI and digital capabilities / workforce training", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2026", "status": "full text reviewed", "locator": "AI and digital capabilities / workforce training"}], "Audit_reviews": [{"url": "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "access": "full text reviewed", "title": "ADNOC Sustainability Report 2025", "publication_date": "2026", "claims_checked": ["In 2025, ADNOC says its Introduction to Agentic AI program engaged >18,000 participants and resulted in >10,000 agents created across operations."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS032", "Source_date": "2026-09-03", "Company": "PETRONAS", "Source_type": "company announcement", "Title": "PETRONAS Advances Malaysia’s Upstream Data Platform with Agentic AI to Accelerate E&P Investment", "URL": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep", "Locator": "Agreement, platform workflow and MPM targets", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2026-09-03", "status": "full text reviewed", "locator": "Agreement, platform workflow and MPM targets"}], "Audit_reviews": [{"url": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep", "access": "full text reviewed", "title": "PETRONAS Advances Malaysia’s Upstream Data Platform with Agentic AI to Accelerate E&P Investment", "publication_date": "2026-09-03", "claims_checked": ["AI-080 source claim(s) checked against cited profile"], "notes": "PETRONAS agentic myPROdata release opened; reviewed stated scope and roadmap language. It establishes product/platform enhancement, not quantified asset outcomes.", "partition": 2}]}, {"Source_ID": "RS033", "Source_date": "2025-Q4", "Company": "Kuwait Oil Company", "Source_type": "company periodical", "Title": "The Kuwait Digest 15 (October–December 2025)", "URL": "https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "Locator": "PDF pp. 14–16, AI Innovation Center / G Agent", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "nocs", "date": "2025-Q4", "status": "full text reviewed", "locator": "PDF pp. 14–16, AI Innovation Center / G Agent"}], "Audit_reviews": [{"url": "https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "access": "full text reviewed", "title": "The Kuwait Digest 15 (October–December 2025)", "publication_date": "2025-Q4", "claims_checked": ["KOC says G Agent uses AI for drilling-rig resource management, systems integration and real-time data analysis; it also describes an agentic drilling-rig scheduling project.", "KOC names Microsoft and Halliburton partnership and says Ghaia AI manages technology programs/training at the Center."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-Q4; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS034", "Source_date": "2026-08", "Company": "Yahoo Finance", "Source_type": "secondary", "Title": "Devon Energy Q2 2026 Earnings Call Transcript", "URL": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "Locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-08", "status": "excerpt reviewed", "locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted"}, {"cohort": "north_america", "date": "2026-08", "status": "excerpt reviewed", "locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted"}], "Audit_reviews": [{"url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "access": "inaccessible", "title": "Devon Energy Q2 2026 Earnings Call Transcript", "publication_date": "2026-08", "claims_checked": ["AI-087 source claim(s) checked against cited profile"], "notes": "Yahoo-hosted Devon transcript directly attempted and returned 429; secondary source not accessible. No Coterra claim certified from it.", "partition": 2}, {"url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "title": "Closed-loop AI adjusts gas-lift settings at operating wells", "publication_date": "2026-08", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS035", "Source_date": "2025-02-19", "Company": "Occidental", "Source_type": "earnings call", "Title": "Q4 2024 Occidental Earnings Call Transcript", "URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf", "Locator": "CEO remarks: AI initiatives and AI Center of Excellence", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2025-02-19", "status": "excerpt reviewed", "locator": "CEO remarks: AI initiatives and AI Center of Excellence"}], "Audit_reviews": [{"url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf", "title": "AI Center of Excellence and cross-functional operating use cases", "publication_date": "2025-02-19", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Oxy named AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization; established an AI Center of Excellence to align cross-company initiatives."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS036", "Source_date": "Undated; see individual source review", "Company": "Suncor", "Source_type": "company announcement", "Title": "The Future of Canadian Mining is Autonomous — and Suncor is Leading the Way", "URL": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "Locator": "Base Plant fleet and command centre", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": null, "status": "full text reviewed", "locator": "Base Plant fleet, command centre and end-of-page target statement"}], "Audit_reviews": [{"url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "title": "Autonomous haulage fleet expansion at Base Plant", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["AHS fleet expansion reached almost 120 trucks across two integrated mines; operators manage flow from control room with real-time vehicle visibility; Suncor planned 140 equipped trucks by year end. Suncor says the system uses AI to optimize for driving conditions and reduce wear/maintenance."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS037", "Source_date": "2026-03-25", "Company": "Canadian Natural Resources", "Source_type": "annual report", "Title": "Canadian Natural 2025 Annual Information Form", "URL": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "Locator": "Artificial Intelligence section, p.21", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-03-25", "status": "excerpt reviewed", "locator": "Artificial Intelligence section, p.21"}], "Audit_reviews": [{"url": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "access": "full text reviewed", "title": "Canadian Natural 2025 Annual Information Form", "publication_date": "2026-03-25", "claims_checked": ["AI is used to expand employee knowledge access, deliver data-driven insight and improve efficiency via incremental, cost-justified deployment; company will not rely on it for time-sensitive/production-critical processes until controls meet standards; employees remain involved/accountable; board risk oversight described."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-03-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS038", "Source_date": "Undated; see individual source review", "Company": "Canadian Natural Resources", "Source_type": "company announcement", "Title": "Governance and Risk Management", "URL": "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/", "Locator": "Use of AI in operations", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": null, "status": "excerpt reviewed", "locator": "Use of AI in operations"}], "Audit_reviews": [{"url": "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/", "access": "full text reviewed", "title": "Governance and Risk Management", "publication_date": "Undated; see individual source review", "claims_checked": ["Corroborates program-level AI scope; refers readers to AIF for policy specifics."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: ; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS039", "Source_date": "2026-02-24", "Company": "Matador Resources", "Source_type": "company announcement", "Title": "Matador Q4 and FY2025 Results and 2026 Operating Plan", "URL": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "Locator": "CEO summary / 2026 focus areas", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Audit_reviews": [{"url": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "title": "Vendor-led, cautious AI integration strategy", "publication_date": "2026-02-25", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Matador lists AI integration among focus areas, without named deployment, vendor, use case, budget or outcome; distinct from natural-gas marketing to AI data centres."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS040", "Source_date": "2026-08-10", "Company": "The Motley Fool", "Source_type": "secondary", "Title": "Diamondback Q2 2026 Earnings Call Transcript", "URL": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "Locator": "Existing register citation; not fully re-opened in primary materials", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": "2026-08-10", "status": "excerpt reviewed", "locator": "Existing register citation; not fully re-opened in primary materials"}], "Audit_reviews": [{"url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "access": "inaccessible", "title": "Diamondback Q2 2026 Earnings Call Transcript", "publication_date": "2026-08-10", "claims_checked": ["Existing record links discussion of lift optimization and downtime to transcript; use case and outcomes remain unverified by issuer source."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-08-10; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS041", "Source_date": "Undated; see individual source review", "Company": "Suncor", "Source_type": "company announcement", "Title": "Dispatching innovation: Suncor's new dispatch system improves mine operations", "URL": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "Locator": "AI dispatch system; human dispatcher workflow", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "north_america", "date": null, "status": "full text reviewed", "locator": "AI dispatch system; human dispatcher workflow"}], "Audit_reviews": [{"url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "title": "Dynamic AI dispatch assigns mine haul trucks", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Dynamic dispatch automates dump/refueling assignments and slowdown rerouting; human dispatchers remain active."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS042", "Source_date": "Undated; accessed 2026-09-27", "Company": "SLB", "Source_type": "company product documentation", "Title": "Tela agentic-AI assistant catalog", "URL": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "Locator": "Tela feature and 33 workflow brochure sections", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Tela feature and 33 workflow brochure sections"}], "Audit_reviews": [{"url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "access": "full text reviewed", "title": "Tela agentic-AI assistant catalog", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "SLB Tela live product catalog opened; 33 is a mutable catalog/capability count, not customer deployments.", "partition": 2}]}, {"Source_ID": "RS043", "Source_date": "2026-06-15", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy", "URL": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "Locator": "Marketplace offerings and governance", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-06-15", "status": "full text reviewed", "locator": "Marketplace offerings and governance"}], "Audit_reviews": [{"url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "access": "full text reviewed", "title": "SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy", "publication_date": "2026-06-15", "claims_checked": ["AI-047 source claim(s) checked against cited profile"], "notes": "SLB Digital Marketplace launch opened; catalog/marketplace scope checked, not proof of deployments.", "partition": 2}]}, {"Source_ID": "RS044", "Source_date": "2026-06-05", "Company": "SLB", "Source_type": "partner case study", "Title": "Petoro and SLB: Pioneering AI-driven well planning on the Norwegian continental shelf", "URL": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "Locator": "Objectives, workflow, impact and deployment sections", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-06-05", "status": "full text reviewed", "locator": "Objectives, workflow, impact and deployment sections"}], "Audit_reviews": [{"url": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "title": "Petoro AI well-data extraction and field-development screening", "publication_date": "2026-06-05", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Petoro and SLB built Lumi/Innovation Factori workflows to extract data from well schematics/reports and rank field-development/sidetrack opportunities.", "Well Schematic Extractor and Well History use LLMs to extract casing, completion and formation information, with deterministic checks and user visual QC.", "SLB reports threefold productivity improvement; preliminary QC rose from two schematics/day to six or seven. This is vendor case-study reporting."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS045", "Source_date": "2025-10-17", "Company": "SLB", "Source_type": "earnings release", "Title": "SLB Announces Third-Quarter 2025 Results", "URL": "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609", "Locator": "Digital highlights / Petoro award", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-10-17", "status": "excerpt reviewed", "locator": "Digital highlights / Petoro award"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609", "title": "SLB Announces Third-Quarter 2025 Results", "publication_date": "2025-10-17", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Q3 earnings: Petoro awarded a generative-AI solution to optimize drilling resources using FDPlan and Lumi."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS046", "Source_date": "2026-01-23", "Company": "SLB", "Source_type": "earnings release", "Title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "URL": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "Locator": "Technology highlights, pp. 18-19", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-01-23", "status": "full text reviewed", "locator": "Technology highlights, pp. 18-19"}, {"cohort": "services", "date": "2026-01-23", "status": "full text reviewed", "locator": "Technology highlights, pp. 18-19"}, {"cohort": "services", "date": "2026-01-23", "status": "full text reviewed", "locator": "Technology highlights, pp. 18-19"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "access": "full text reviewed", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "publication_date": "2026-01-23", "claims_checked": ["Q4 release also records the bp partnership update.", "Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "access": "full text reviewed", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "publication_date": "2026-01-23", "claims_checked": ["AI-013 source claim(s) checked against cited profile"], "notes": "SLB Q4 2025 earnings PDF fully reviewed; includes OQ Neuro Oman, KOC AI work and broad digital results. Vendor evidence; no operator corroboration assumed.", "partition": 2}]}, {"Source_ID": "RS047", "Source_date": "Undated; accessed 2026-09-27", "Company": "SLB", "Source_type": "company case study", "Title": "SLB digital solutions save 1,500 personnel-on-board days and boost efficiency", "URL": "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "Locator": "Summary, objective and results; excerpt reviewed", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "excerpt reviewed", "locator": "Summary, objective and results; excerpt reviewed"}], "Audit_reviews": [{"url": "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "access": "full text reviewed", "title": "SLB digital solutions save 1,500 personnel-on-board days and boost efficiency", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["Performance Live with DrillOps and Neuro reduced offshore well-placement crew size 66% and saved >1,500 POB days across several Western Asia platforms.", "Case attributes benefits to remote operations and autonomous directional drilling; exact operator not identified in the case page."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: Undated; accessed 2026-09-27; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS048", "Source_date": "2025-01-06", "Company": "Halliburton", "Source_type": "company announcement", "Title": "Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program", "URL": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "Locator": "Deployment and results paragraphs", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-01-06", "status": "full text reviewed", "locator": "Deployment and results paragraphs"}], "Audit_reviews": [{"url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "access": "full text reviewed", "title": "Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program", "publication_date": "2025-01-06", "claims_checked": ["AI-109 source claim(s) checked against cited profile"], "notes": "Halliburton/Coterra announcement reviewed; source establishes automated fracturing, not an AI method.", "partition": 2}]}, {"Source_ID": "RS049", "Source_date": "2025-04-12", "Company": "Halliburton", "Source_type": "company announcement", "Title": "Chevron and Halliburton Enable Intelligent Hydraulic Fracturing", "URL": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "Locator": "Closed-loop intelligent fracturing description", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-04-12", "status": "full text reviewed", "locator": "Closed-loop intelligent fracturing description"}], "Audit_reviews": [{"url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "access": "full text reviewed", "title": "Chevron and Halliburton Enable Intelligent Hydraulic Fracturing", "publication_date": "2025-04-12", "claims_checked": ["AI-110 source claim(s) checked against cited profile"], "notes": "Halliburton/Chevron release reviewed; intelligent/closed-loop control is described, but source does not establish AI/ML.", "partition": 2}]}, {"Source_ID": "RS050", "Source_date": "2025-02-26", "Company": "Halliburton", "Source_type": "company announcement", "Title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "URL": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "Locator": "Deployment description", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-02-26", "status": "full text reviewed", "locator": "Deployment description"}], "Audit_reviews": [{"url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "access": "full text reviewed", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "publication_date": "2025-02-26", "claims_checked": ["AI-157 source claim(s) checked against cited profile"], "notes": "Halliburton/Sekal announcement reviewed; on-bottom autonomous drilling outcome and control workflow checked; no explicit AI method.", "partition": 2}]}, {"Source_ID": "RS051", "Source_date": "2025-10-01", "Company": "Weatherford", "Source_type": "company announcement", "Title": "Weatherford Launches Industrial Intelligence at FWRD 2025", "URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "Locator": "Product portfolio descriptions", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-10-01", "status": "full text reviewed", "locator": "Product portfolio descriptions"}], "Audit_reviews": [{"url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "title": "Industrial Intelligence portfolio for AI and edge operations", "publication_date": "2026-04-01", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Portfolio includes MetaFlow AI/physics-based virtual flow meter, Engineer’s Briefcase AI document transcription/recognition/translation tools and StrataCore production foundational model.", "It also includes unified data model, edge and SaaS software; portfolio launch does not itself verify customer deployment."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS052", "Source_date": "2026-04-01", "Company": "Weatherford", "Source_type": "sustainability report", "Title": "Weatherford 2025 Sustainability Report: Artificial Intelligence", "URL": "https://www.weatherford.com/documents/sustainability/sustainability-report/", "Locator": "Artificial Intelligence section", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-04-01", "status": "excerpt reviewed", "locator": "Artificial Intelligence section"}], "Audit_reviews": [{"url": "https://www.weatherford.com/documents/sustainability/sustainability-report/", "title": "Weatherford 2025 Sustainability Report: Artificial Intelligence", "publication_date": "2026-04-01", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["Weatherford reports an AI Impact Hub and agents embedded in internal workflows in 2025.", "Named agents review material test reports, cross-reference parts/inventory, classify spend and simplify bidding; training and human involvement stated."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS053", "Source_date": "2025-06-27", "Company": "NOV", "Source_type": "company announcement", "Title": "NOV supports record bit run and Egypt’s first remote operations center", "URL": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center", "Locator": "DBA functionality, field result and RTOC", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-06-27", "status": "full text reviewed", "locator": "DBA functionality, field result and RTOC"}], "Audit_reviews": [{"url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center", "access": "full text reviewed", "title": "NOV supports record bit run and Egypt’s first remote operations center", "publication_date": "2025-06-27", "claims_checked": ["AI-158 source claim(s) checked against cited profile"], "notes": "NOV Egypt/remote operations announcement opened; treated as supplier-authored deployment report; AI basis kept separate from remote automation.", "partition": 2}]}, {"Source_ID": "RS054", "Source_date": "2025-04-10", "Company": "Nabors", "Source_type": "company announcement", "Title": "Nabors and Corva Expand Strategic Alliance to Accelerate Digital Innovation", "URL": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "Locator": "RigCLOUD Powered by Corva scope", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-04-10", "status": "full text reviewed", "locator": "RigCLOUD Powered by Corva scope"}], "Audit_reviews": [{"url": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "title": "RigCLOUD Powered by Corva drilling intelligence", "publication_date": "2025-04-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Expanded alliance integrates RigCLOUD edge/cloud platform with Corva AI analytics and apps for real-time drilling data, predictive insights and optimization.", "Operators and contractors can use the tools at rigsite and remotely; SmartROS controls are optional. Announcement describes solution capability, not specific named customer rollout."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS055", "Source_date": "Undated; accessed 2026-09-27", "Company": "Nabors", "Source_type": "company product documentation", "Title": "Predictive Drilling - Nabors", "URL": "https://www.nabors.com/drilling-automation/predictive-drilling/", "Locator": "Success in Action and availability", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Success in Action and availability"}], "Audit_reviews": [{"url": "https://www.nabors.com/drilling-automation/predictive-drilling/", "title": "Predictive Drilling - Nabors", "publication_date": "Undated; accessed 2026-09-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Nabors reports the Predictive Drilling ROP optimizer combines AI/ML with SmartROS/RigCLOUD for remote autodriller setpoint control.", "First deployment reported 36% higher average ROP, 9.7% lower vibration and 80% lateral footage utilization; historic case source (2023) is outside window."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS056", "Source_date": "2026-02-04", "Company": "H&P", "Source_type": "earnings release", "Title": "Helmerich & Payne, Inc. Announces Fiscal First Quarter Results", "URL": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "Locator": "FlexRobotics deployment", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-02-04", "status": "full text reviewed", "locator": "FlexRobotics deployment"}], "Audit_reviews": [{"url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "title": "FlexRobotics rig-floor handling with AI-enabled zone monitoring", "publication_date": "2026-02-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["H&P reports FlexRobotics deployed on a Permian rig for a supermajor across multiple pads; growing customer interest.", "Rig-floor system is collaborative robotics for tubular handling. H&P product page says AI-enabled zone monitoring is an ancillary feature; core handling automation is not AI."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS057", "Source_date": "Undated; accessed 2026-09-27", "Company": "H&P", "Source_type": "company product documentation", "Title": "FlexRobotics Hands-Free Rig Floor", "URL": "https://www.hpinc.com/technologies/flexrobotics", "Locator": "Description, operation and AI-enabled zone monitoring", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Description, operation and AI-enabled zone monitoring"}], "Audit_reviews": [{"url": "https://www.hpinc.com/technologies/flexrobotics", "title": "FlexRobotics Hands-Free Rig Floor", "publication_date": "Undated; accessed 2026-09-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three robot arms handle pipe tailing, cleaning/alignment and racking. H&P lists 1:40 slip-to-slip connection times and red-zone personnel exposure reduction.", "System was developed with NOV and uses Yaskawa robot manipulators. Cameras support AI-enabled zone monitoring."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS058", "Source_date": "2026-02-04", "Company": "Patterson-UTI", "Source_type": "earnings release", "Title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025", "URL": "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm", "Locator": "Drilling Services", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-02-04", "status": "full text reviewed", "locator": "Drilling Services"}], "Audit_reviews": [{"url": "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm", "access": "full text reviewed", "title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025", "publication_date": "2026-02-04", "claims_checked": ["Patterson-UTI said nearly all rigs were equipped with proprietary Cortex automation applications by year-end 2025.", "Management cited continued development of automation software and demand for performance-enhancing products; this is scale of automation suite, not AI-specific model usage."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS059", "Source_date": "2025-04", "Company": "Precision Drilling", "Source_type": "sustainability report", "Title": "Precision Drilling 2024 ESG Report", "URL": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "Locator": "Alpha Suite technologies", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-04", "status": "excerpt reviewed", "locator": "Alpha Suite technologies"}], "Audit_reviews": [{"url": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "access": "inaccessible", "title": "Precision Drilling 2024 ESG Report", "publication_date": "2025-04", "claims_checked": ["Precision reports 78 rigs with Alpha digital technology at end-2024.", "Alpha suite combines digital automation and analytics with AI algorithms for drilling performance pattern recognition, real-time adjustments and predictive maintenance; report does not break out AI-specific results."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS060", "Source_date": "2026-04", "Company": "Saipem", "Source_type": "sustainability report", "Title": "Saipem Sustainability Overview 2025", "URL": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "Locator": "Harnessing AI for engineering section", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-04", "status": "excerpt reviewed", "locator": "Harnessing AI for engineering section"}], "Audit_reviews": [{"url": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "title": "AI Project Schedule Onshore for EPC risk forecasting", "publication_date": "2026-04", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Saipem’s AI for Engineering is a two-year cross-functional effort launched 2024; 13 use cases explored by 2025.", "AI Project Schedule Onshore analyzes historical project data with predictive models for EPC construction risks/delays and bottlenecks; 12 additional apps were planned for 2026."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS061", "Source_date": "2026-02-19", "Company": "Transocean", "Source_type": "annual report", "Title": "Transocean Ltd. 2025 Form 10-K", "URL": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "Locator": "Technology and AI discussion", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-02-19", "status": "full text reviewed", "locator": "Technology and AI discussion"}], "Audit_reviews": [{"url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "access": "inaccessible", "title": "Transocean Ltd. 2025 Form 10-K", "publication_date": "2026-02-19", "claims_checked": ["Transocean explicitly says AI is not currently used in offshore drilling operations; it has introduced AI tools for high-volume office/administrative tasks and is exploring productivity use cases.", "Company describes AI governance committee; fleet automation/robotics systems are separate from AI."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-02-19; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS062", "Source_date": "Undated; accessed 2026-09-27", "Company": "Halliburton", "Source_type": "partner case study", "Title": "Enhanced well delivery with LOGIX automation and remote operations", "URL": "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop", "Locator": "Overview, solution, results", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Overview, solution, results"}], "Audit_reviews": [{"url": "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop", "access": "full text reviewed", "title": "Enhanced well delivery with LOGIX automation and remote operations", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-016 source claim(s) checked against cited profile"], "notes": "Halliburton LOGIX source opened; automation, ROP and well placement content checked; it describes autonomous/automated drilling, not a demonstrated AI model.", "partition": 2}]}, {"Source_ID": "RS063", "Source_date": "2026-07-21", "Company": "Halliburton", "Source_type": "company announcement", "Title": "Halliburton and Shape Digital form strategic collaboration to advance digital asset performance management", "URL": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management", "Locator": "Collaboration scope and connected production planning", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-07-21", "status": "full text reviewed", "locator": "Collaboration scope and connected production planning"}], "Audit_reviews": [{"url": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management", "access": "full text reviewed", "title": "Halliburton and Shape Digital form strategic collaboration to advance digital asset performance management", "publication_date": "2026-07-21", "claims_checked": ["AI-018 source claim(s) checked against cited profile"], "notes": "Halliburton/Shape collaboration release opened; future strategic collaboration, no customer deployment or achieved result.", "partition": 2}]}, {"Source_ID": "RS064", "Source_date": "2025-05-28", "Company": "C3 AI", "Source_type": "company announcement", "Title": "C3 AI and Baker Hughes Renew and Expand Joint Venture Agreement", "URL": "https://ir.c3.ai/node/10916/pdf", "Locator": "Agreement scope", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-05-28", "status": "full text reviewed", "locator": "Agreement scope"}], "Audit_reviews": [{"url": "https://ir.c3.ai/node/10916/pdf", "title": "Baker Hughes–C3 AI enterprise AI joint venture renewal", "publication_date": "2025-05-28", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["C3 AI and Baker Hughes renewed and expanded their joint venture agreement for multiple years through June 2028."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS065", "Source_date": "2026-06-09", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "URL": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "Locator": "Collaboration scope", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-06-09", "status": "full text reviewed", "locator": "Collaboration scope"}], "Audit_reviews": [{"url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "access": "full text reviewed", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "publication_date": "2026-06-09", "claims_checked": ["AI-046 source claim(s) checked against cited profile"], "notes": "SLB Agora/Qualcomm release opened; edge-AI product/collaboration scope checked, no named deployed customer.", "partition": 2}, {"url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "access": "full text reviewed", "claims_checked": ["SLB/Qualcomm edge-AI collaboration and product/workflow development scope."], "notes": "Same exact initiative and announcement as assigned existing AI-046; a second addition would duplicate the existing record.", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "publication_date": "2026-06-09"}]}, {"Source_ID": "RS066", "Source_date": "2026-06-18", "Company": "SLB", "Source_type": "company feature", "Title": "Inside the Future of Production Operations", "URL": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio", "Locator": "Autonomous pad and visitor details", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-06-18", "status": "full text reviewed", "locator": "Autonomous pad and visitor details"}], "Audit_reviews": [{"url": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio", "access": "full text reviewed", "title": "Inside the Future of Production Operations", "publication_date": "2026-06-18", "claims_checked": ["AI-048 source claim(s) checked against cited profile"], "notes": "SLB Intelligent Production Studio feature opened; product workflow descriptions checked; case results require careful attribution to SLB digital suite.", "partition": 2}]}, {"Source_ID": "RS067", "Source_date": "Undated; accessed 2026-09-27", "Company": "SLB", "Source_type": "company product documentation", "Title": "Intelligent Production Studio", "URL": "https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "Locator": "Inside the studio and FAQ", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "Undated; accessed 2026-09-27", "status": "full text reviewed", "locator": "Inside the studio and FAQ"}], "Audit_reviews": [{"url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "access": "full text reviewed", "title": "Intelligent Production Studio", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-048 source claim(s) checked against cited profile"], "notes": "SLB product page opened; current capability is vendor offer, page is undated.", "partition": 2}]}, {"Source_ID": "RS068", "Source_date": "2026-03-25", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Industrializes AI for the Energy Industry with NVIDIA", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/", "Locator": "Three strategic elements", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-03-25", "status": "full text reviewed", "locator": "Three strategic elements"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/", "access": "full text reviewed", "title": "SLB Industrializes AI for the Energy Industry with NVIDIA", "publication_date": "2026-03-25", "claims_checked": ["AI-105 source claim(s) checked against cited profile"], "notes": "SLB/NVIDIA AI Factory announcement opened; commercialization partnership/product roadmap, not upstream AI deployment.", "partition": 2}]}, {"Source_ID": "RS069", "Source_date": "2026-05-29", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB to Acquire Tachyus", "URL": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update", "Locator": "Aqueon workflows, deployment and planned integration", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-05-29", "status": "full text reviewed", "locator": "Aqueon workflows, deployment and planned integration"}], "Audit_reviews": [{"url": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update", "access": "full text reviewed", "title": "SLB to Acquire Tachyus", "publication_date": "2026-05-29", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "SLB Tachyus acquisition update opened; agreement/transaction timing checked, no purchase price or close date inferred.", "partition": 2}]}, {"Source_ID": "RS070", "Source_date": "2026-07-24", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Announces Second-Quarter 2026 Results", "URL": "https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "Locator": "Acquisition completion", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-07-24", "status": "full text reviewed", "locator": "Acquisition completion"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "access": "full text reviewed", "title": "SLB Announces Second-Quarter 2026 Results", "publication_date": "2026-07-24", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "SLB Q2 2026 earnings Form 8-K opened; portfolio and transaction references checked.", "partition": 2}]}, {"Source_ID": "RS071", "Source_date": "2026-04-24", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Advances Digital Strategy with Acquisition of S&P Global Energy’s Upstream Software Portfolio", "URL": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition", "Locator": "Transaction scope and planned AI collaboration", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-04-24", "status": "full text reviewed", "locator": "Transaction scope and planned AI collaboration"}], "Audit_reviews": [{"url": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition", "access": "full text reviewed", "title": "SLB Advances Digital Strategy with Acquisition of S&P Global Energy’s Upstream Software Portfolio", "publication_date": "2026-04-24", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "SLB S&P acquisition release opened; agreement distinguished from completion.", "partition": 2}]}, {"Source_ID": "RS072", "Source_date": "2026-09-01", "Company": "S&P Global", "Source_type": "company announcement", "Title": "S&P Global Completes Divestiture of Upstream Energy Software Portfolio to SLB", "URL": "https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "Locator": "Acquisition completion", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-09-01", "status": "full text reviewed", "locator": "Acquisition completion"}], "Audit_reviews": [{"url": "https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "access": "full text reviewed", "title": "S&P Global Completes Divestiture of Upstream Energy Software Portfolio to SLB", "publication_date": "2026-09-01", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "S&P Global completion release opened; completion date checked.", "partition": 2}]}, {"Source_ID": "RS073", "Source_date": "2026-04-01", "Company": "Halliburton", "Source_type": "company announcement", "Title": "Halliburton Acquires Sekal AS From Sumitomo Corporation to Accelerate Drilling Automation Services", "URL": "https://ir.halliburton.com/node/36266/pdf", "Locator": "Acquisition and portfolio combination", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-04-01", "status": "full text reviewed", "locator": "Acquisition and portfolio combination"}], "Audit_reviews": [{"url": "https://ir.halliburton.com/node/36266/pdf", "access": "full text reviewed", "title": "Halliburton Acquires Sekal AS From Sumitomo Corporation to Accelerate Drilling Automation Services", "publication_date": "2026-04-01", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Halliburton Sekal acquisition release opened; acquisition closed/disclosed date distinct from product automation claims.", "partition": 2}]}, {"Source_ID": "RS074", "Source_date": "2025-02-26", "Company": "Halliburton", "Source_type": "partner case study", "Title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "URL": "https://ir.halliburton.com/node/35181/pdf", "Locator": "Integrated system and Equinor deployment", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2025-02-26", "status": "full text reviewed", "locator": "Integrated system and Equinor deployment"}], "Audit_reviews": [{"url": "https://ir.halliburton.com/node/35181/pdf", "access": "full text reviewed", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "publication_date": "2025-02-26", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Halliburton/Sekal Equinor case PDF opened in full; Feb 2025 on-bottom system, integration, autonomy and safety limits checked; AI itself is not explicit.", "partition": 2}]}, {"Source_ID": "RS075", "Source_date": "2026-06-30", "Company": "SLB", "Source_type": "company announcement", "Title": "SLB Awarded Seven-Year Contract Under Kuwait Oil Company’s Ahmadi Innovation Valley Initiative", "URL": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1", "Locator": "Contract scope, key points", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "services", "date": "2026-06-30", "status": "full text reviewed", "locator": "Contract scope, key points"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1", "access": "full text reviewed", "title": "SLB Awarded Seven-Year Contract Under Kuwait Oil Company’s Ahmadi Innovation Valley Initiative", "publication_date": "2026-06-30", "claims_checked": ["KOC awarded SLB a seven-year AIV contract to evaluate/test/deploy technologies; scope spans nearly 100 projects across AI, IIoT, production optimization, reservoirs, water and energy transition.", "A dedicated facility is planned, with construction expected to start 2026 and opening in 2028; the announcement names no specific AI workflow or outcome."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-06-30; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS076", "Source_date": "2025-05-02", "Company": "Quantum Capital Group", "Source_type": "annual report", "Title": "2024 ESG Report", "URL": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "Locator": "AI/data infrastructure section (search excerpt; PDF open failed)", "Short_excerpt": "", "Verification": "Document accessible on 29 September 2026. New review: Printed/PDF page 91, Sentinel Peak Resources. Earlier claim-level findings remain scoped to their original review.", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-29", "Research_reviews": [{"cohort": "software_capital", "date": "2025-05-02", "status": "excerpt reviewed", "locator": "AI/data infrastructure section (search excerpt; PDF open failed)"}], "Audit_reviews": [{"url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "title": "2024 ESG Report", "publication_date": "2025-05-02", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}], "Followup_source_review": {"date": "2026-09-29", "status": "Latest access review: original document available; Printed/PDF page 91, Sentinel Peak Resources", "note": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained."}}, {"Source_ID": "RS077", "Source_date": "2025-09-10", "Company": "Novi Labs", "Source_type": "company announcement", "Title": "Novi Labs unveils Novi Intelligence", "URL": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "Locator": "Novi Intelligence product", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-09-10", "status": "full text reviewed", "locator": "Novi Intelligence product"}], "Audit_reviews": [{"url": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "access": "full text reviewed", "title": "Novi Labs unveils Novi Intelligence", "publication_date": "2025-09-10", "claims_checked": ["Novi Intelligence joins operator-sourced data, machine learning and forecasts with proprietary metrics such as rock quality, completion intensity, depletion and well density.", "Product exposes underlying data for client validation; accuracy claims are vendor claims."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-09-10; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS078", "Source_date": "2026-04-29", "Company": "Novi Labs", "Source_type": "company announcement", "Title": "Novi Labs acquires Turing Analytics", "URL": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "Locator": "Acquisition rationale and geographic coverage", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-04-29", "status": "excerpt reviewed", "locator": "Acquisition rationale and geographic coverage"}], "Audit_reviews": [{"url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "access": "full text reviewed", "title": "Novi Labs acquires Turing Analytics", "publication_date": "2026-04-29", "claims_checked": ["Novi acquired Turing Analytics/EVA to combine US and Canadian upstream datasets and analytics for investment evaluation.", "Novi says Turing clients represent over 50% of Canadian oil and gas production; company-reported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-04-29; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS079", "Source_date": "2026-05-20", "Company": "PakEnergy", "Source_type": "company announcement", "Title": "PakEnergy launches AI-powered Ticket-N-Receipt Management", "URL": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "Locator": "PakCAPTURE workflow", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-05-20", "status": "full text reviewed", "locator": "PakCAPTURE workflow"}], "Audit_reviews": [{"url": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "title": "PakCAPTURE ticket and receipt extraction", "publication_date": "2026-05-20", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Drivers photograph receipts or paper tickets; PakAI extracts data, populates pickup/drop forms and associates information with a load.", "Feature embedded in PakTransportation and available to customers."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS080", "Source_date": "undated", "Company": "Collide", "Source_type": "company webpage", "Title": "Collide product: Operations System", "URL": "https://collide.io/product", "Locator": "Architecture/capability page", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "undated", "status": "excerpt reviewed", "locator": "Architecture/capability page"}], "Audit_reviews": [{"url": "https://collide.io/product", "title": "Collide product: Operations System", "publication_date": "undated", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Collide describes petroleum-specific connected asset model and grounded answers with source citations.", "Company describes deterministic tested calculations and deployment in customer cloud/data platform; page undated."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS081", "Source_date": "2025-05-28", "Company": "Applied Computing", "Source_type": "company announcement", "Title": "Applied Computing raises £9m seed to build Orbital AI", "URL": "https://appliedcomputing.com/press-release-28052025", "Locator": "Orbital architecture, financing and initial refinery", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-05-28", "status": "full text reviewed", "locator": "Orbital architecture, financing and initial refinery"}], "Audit_reviews": [{"url": "https://appliedcomputing.com/press-release-28052025", "access": "full text reviewed", "title": "Applied Computing raises £9m seed to build Orbital AI", "publication_date": "2025-05-28", "claims_checked": ["Applied Computing raised £9m seed funding for Orbital, a process-industry model using language, time-series, physics and chemical engineering models.", "Company says it is on edge, can work on sensitive process data, and reports one large refinery live.", "Cost savings, energy reduction and payback cited are company claims/potential, not independently supported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS082", "Source_date": "2025-09-09", "Company": "Cognite", "Source_type": "company announcement", "Title": "Cognite Atlas AI major release accelerates industrial AI", "URL": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "Locator": "Agent workbench, templates, governance", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-09-09", "status": "full text reviewed", "locator": "Agent workbench, templates, governance"}], "Audit_reviews": [{"url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "access": "full text reviewed", "title": "Cognite Atlas AI major release accelerates industrial AI", "publication_date": "2025-09-09", "claims_checked": ["AI-120 source claim(s) checked against cited profile"], "notes": "Cognite Atlas AI product release opened; product architecture/workbench claims checked; not proof of customer deployment.", "partition": 2}]}, {"Source_ID": "RS083", "Source_date": "2026-03-03", "Company": "Seeq", "Source_type": "company announcement", "Title": "Introducing Seeq Intelligence", "URL": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "Locator": "Agent Q, build-your-own agents and integrations", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-03-03", "status": "full text reviewed", "locator": "Agent Q, build-your-own agents and integrations"}], "Audit_reviews": [{"url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "access": "full text reviewed", "title": "Introducing Seeq Intelligence", "publication_date": "2026-03-03", "claims_checked": ["AI-128 source claim(s) checked against cited profile"], "notes": "Seeq Intelligence release opened; product/Agent Q scope checked; no named O&G deployment or outcome.", "partition": 2}]}, {"Source_ID": "RS084", "Source_date": "undated; CY2025 economics", "Company": "C3 AI", "Source_type": "partner case study", "Title": "Baker Hughes achieves bottom line savings with AI Sourcing Optimization", "URL": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "Locator": "AI sourcing optimization; value and scale", "Short_excerpt": "", "Verification": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "Notes": "Research expansion. Company and vendor claims are not independently validated. Follow-up source review on 29 September 2026; performance remains source-attributed.", "Accessed": "2026-09-29", "Research_reviews": [{"cohort": "software_capital", "date": "undated; CY2025 economics", "status": "excerpt reviewed", "locator": "AI sourcing optimization; value and scale"}], "Audit_reviews": [{"url": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "title": "AI sourcing optimization across procurement data", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Case study says C3 AI Sourcing Optimization supported 900,000 SKUs across 800+ sites and realized millions in economic benefit in CY2025.", "Workflow addresses fragmented procurement data and manual spreadsheet reconciliation.", "Vendor-published customer story; not independent assurance."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS085", "Source_date": "2025-04-29", "Company": "Microsoft", "Source_type": "partner case study", "Title": "Encino Energy streamlines operations with Microsoft Power Platform", "URL": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "Locator": "Copilot Studio/data agent and invoice AI Builder details", "Short_excerpt": "", "Verification": "Document accessible on 1 October 2026. New review: Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83.. Earlier claim-level findings remain scoped to their original review.", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-10-01", "Research_reviews": [{"cohort": "software_capital", "date": "2025-04-29", "status": "full text reviewed", "locator": "Copilot Studio/data agent and invoice AI Builder details"}], "Audit_reviews": [{"url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "access": "full text reviewed", "title": "Encino Energy streamlines operations with Microsoft Power Platform", "publication_date": "2025-04-29", "claims_checked": ["AI-166 source claim(s) checked against cited profile"], "notes": "Microsoft Encino customer case opened; Power Platform process workflows reviewed; does not establish AI as the enabling technology.", "partition": 2}], "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83.", "note": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study."}}, {"Source_ID": "RS086", "Source_date": "2025-10-22", "Company": "Microsoft", "Source_type": "partner case study", "Title": "TotalEnergies from Microsoft 365 Copilot to Copilot Studio", "URL": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "Locator": "MARGE and BuyerCompanion", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-10-22", "status": "full text reviewed", "locator": "MARGE and BuyerCompanion"}], "Audit_reviews": [{"url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "title": "MARGE maintenance report classification and BuyerCompanion procurement agent", "publication_date": "2025-10-22", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["MARGE uses Copilot Studio to classify intervention reports; company reports 25-45% improvement in fault-cause identification.", "BuyerCompanion streamlines purchases up to €50,000, operational in France, with company target of 10% savings on certain purchases.", "30,000 Copilot licenses, training and internal agent-validation committee described."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS087", "Source_date": "2025-08-05", "Company": "Ambyint", "Source_type": "partner case study", "Title": "Infinity PL drives efficiency gains for Encino Energy", "URL": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "Locator": "InfinityPL customer deployment", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-08-05", "status": "full text reviewed", "locator": "Customer background; solutions; key outcomes"}], "Audit_reviews": [{"url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "access": "full text reviewed", "title": "Infinity PL drives efficiency gains for Encino Energy", "publication_date": "2025-08-05", "claims_checked": ["AI-168 source claim(s) checked against cited profile"], "notes": "Ambyint vendor case opened; vendor-reported InfinityPL lift optimization/results, no independent operator corroboration.", "partition": 2}, {"url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "title": "Infinity PL Drives Efficiency Gains for Encino Energy", "publication_date": "2025-08-05", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Page identifies Encino, Appalachian Basin, 344 platform wells and project timeframe 2023 to present.", "Page describes real-time surveillance, anomaly detection, autonomous setpoint control, dashboard monitoring and operator intervention.", "Page reports decline, setpoint and venting metrics, but does not show cohort sizes or validation method; outcomes remain vendor-reported."], "notes": "Direct page open succeeded in this pass; full case text and outcome section reviewed. This is vendor case evidence, not independent validation.", "partition": 3}]}, {"Source_ID": "RS088", "Source_date": "2025-02-19", "Company": "Imubit", "Source_type": "company announcement", "Title": "Preem partners with Imubit for closed-loop AI optimization", "URL": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "Locator": "Lysekil FCC deployment", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2025-02-19", "status": "full text reviewed", "locator": "Lysekil FCC deployment"}], "Audit_reviews": [{"url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "access": "full text reviewed", "title": "Preem partners with Imubit for closed-loop AI optimization", "publication_date": "2025-02-19", "claims_checked": ["AI-169 source claim(s) checked against cited profile"], "notes": "Imubit/Preem announcement opened; says Optimizing Brain will be deployed, so outcomes are expected, not achieved.", "partition": 2}]}, {"Source_ID": "RS089", "Source_date": "2026-03-05", "Company": "Winn Resources (issuer release syndicated by PRNewswire/StreetInsider)", "Source_type": "company announcement", "Title": "Winn Resources partners with Collide to turn operational data into a competitive advantage", "URL": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html", "Locator": "W-10/G-10 filing automation", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "2026-03-05", "status": "full text reviewed", "locator": "W-10/G-10 filing automation"}], "Audit_reviews": [{"url": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html", "title": "Winn Resources Texas RRC W-10/G-10 filing workflow", "publication_date": "2026-03-05", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Winn says Collide automates monthly Texas RRC W-10/G-10 filings using existing well test/production data; task moves from multi-hour effort to under 30 minutes.", "Winn says workflow surfaces missing-data gaps earlier and reports >95% time reduction. This is customer-reported; no independent audit."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS090", "Source_date": "undated", "Company": "Collide", "Source_type": "partner case study", "Title": "How Winn Resources Slashed Regulatory Filing Time by 95%", "URL": "https://collide.io/case-studies/winn-resources-ai-regulatory-filing-oil-gas", "Locator": "Case-study page title; body extraction unavailable", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "software_capital", "date": "undated", "status": "excerpt reviewed", "locator": "Case-study page title; body extraction unavailable"}], "Audit_reviews": [{"url": "https://collide.io/case-studies/winn-resources-ai-regulatory-filing-oil-gas", "title": "How Winn Resources Slashed Regulatory Filing Time by 95%", "publication_date": "undated", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS091", "Source_date": "2025-08-05", "Company": "Williams", "Source_type": "earnings call transcript", "Title": "Williams Q2 2025 earnings call transcript", "URL": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-08-05", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "access": "full text reviewed", "title": "Williams Q2 2025 earnings call transcript", "publication_date": "2025-08-05", "claims_checked": ["Sequent gas-marketing AI anecdote; CEO describes one-month DFW test, model score 96% vs trader 94% against a 100% hindsight benchmark; narrow test only."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS092", "Source_date": "2025-08-05", "Company": "Williams", "Source_type": "investor relations event page", "Title": "Williams Q2 2025 earnings event page", "URL": "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-08-05", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025", "access": "full text reviewed", "title": "Williams Q2 2025 earnings event page", "publication_date": "2025-08-05", "claims_checked": ["Confirms issuer-hosted earnings materials including transcript and presentation."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS093", "Source_date": "2025", "Company": "TC Energy", "Source_type": "annual report", "Title": "TC Energy 2025 Report on Sustainability", "URL": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "access": "inaccessible", "title": "TC Energy 2025 Report on Sustainability", "publication_date": "2025", "claims_checked": ["Autonomous Pipeline analytics and machine learning for commercial daily/weekly asset strategies and transport/PAL opportunities; RAMP described as advanced algorithms only."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS094", "Source_date": "2026-03-05", "Company": "TC Energy", "Source_type": "proxy statement", "Title": "TC Energy 2026 Management Information Circular", "URL": "https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026-03-05", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf", "access": "full text reviewed", "title": "TC Energy 2026 Management Information Circular", "publication_date": "2026-03-05", "claims_checked": ["AI governance, training, board oversight; governance does not evidence operating deployment."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-03-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS095", "Source_date": "Undated source; see chronology", "Company": "Amazon Web Services", "Source_type": "partner case study", "Title": "TC Energy Pipeline IQ on AWS", "URL": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "Locator": "PIPER, SCOUT and outcome sections", "Short_excerpt": "", "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-10-06", "Research_reviews": [{"cohort": "midstream", "date": "undated", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "access": "full text reviewed", "title": "TC Energy Pipeline IQ on AWS", "publication_date": "undated", "claims_checked": ["PIPER reviews engineering documentation/material test reports against weld maps and pressure-test docs; SCOUT discovery; >100 users, 85% review-time reduction, expected $18.7m five-year maintenance-capital avoidance. Page is undated, date/start of deployment not established."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: undated; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS096", "Source_date": "2025", "Company": "Enbridge", "Source_type": "sustainability report", "Title": "Enbridge 2025 Sustainability Report", "URL": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "access": "inaccessible", "title": "Enbridge 2025 Sustainability Report", "publication_date": "2025", "claims_checked": ["Describes developing machine-learning prioritization of potentially injurious deformation features from in-line inspection and field data; engineers assess and mitigate priority features."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS097", "Source_date": "2026", "Company": "Enbridge", "Source_type": "company technical publication", "Title": "Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability", "URL": "https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "access": "full text reviewed", "title": "Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability", "publication_date": "2026", "claims_checked": ["States ML predictive analysis in complex pipeline defect assessment to prioritize repairs and allocate resources; no quantified outcome."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS098", "Source_date": "2026", "Company": "Enbridge", "Source_type": "company policy", "Title": "Enbridge 2026 AI Policy", "URL": "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152", "access": "full text reviewed", "title": "Enbridge 2026 AI Policy", "publication_date": "2026", "claims_checked": ["Company policy describes AI governance and general AI uses; does not by itself date or quantify named operating workflow."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS099", "Source_date": "undated", "Company": "ONEOK", "Source_type": "company sustainability webpage", "Title": "ONEOK Innovation Spotlight", "URL": "https://sustainability.oneok.com/innovation-spotlight", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "undated", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://sustainability.oneok.com/innovation-spotlight", "title": "AI virtual river gauges for extreme-weather response", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reports AI-powered virtual river gauges and 48,872 deployed; weather-response risk triage. Undated page does not establish launch date or confirm in-window deployment."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS100", "Source_date": "2026", "Company": "ONEOK", "Source_type": "earnings release", "Title": "ONEOK Q4 2025 earnings release", "URL": "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638", "title": "ONEOK Q4 2025 earnings release", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reviewed alongside company materials; no specific AI workflow found in release."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS101", "Source_date": "2025-10-23", "Company": "Valero", "Source_type": "investor relations event page", "Title": "Valero Q3 2025 earnings call event", "URL": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-10-23", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "access": "full text reviewed", "title": "Valero Q3 2025 earnings call event", "publication_date": "2025-10-23", "claims_checked": ["AI-102 source claim(s) checked against cited profile"], "notes": "Valero issuer earnings event page opened; event metadata only, not transcript content.", "partition": 2}]}, {"Source_ID": "RS102", "Source_date": "2025-01-14", "Company": "Flyscan Systems", "Source_type": "partner announcement", "Title": "Flyscan announces Kinder Morgan as strategic investor", "URL": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-01-14", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "title": "Strategic investment in Flyscan AI-assisted aerial inspection", "publication_date": "2025-04-16", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["KMI joins strategic investor group; vendor describes remote sensing and AI visual patrol/leak detection. This confirms investment/partnership, not KMI rollout."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS103", "Source_date": "2025-04-16", "Company": "Kinder Morgan", "Source_type": "earnings release", "Title": "Kinder Morgan Q1 2025 earnings release", "URL": "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-04-16", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx", "title": "Kinder Morgan Q1 2025 earnings release", "publication_date": "2025-04-16", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reviewed as in-window operating/financial source; no Flyscan or internal AI deployment disclosed in this release."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS104", "Source_date": "2025-11-04", "Company": "Energy Transfer", "Source_type": "company announcement", "Title": "Entergy Louisiana and Energy Transfer sign agreement supporting Meta data center", "URL": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-11-04", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "title": "Long-term gas transportation for Meta Louisiana data center", "publication_date": "2025-11-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Long-term gas transportation infrastructure for future Meta site; separate AI-demand category, not internal AI."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS105", "Source_date": "2026", "Company": "Energy Transfer", "Source_type": "earnings release", "Title": "Energy Transfer Q1 2026 results and updates", "URL": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "title": "Nexus Hubbard Campus gas transport for behind-the-meter AI campus", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reports agreements to transport gas to Nexus Hubbard Campus, an AI data center campus with on-site gas generation; infrastructure supporting AI demand."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS106", "Source_date": "2025-08-04", "Company": "Williams", "Source_type": "earnings release", "Title": "Williams Q2 2025 results", "URL": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-08-04", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "access": "full text reviewed", "title": "Williams Q2 2025 results", "publication_date": "2025-08-04", "claims_checked": ["Describes $1.6bn Socrates power innovation project to serve growing AI demand; infrastructure, not internal AI."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS107", "Source_date": "2025-07-30", "Company": "Bridgepoint Group / ECP", "Source_type": "company announcement", "Title": "ECP and KKR announce Texas hyperscale data-center campus", "URL": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2025-07-30", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "access": "full text reviewed", "title": "ECP and KKR announce Texas hyperscale data-center campus", "publication_date": "2025-07-30", "claims_checked": ["AI-059 source claim(s) checked against cited profile"], "notes": "ECP/KKR campus announcement opened; AI data-center demand infrastructure, not oil/gas company AI adoption.", "partition": 2}]}, {"Source_ID": "RS108", "Source_date": "2026-05-07", "Company": "Targa Resources", "Source_type": "job posting", "Title": "Targa Principal Engineer, AI & Emerging Technology job posting", "URL": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "Locator": null, "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "midstream", "date": "2026-05-07", "status": "full text reviewed", "locator": null}], "Audit_reviews": [{"url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "access": "full text reviewed", "title": "Targa Principal Engineer, AI & Emerging Technology job posting", "publication_date": "2026-05-07", "claims_checked": ["AI-176 source claim(s) checked against cited profile"], "notes": "Targa job posting opened; establishes hiring intent/role, not deployed operational AI.", "partition": 2}]}, {"Source_ID": "RS109", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Investor presentation", "Title": "Saudi Aramco FY 2025 Results presentation", "URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "Locator": "“Pioneering AI & technology: progressing value-creation opportunities,” slide 10 / PDF p. 10; portfolio TRV value-type mix (not allocated to AI)", "Short_excerpt": "", "Verification": "Full primary text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated. 6 October 2026: earlier source entry linked the companion text script, which does not contain the value-type chart. Corrected to the original presentation; prior URL retained in Original_URL.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2026-03-10", "status": "Full primary text reviewed", "locator": "Slide 10; Technology Realized Value"}], "Audit_reviews": [{"url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "access": "full text reviewed", "title": "FY 2025 results presentation and script", "publication_date": "2026-03-10", "claims_checked": ["AI-037: $5.3bn total 2025 TRV; CEO says over half enabled by AI; mixed-technology components are not an AI-only breakdown."], "notes": "Primary Aramco FY2025 remarks/presentation reviewed. CEO remarks explicitly state total TRV was $5.3bn, with over half enabled by AI; this is high-level company attribution, not independently validated project-level value.", "partition": 2}], "Original_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "Correction_note": "6 October 2026: earlier source entry linked the companion text script, which does not contain the value-type chart. Corrected to the original presentation; prior URL retained in Original_URL.", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Official operator FY2025 results presentation independently reviewed", "publisher": "Aramco", "locator": "“Pioneering AI & technology: progressing value-creation opportunities,” slide 10 / PDF p. 10; portfolio TRV value-type mix (not allocated to AI)"}, "Publisher": "Aramco"}, {"Source_ID": "RS110", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Company announcement / remarks", "Title": "Aramco announces fourth quarter and full-year 2025 results", "URL": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "Locator": "HUMAIN investment plans; TRV highlights", "Short_excerpt": "", "Verification": "Full primary text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2026-03-10", "status": "Full primary text reviewed", "locator": "HUMAIN investment plans; TRV highlights"}], "Audit_reviews": [{"url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "access": "full text reviewed", "title": "Aramco announces fourth quarter and full-year 2025 results", "publication_date": "2026-03-10", "claims_checked": ["AI-039 source claim(s) checked against cited profile"], "notes": "Aramco FY2025 result release opened; company financial/AI statements checked.", "partition": 2}]}, {"Source_ID": "RS111", "Source_date": "2025-02-09", "Company": "Aramco", "Source_type": "Company announcement / remarks", "Title": "Remarks by Ahmad O. Al-Khowaiter at LEAP 2025", "URL": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "Locator": "MetaBrain and MetaBrain for Plants paragraphs", "Short_excerpt": "", "Verification": "Full primary text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "root_backfill", "date": "2025-02-09", "status": "Full primary text reviewed", "locator": "MetaBrain and MetaBrain for Plants paragraphs"}], "Audit_reviews": [{"url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "access": "full text reviewed", "title": "Remarks by Ahmad O. Al-Khowaiter at LEAP 2025", "publication_date": "2025-02-09", "claims_checked": ["AI-177 source claim(s) checked against cited profile"], "notes": "Aramco LEAP remarks opened; MetaBrain for Plants described with future tense; workflow not treated as achieved plant operation.", "partition": 2}]}, {"Source_ID": "RS112", "Source_date": "2025-10-23", "Company": "ExxonMobil", "Source_type": "company webpage", "Title": "ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence", "URL": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "Locator": "Production bottleneck monitoring", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-10-23", "status": "full text reviewed", "locator": "Production bottleneck monitoring"}], "Audit_reviews": [{"url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "access": "full text reviewed", "title": "ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence", "publication_date": "2025-10-23", "claims_checked": ["ExxonMobil India describes using AI with production staff to monitor equipment in real time and identify current or prospective bottlenecks affecting production.", "Engineers and asset operators use domain knowledge to investigate and remove bottlenecks; no basin/site counts or metrics."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-10-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS113", "Source_date": "2025-06-10", "Company": "PETRONAS", "Source_type": "company announcement", "Title": "PETRONAS-incubated AIngineer joins forces with Microsoft", "URL": "https://www.petronas.com/bm/node/1668", "Locator": "AIngineer industrial engineer copilot", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-06-10", "status": "full text reviewed", "locator": "AIngineer industrial engineer copilot"}], "Audit_reviews": [{"url": "https://www.petronas.com/bm/node/1668", "title": "Industrial AI copilot integrated with engineering technical references", "publication_date": "2025-06-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["AIngineer is a PETRONAS-incubated industrial AI copilot using LLMs and ML and integrating technical databases and standards.", "PETRONAS says it has technical use cases and prior deployments in energy/government; this release does not name a customer asset or quantify results.", "Microsoft collaboration includes cloud scaling and prospective marketplace commercialization."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS114", "Source_date": "June/July 2025 listing; exact publication date unavailable", "Company": "Xecta", "Source_type": "company webpage", "Title": "Xecta news and resources index", "URL": "https://xecta.com/news-resources/", "Locator": "Santos Eastern Australia and PNG digital oilfield listing", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "June/July 2025 listing; exact publication date unavailable", "status": "excerpt reviewed", "locator": "Santos Eastern Australia and PNG digital oilfield listing"}], "Audit_reviews": [{"url": "https://xecta.com/news-resources/", "access": "excerpt reviewed", "title": "Xecta news and resources index", "publication_date": "June/July 2025 listing; exact publication date unavailable", "claims_checked": ["Xecta index identifies Santos integrated production-system model/digital oilfield across Eastern Australia and PNG.", "Underlying dated technical release did not open; current confirmation limited to vendor page index."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: June/July 2025 listing; exact publication date unavailable; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS115", "Source_date": "2025-06-12", "Company": "TotalEnergies", "Source_type": "company announcement", "Title": "TotalEnergies to collaborate with Mistral AI", "URL": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "Locator": "Digital Factory baseline and joint AI lab", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "majors_backfill", "date": "2025-06-12", "status": "full text reviewed", "locator": "Digital Factory baseline and joint AI lab"}], "Audit_reviews": [{"url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "title": "Digital Factory portfolio of AI and digital solutions", "publication_date": "2025-06-12", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["TotalEnergies says its Digital Factory had 300 digital experts and developed >100 solutions, 60 using technologies from ML to GenAI.", "The announcement creates a distinct Mistral joint innovation lab for researcher assistant, industrial asset decision support, customer and renewable use cases; it says those use cases will be designed/tested, not deployed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}]}, {"Source_ID": "RS116", "Source_date": "2025 (FY2024-25)", "Company": "Indian Oil Corporation", "Source_type": "annual report webpage", "Title": "IndianOil – Intellectual Capital, Integrated Annual Report 2024-25", "URL": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "Locator": "Leveraging Data and AI for Operational Efficiency", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2025 (FY2024-25)", "status": "excerpt reviewed", "locator": "Leveraging Data and AI for Operational Efficiency"}], "Audit_reviews": [{"url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "access": "full text reviewed", "title": "IndianOil – Intellectual Capital, Integrated Annual Report 2024-25", "publication_date": "2025 (FY2024-25)", "claims_checked": ["AI-180 source claim(s) checked against cited profile"], "notes": "IndianOil FY2024-25 Intellectual Capital page opened; broad AI/ML predictive-maintenance statement checked; no sites, asset types or outcomes disclosed.", "partition": 2}]}, {"Source_ID": "RS117", "Source_date": "2025 (FY2024-25)", "Company": "Bharat Petroleum", "Source_type": "annual report", "Title": "Annual Report 2024-25", "URL": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "Locator": "Digital initiatives, p. 135", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2025 (FY2024-25)", "status": "excerpt reviewed", "locator": "Digital initiatives, p. 135"}], "Audit_reviews": [{"url": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "access": "inaccessible", "title": "Annual Report 2024-25", "publication_date": "2025 (FY2024-25)", "claims_checked": ["AI predictive analytics for asset reliability/performance was initiated across Mumbai, Kochi and Bina refineries, extending a successful Kochi trial; cloud predictive alerts target failures/degradation in identified critical equipment.", "AI video algorithms detect non-compliance and safety hazards from field video; Ultra Critical Video Analytics was implemented across all three refineries and integrated with the in-house Work Permit System."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025 (FY2024-25); baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS118", "Source_date": "2026-01-05", "Company": "Hindustan Petroleum", "Source_type": "company announcement", "Title": "HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery", "URL": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "Locator": "paragraphs 24-28", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2026-01-05", "status": "full text reviewed", "locator": "paragraphs 24-28"}], "Audit_reviews": [{"url": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "access": "full text reviewed", "title": "HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery", "publication_date": "2026-01-05", "claims_checked": ["At commissioning of the Visakh residue-upgradation facility, HPCL said its LC-Max Digital Suite integrates real-time monitoring, predictive analytics using proprietary thermodynamic models and AI-driven optimization.", "The source reports RUF physical-process conversion/yield figures separately; it does not attribute those outcomes to AI."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS119", "Source_date": "2026-06", "Company": "Reliance Industries", "Source_type": "shareholder statement", "Title": "Chairman’s Statement, 49th Annual General Meeting", "URL": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "Locator": "O2C business performance", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2026-06", "status": "excerpt reviewed", "locator": "O2C business performance"}], "Audit_reviews": [{"url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "access": "full text reviewed", "title": "Chairman’s Statement, 49th Annual General Meeting", "publication_date": "2026-06", "claims_checked": ["AI-183 source claim(s) checked against cited profile"], "notes": "RIL chairman statement PDF opened; crude-feedstock AI optimizer statement checked; no quantified result, scale or method disclosed.", "partition": 2}]}, {"Source_ID": "RS120", "Source_date": "2026-07-20", "Company": "PTT Global Chemical", "Source_type": "company announcement", "Title": "GC Advances Holistic Optimization Through Digital Technology", "URL": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "Locator": "Smart Plant / Supply Chain & Operations Optimization", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2026-07-20", "status": "full text reviewed", "locator": "Smart Plant / Supply Chain & Operations Optimization"}, {"cohort": "regional", "date": "2026-07-20", "status": "full text reviewed", "locator": "Smart Plant / Supply Chain & Operations Optimization"}], "Audit_reviews": [{"url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "access": "full text reviewed", "title": "GC Advances Holistic Optimization Through Digital Technology", "publication_date": "2026-07-20", "claims_checked": ["GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale.", "GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-20; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS121", "Source_date": "2025-07-17", "Company": "Rosneft", "Source_type": "company announcement", "Title": "Rosneft improves well research methods (EchoTools)", "URL": "https://rosneft.ru/press/news/item/222577", "Locator": "EchoTools", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2025-07-17", "status": "excerpt reviewed", "locator": "EchoTools"}], "Audit_reviews": [{"url": "https://rosneft.ru/press/news/item/222577", "access": "inaccessible", "title": "Rosneft improves well research methods (EchoTools)", "publication_date": "2025-07-17", "claims_checked": ["AI interprets fluid level and sound speed in the well annulus; the derived values support bottomhole-pressure calculation and selection of an operating mode.", "Rosneft says the software processes over 10,000 well indicators in about two minutes versus over three days manually, and recommendations yield an average additional 1 tonne/day per well."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-07-17; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS122", "Source_date": "2026-08-20", "Company": "Rosneft", "Source_type": "company announcement", "Title": "Rosneft increases production through innovative software (EchoTools update)", "URL": "https://www.rosneft.ru/press/news/item/224545/", "Locator": "EchoTools rollout update", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2026-08-20", "status": "excerpt reviewed", "locator": "EchoTools rollout update"}], "Audit_reviews": [{"url": "https://www.rosneft.ru/press/news/item/224545/", "access": "inaccessible", "title": "Rosneft increases production through innovative software (EchoTools update)", "publication_date": "2026-08-20", "claims_checked": ["Rosneft reports that in 1H 2026 EchoTools analyzed information from more than 2,000 wells and generated a ranking of over 500 candidate wells for operating-mode optimization.", "This is a later rollout update to the July 2025 named workflow, not its launch date."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-08-20; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS123", "Source_date": "2025-07-28", "Company": "Rosneft", "Source_type": "company announcement", "Title": "Rosneft introduces new technologies using artificial intelligence", "URL": "https://rosneft.ru/press/news/item/222604/", "Locator": "LLM virtual assistant", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "regional", "date": "2025-07-28", "status": "excerpt reviewed", "locator": "LLM virtual assistant"}], "Audit_reviews": [{"url": "https://rosneft.ru/press/news/item/222604/", "access": "inaccessible", "title": "Rosneft introduces new technologies using artificial intelligence", "publication_date": "2025-07-28", "claims_checked": ["Company-developed LLM-based virtual assistant is used for text search, verification and formatting tasks, with applications from geology to field-development design.", "Rosneft claims 70-80% lower routine workload for these tasks; no independent validation or financial return is disclosed."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-07-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}]}, {"Source_ID": "RS124", "Source_date": "2025-11-05", "Company": "Permian Resources", "Source_type": "investor presentation", "Title": "Q3 2025 Earnings Presentation", "URL": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "Locator": "Slide 5, “Leading Execution, Cutting Edge Technology Drive Robust Results”", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-1", "date": "2025-11-05", "status": "full text reviewed", "locator": "Slide 5, “Leading Execution, Cutting Edge Technology Drive Robust Results”"}], "Audit_reviews": [{"url": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "access": "full text reviewed", "title": "Q3 2025 Earnings Presentation", "publication_date": "2025-11-05", "claims_checked": ["AI-driven subsurface insight and precision wellbore placement on printed slide 5; model and scale not disclosed; Haley results separately presented."], "notes": "Independent lead-agent source check; public company-reported description, not measured AI return."}]}, {"Source_ID": "RS125", "Source_date": "2026-05-22", "Company": "KazMunayGas", "Source_type": "company announcement", "Title": "КМГ внедряет предиктивную аналитику на НПЗ", "URL": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "Locator": "Russian-language release; refinery predictive analytics program", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-2", "date": "2026-05-22", "status": "full text reviewed", "locator": "Russian-language release; refinery predictive analytics program"}], "Audit_reviews": [{"url": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "access": "full text reviewed", "publication_date": "2026-05-22", "notes": "Official KMG Russian-language release. Confirms 2025 pilots; 2026 rollout plans, Shymkent phase, and polypropylene plant 10 units.", "title": "КМГ внедряет предиктивную аналитику на НПЗ"}]}, {"Source_ID": "RS126", "Source_date": "2025-12-25", "Company": "Atyrau Oil Refinery (ANPZ)", "Source_type": "company announcement", "Title": "AI control of equipment at Atyrau refinery: predictive analytics in action", "URL": "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/", "Locator": "AI predictive analytics pilot; connected to Honeywell PHD/MES", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-2", "date": "2025-12-25", "status": "full text reviewed", "locator": "AI predictive analytics pilot; connected to Honeywell PHD/MES"}], "Audit_reviews": [{"url": "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/", "access": "full text reviewed", "publication_date": "2025-12-25", "notes": "Official Atyrau refinery Russian-language release. October pilot start, Honeywell, equipment examples, operational anomaly cases, and 28-unit 2026 target reviewed.", "title": "AI control of equipment at Atyrau refinery: predictive analytics in action"}]}, {"Source_ID": "RS127", "Source_date": "2025-07", "Company": "QatarEnergy LNG", "Source_type": "company magazine / strategy document", "Title": "The Pioneer 167 — QatarEnergy LNG", "URL": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "Locator": "The Pioneer 167, PDF p. 5, Corporate Strategic Initiatives: AI Solutions for Machinery Fleet", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-2", "date": "2025-07", "status": "full text reviewed", "locator": "The Pioneer 167, PDF p. 5, Corporate Strategic Initiatives: AI Solutions for Machinery Fleet"}], "Audit_reviews": [{"url": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "access": "full text reviewed", "publication_date": "2025-07 (Issue 167)", "notes": "Primary company magazine, complete PDF opened; page 5 strategy initiative reviewed.", "title": "The Pioneer 167 — QatarEnergy LNG"}]}, {"Source_ID": "RS128", "Source_date": "2025-05-09", "Company": "Ambyint", "Source_type": "company announcement", "Title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction", "URL": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "Locator": "Full page", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": "2025-05-09", "status": "full text reviewed", "locator": "Full page"}], "Audit_reviews": [{"url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction", "publication_date": "2025-05-09", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Ambyint says $4.1m from ERA supports a next-generation InfinityPL platform upgrade.", "Company says the existing platform is deployed by 17 producers, and the project is expected to reduce methane venting by up to 90%.", "Tourmaline is named as key planned deployment partner at plunger-lift sites in Alberta."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "access": "full text reviewed", "publication_date": "2025-05-09", "notes": "Official Ambyint release says $4.1m, 17 existing producer deployments, up-to-90% expected venting reduction, Tourmaline planned deployment, and new capabilities.", "title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction"}]}, {"Source_ID": "RS129", "Source_date": "2025-04-16", "Company": "Emissions Reduction Alberta / Government of Alberta", "Source_type": "public funding announcement", "Title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector", "URL": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "Locator": "Ambyint project description", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": "2025-04-16", "status": "full text reviewed", "locator": "Ambyint project description"}], "Audit_reviews": [{"url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector", "publication_date": "2025-04-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["ERA confirms $4.1m funding for Ambyint’s upgrade to the InfinityPL AI platform using advanced ML and reinforcement learning.", "ERA describes prediction of inefficiencies, complex-anomaly detection and reduced manual intervention; up to-90% methane-venting reduction remains expected.", "Tourmaline plans deployment across a substantial number of Alberta plunger-lift well sites."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "access": "full text reviewed", "publication_date": "2025-04-16", "notes": "Funder primary release confirms $4.1m, advanced ML/reinforcement learning, expected up-to-90% reduction, and planned Tourmaline deployment. Does not state exact $4.08m or setpoint-management detail.", "title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector"}]}, {"Source_ID": "RS130", "Source_date": "2025-10-16", "Company": "Kpler", "Source_type": "company announcement", "Title": "Navigating Risk & Harnessing AI Briefing", "URL": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "Locator": "Harness AI in Chartering", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": "2025-10-16", "status": "full text reviewed", "locator": "Harness AI in Chartering"}], "Audit_reviews": [{"url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "title": "Navigating Risk & Harnessing AI Briefing", "publication_date": "2025-10-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Kpler described AI-powered parsing of emails, PDFs and spreadsheets into structured cargo and position lists.", "Kpler previewed an AI chartering agent to recommend vessel candidates based on freight economics, risk and custom filters, and said it was launching Q4."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "access": "full text reviewed", "publication_date": "2025-10-16 (event page)", "notes": "Official Kpler event content directly states structured email/PDF/spreadsheet parsing and chartering-agent Q4 plan.", "title": "Navigating Risk & Harnessing AI Briefing"}]}, {"Source_ID": "RS131", "Source_date": "Undated; see individual source review", "Company": "Kpler", "Source_type": "product page", "Title": "Kpler AI", "URL": "https://www.kpler.com/solutions/ai", "Locator": "Kpler Copilot / Kpler agents / FAQ", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": null, "status": "full text reviewed", "locator": "Kpler Copilot / Kpler agents / FAQ"}], "Audit_reviews": [{"url": "https://www.kpler.com/solutions/ai", "title": "Kpler AI", "publication_date": null, "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Current Kpler page describes natural-language questions, AI-assisted market analysis, and chartering/cargo/compliance agent categories.", "Page describes sources including proprietary history, AIS, satellite tracking, customs/trade flows and refinery information.", "Page contains vendor-stated capabilities and accuracy/time benefits; publication date and independent results are not provided."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.kpler.com/solutions/ai", "access": "full text reviewed", "publication_date": "undated current page", "notes": "Official current product page reviewed and searched for accuracy and 99%; general AI/data features appear, but no email-parsing accuracy metric was located.", "title": "Kpler AI"}]}, {"Source_ID": "RS132", "Source_date": "2025-11-27", "Company": "IFS", "Source_type": "company announcement", "Title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI", "URL": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "Locator": "Kodiak customer disclosure / CIO quote", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": "2025-11-27", "status": "full text reviewed", "locator": "Kodiak customer disclosure / CIO quote"}], "Audit_reviews": [{"url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI", "publication_date": "2025-11-27", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["IFS reported Kodiak rolled out a Material Replenisher agent for technicians to find and order parts conversationally.", "CIO’s $3m and 90,000-hour figures are conditional on half the workforce engaging once a day; they are modeled potential, not achieved results."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "access": "full text reviewed", "publication_date": "2025-11-27", "notes": "IFS release directly quotes the conditional half-workforce daily-use scenario; does not establish deployment by this date.", "title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI"}]}, {"Source_ID": "RS133", "Source_date": "2026-04-23", "Company": "IFS", "Source_type": "company announcement", "Title": "IFS Loops launches Agent Studio", "URL": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "Locator": "Kodiak paragraph", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-3", "date": "2026-04-23", "status": "full text reviewed", "locator": "Kodiak paragraph"}], "Audit_reviews": [{"url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "title": "IFS Loops launches Agent Studio", "publication_date": "2026-04-23", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["IFS again describes Kodiak Material Replenisher as rolled out; technicians find and order parts through conversation.", "Release repeats potential $3m annual ROI and 90,000 hours; not treated as realized outcomes."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "access": "full text reviewed", "publication_date": "2026-04-23", "notes": "IFS says Kodiak rolled out Material Replenisher and describes parts requests conversationally; calls $3m and 90k hours the result. Vendor claim, not independently validated.", "title": "IFS Loops launches Agent Studio"}]}, {"Source_ID": "RS134", "Source_date": "2026-03-03", "Company": "Expand Energy", "Source_type": "company announcement", "Title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "URL": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "Locator": "Leucipa deployment and Snowflake data-platform sections", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-root", "date": "2026-03-03", "status": "full text reviewed", "locator": "Leucipa deployment and Snowflake data-platform sections"}, {"cohort": "round2/additions-root", "date": "2026-03-03", "status": "full text reviewed", "locator": "Leucipa deployment and Snowflake data-platform sections"}], "Audit_reviews": [{"url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "access": "full text reviewed", "claims_checked": ["Leucipa AI-powered production management, field optimization and forecasting; Lucy described as a planned pilot with GenAI chat interface and real-time production insights; deployment beginning at Operations Support Center.", "Snowflake data platform adoption; ML model forecasting; real-time drilling data/models hosted; SCADA signals used to detect issues and alert Operations Support Center."], "notes": "The issuer’s announcement supports both packets but distinguishes an operating platform/deployment from the planned Lucy pilot.", "title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "publication_date": "2026-03-03"}]}, {"Source_ID": "RS135", "Source_date": "2026-06-10", "Company": "Viridien", "Source_type": "company announcement", "Title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "URL": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "Locator": "Product launch description", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-root", "date": "2026-06-10", "status": "full text reviewed", "locator": "Product launch description"}], "Audit_reviews": [{"url": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "access": "full text reviewed", "claims_checked": ["Evergreen Data Solution product launch and AI-assisted classification, curation, contextualization and validation of subsurface data."], "notes": "Commercial product capability, not customer deployment proof.", "title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "publication_date": "2026-06-10"}]}, {"Source_ID": "RS136", "Source_date": "2026-08-18", "Company": "TGS", "Source_type": "company announcement", "Title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "URL": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "Locator": "Opening four paragraphs", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-root", "date": "2026-08-18", "status": "full text reviewed", "locator": "Opening four paragraphs"}], "Audit_reviews": [{"url": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "access": "full text reviewed", "claims_checked": ["TGS multi-year AI seismic foundation-model contract award to an unnamed international operator; scope includes structural and stratigraphic interpretation and TGS multi-client data."], "notes": "Customer remains unnamed; no contract value or field result disclosed.", "title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "publication_date": "2026-08-18"}]}, {"Source_ID": "RS137", "Source_date": "2025-04-25", "Company": "SLB", "Source_type": "earnings release", "Title": "SLB Announces First-Quarter 2025 Results", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "Locator": "Core — Woodside Trion contract bullet", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-root", "date": "2025-04-25", "status": "excerpt reviewed", "locator": "Core — Woodside Trion contract bullet"}], "Audit_reviews": [{"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "access": "excerpt reviewed", "claims_checked": ["SLB Q1 2025 earnings release lines 285–286: Woodside award for integrated delivery of 18 Trion ultra-deepwater wells; AI-enabled drilling capabilities included; service start scheduled early 2026 and delivery via Performance Live."], "notes": "The cited release supports the packet’s timing as a plan and the contracted scope; it does not prove services commenced or outcomes were achieved.", "title": "SLB Announces First-Quarter 2025 Results", "publication_date": "2025-04-25"}]}, {"Source_ID": "RS138", "Source_date": "Undated; see individual source review", "Company": "Technip Energies", "Source_type": "annual report", "Title": "Technip Energies shareholder information — 2025 chair and CEO messages", "URL": "https://investors.technipenergies.com/shareholder-information", "Locator": "Digital technology; Driving Innovation and Technological Leadership; Looking Ahead", "Short_excerpt": "", "Verification": "excerpt reviewed", "Notes": "Research expansion. Company and vendor claims are not independently validated.", "Accessed": "2026-09-27", "Research_reviews": [{"cohort": "round2/additions-root", "date": null, "status": "excerpt reviewed", "locator": "Digital technology; Driving Innovation and Technological Leadership; Looking Ahead"}], "Audit_reviews": [{"url": "https://investors.technipenergies.com/shareholder-information", "access": "excerpt reviewed", "claims_checked": ["2025 shareholder/management messages say DAP leverages advanced analytics and large language models; DAP launched in 2025; target €100m cost savings and incremental revenues by 2028; AI for All implemented."], "notes": "The €100m is a 2028 target, not realized savings. The source supplies no named AI workflow or AI-specific attribution for savings.", "title": "Technip Energies shareholder information — 2025 chair and CEO messages", "publication_date": "Undated; see individual source review"}]}, {"Source_ID": "FU001", "Title": "Devon Q1 2026 earnings call — Smart Gas Lift", "URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Source_date": "2026-05-06", "Company": "Devon Energy", "Source_type": "Issuer-hosted earnings transcript", "Locator": "Prepared remarks p. 4; Q&A pp. 12–14; 21 October additional locator: PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI", "Verification": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation. Earlier claim reviews retain their original scope.", "Notes": "Research follow-up, 28 September 2026", "Accessed": "2026-10-21", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; pp. 3-4 (PDF pp. 3-4), CEO Clay Gaspar prepared remarks, lines 108-128; fully read 20-page transcript", "note": "Original source opened; relevant passage read."}, "Publisher": "Devon Energy / FactSet CallStreet", "Current_review": {"last_checked": "2026-10-21", "source_read": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.", "independent_validation": "Not established", "note": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status."}}, {"Source_ID": "FU002", "Title": "SLB introduces autonomous AI geosteering", "URL": "https://www.slb.com/newsroom/press-release/2024/slb-adds-ai-geosteering", "Source_date": "2024-12-09", "Company": "SLB", "Source_type": "Provider press release", "Locator": "Full article / relevant case section", "Verification": "Source text reviewed; reported performance not independently validated", "Notes": "Research follow-up, 28 September 2026", "Accessed": "2026-09-28"}, {"Source_ID": "FU003", "Title": "Field deployments of AI-based IOCaaS", "URL": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "Source_date": "2026-04-01", "Company": "SPE / JPT; OPX Ai authors", "Source_type": "Vendor-authored technical case article", "Locator": "Kaybob and Montney deployment sections; integration and measurement-period paragraphs", "Verification": "Full vendor-authored JPT article rechecked 6 October 2026", "Notes": "Research follow-up, 28 September 2026", "Accessed": "2026-10-18", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Montney case; architecture, implementation, first-winter results and authorship.", "note": "Original full article opened/read, including the Montney deployment, architecture, timeline, outcomes, and authorship. The source gives operator/team attributions but does not provide underlying data or independent validation."}, "Current_review": {"last_checked": "2026-10-18", "source_read": "Original JPT case sections read", "independent_validation": "Not established", "note": "OPX Ai founders coauthor. Both Kaybob AI-201 and Montney AI-202 already covered; library indexing adds no adoption or outcome count."}}, {"Source_ID": "FR001", "Title": "Upstream AI opportunity scenarios", "URL": "https://www.mckinsey.com/industries/oil-and-gas/our-insights/can-upstream-oil-and-gas-produce-value-from-ais-230-billion-dollar-pay-zone", "Source_date": "25 August 2026", "Company": "McKinsey", "Source_type": "Modeled potential", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "More than 550 use cases; addressability and deployment assumptions, overlap correction and over $30bn in annual implementation costs. Full potential assumes wider autonomy.", "Accessed": "2026-09-28"}, {"Source_ID": "FR002", "Title": "Gen AI adoption in oilfield services", "URL": "https://www.mckinsey.com/industries/oil-and-gas/our-insights/gen-ai-in-the-ofse-industry-progress-lags-behind-intent", "Source_date": "1 April 2026; survey May 2025", "Company": "McKinsey", "Source_type": "Executive survey", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "164 global OFSE executives. Self-reported responses, fielded almost a year before publication; not a representative census of all operator AI use.", "Accessed": "2026-09-28"}, {"Source_ID": "FR003", "Title": "How operating models must adapt", "URL": "https://www.mckinsey.com/industries/oil-and-gas/our-insights/four-shifts-redefining-the-oil-and-gas-operating-model-of-the-future", "Source_date": "5 January 2026", "Company": "McKinsey", "Source_type": "Organizational framework", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Management framework; not a deployment survey or a measured AI-return series.", "Accessed": "2026-09-28"}, {"Source_ID": "FR004", "Title": "Technology transformation before the current AI cycle", "URL": "https://www.mckinsey.com/capabilities/operations/our-insights/harnessing-volatility-technology-transformation-in-oil-and-gas", "Source_date": "1 September 2022 · historical context", "Company": "McKinsey", "Source_type": "Earlier transformation evidence", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Older practice synthesis with anonymized cases; broad technology programs rather than isolated AI treatment effects.", "Accessed": "2026-09-28"}, {"Source_ID": "FR005", "Title": "Energy and AI: scaling beyond individual cases", "URL": "https://www.iea.org/reports/energy-and-ai/ai-for-energy-optimisation-and-innovation", "Source_date": "10 April 2025", "Company": "IEA", "Source_type": "Independent system analysis", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Energy-system analysis and case-based scenarios. Region, digital infrastructure and adoption barriers affect attainable scale.", "Accessed": "2026-09-28"}, {"Source_ID": "FR008", "Title": "Libyan gas-lift optimization field paper", "URL": "https://ejp.researchcommons.org/journal/vol34/iss3/8/", "Source_date": "2025", "Company": "Egyptian Journal of Petroleum", "Source_type": "Peer-reviewed paper; abstract reviewed", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Only abstract and bibliographic record were reviewed. Cohort, controls, intervention history and AI-specific attribution require the full methods.", "Accessed": "2026-09-28"}, {"Source_ID": "FR009", "Title": "Machine-learning gas-lift forecast and injection optimization", "URL": "https://jpt.spe.org/urtec-exxonmobils-machine-learning-workflow-boosts-bakken-output", "Source_date": "Undated or historical; see research notes", "Company": "Supplemental evidence", "Source_type": "Contextual case study", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Historical disclosure: 24 June 2024, before the research window. SPE/JPT reports a 30-well ML pilot and subsequent 200+ well deployment; closed-loop valve actuation is not established.", "Accessed": "2026-09-28"}, {"Source_ID": "FR010", "Title": "Leucipa ESP failure prediction and recovery recommendations", "URL": "https://dam.bakerhughes.com/m/6d9e02a632fdc7e4/original/Case-Study_Leucipa-delivers-operational-efficiency-for-ExxonMobil.pdf", "Source_date": "Undated or historical; see research notes", "Company": "Supplemental evidence", "Source_type": "Contextual case study", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Undated Baker Hughes PDF, copyright 2025. Vendor reports a 157-well pilot, 80% fewer failures since Q3 2024 and $1.8m+ savings. Publication date, baseline and ROI methodology are unresolved. This is ESP reliability, not autonomous setpoint control.", "Accessed": "2026-09-28"}, {"Source_ID": "FR011", "Title": "ML and capacitance-resistance models for waterflood pattern surveillance", "URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/enhancing-waterflooding-efficiency-through-integrated-workflows-and-advanced-analytics", "Source_date": "Undated or historical; see research notes", "Company": "Supplemental evidence", "Source_type": "Contextual case study", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Undated vendor case, unnamed Ecuador operator. ML and physics-based surveillance; reported analysis time fell from 23 to 5 hours. The page’s separate 85% efficiency claim does not equal the 78% time reduction implied by those endpoints. Results through 2024 do not establish publication date.", "Accessed": "2026-09-28"}, {"Source_ID": "FR012", "Title": "AI/ML geosteering recommendations integrated with automated directional drilling", "URL": "https://www.halliburton.com/en/resources/first-horizontal-well-via-fully-automated-ai-driven-drilling-technology", "Source_date": "Undated or historical; see research notes", "Company": "Supplemental evidence", "Source_type": "Contextual case study", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Undated provider case for an unnamed operator. A one-well trial reports 87.4% of section footage without human intervention. Geological-team decisions remain part of the described workflow; the page does not establish end-to-end autonomous interpretation.", "Accessed": "2026-09-28"}, {"Source_ID": "FR013", "Title": "Bakken rod-pump optimization case", "URL": "https://www.ambyint.com/case-studies/oil-well-optimization", "Source_date": "Undated or historical; see research notes", "Company": "Supplemental evidence", "Source_type": "Contextual case study", "Locator": "Relevant article / case section", "Verification": "Research follow-up source reviewed; limitations stated in research notes", "Notes": "Provider case dated 22 March 2024, outside the disclosure window. Useful historical lead; excluded from the current initiative count unless a qualifying update is found.", "Accessed": "2026-09-28"}, {"Source_ID": "EQTD01", "Source_date": "2026-02-18", "Company": "EQT", "Source_type": "Issuer-hosted corrected earnings-call transcript", "Title": "EQT Q4 2025 earnings call — AI-assisted procurement targeting", "URL": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "Locator": "Printed pp. 20–21; Toby Rice, service rebids and procurement", "Short_excerpt": "", "Verification": "full text reviewed", "Notes": "Review on 29 September 2026 establishes the management statement. Prospective opportunity; no realized or AI-isolated savings.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-001", "Company": "Marathon Pipe Line (MPLX / Marathon Petroleum)", "Title": "Our People: How Paige Chenevey drives success at Marathon Pipe Line", "URL": "https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/", "Source_date": "2025-04-09", "Source_type": "Operator interview / company news", "Locator": "Paragraphs under ‘What drives MPL’s success in safety, innovation and sustainability?’, especially lines 43–44 in web text; page dates article Apr 9 2025", "Verification": "Full company page opened; article text directly accessible.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-002", "Company": "ORLEN Group", "Title": "ORLEN’s transformation to gain momentum with CEE’s largest AI deployment", "URL": "https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment", "Source_date": "2025-07-08", "Source_type": "Operator press release", "Locator": "Web lines 312–323: AI deployment objectives; Copilot and Agentic AI; use cases monitoring global oil and gas market data and streamlining supply-chain logistics; internal competence centre to identify, develop and scale projects.", "Verification": "Full English company announcement opened; exact application wording accessible.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-003", "Company": "Apache Corporation / APA Corporation", "Title": "Egypt Leading Oil & Gas Excellence Supplement — November 2024 (Apache interview)", "URL": "https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence-1.pdf", "Source_date": "2024-11", "Source_type": "Operator executive interview in industry supplement / original report", "Locator": "Printed p. 14 (PDF page 14 / extracted web PDF P13), interview with Greg McDaniel, VP Apache Egypt Assets and Country Manager; technology question/answer lines 826–837. Supplement imprint: November 2024 at printed p. 3 / PDF P2; publisher publication listing corroborates.", "Verification": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-004", "Company": "Murphy Oil Corporation", "Title": "2025 Murphy Sustainability Report", "URL": "https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf", "Source_date": "2025-08-06", "Source_type": "Operator sustainability report", "Locator": "Printed p. 78 (PDF page index P77), “Risk Management Strategy” Technology column: continuous monitoring and analytics built on advanced machine learning and AI; tools include endpoint/email protection and Managed Detection and Response. Printed p. 79 (PDF index P78) describes the Murphy Generative AI Hub and enterprise Microsoft Copilot rollout. Report printed p. 3 (PDF P2) gives publication date August 6, 2025; issuer Q2 announcement/update also dates Aug. 6 and confirms the report was published.", "Verification": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-005", "Company": "TotalEnergies", "Title": "TotalEnergies and Honeywell Pilot AI-Assisted Control Room to Accelerated Shift to Industrial Autonomy", "URL": "https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial", "Source_date": "2025-11-11", "Source_type": "Operator-hosted joint pilot announcement", "Locator": "Lines 11–19; five event predictions and 12-minute lead at line 17", "Verification": "Full primary announcement reviewed.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-006", "Company": "Snam / Engineering Group", "Title": "Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants", "URL": "https://arxiv.org/html/2511.08609v1", "Source_date": "2025-11-05", "Source_type": "Operator/provider-authored technical preprint", "Locator": "Sections 3.1, 4 (Tables 1–3) and 5; abstract uses different summary accuracy figures", "Verification": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-007", "Company": "Diversified Energy Company (DEC)", "Title": "Diversified Energy 2025 Sustainability Report", "URL": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "Source_date": "2026-05-28", "Source_type": "Operator-published sustainability report; issuer release confirms report date", "Locator": "Printed p. 64 (PDF physical page 63), Artificial Intelligence / AI Efficiency at Scale; issuer release dated 28 May 2026", "Verification": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-008", "Company": "Trafigura", "Title": "Trafigura announces strategic alliance with maritime technology provider ZeroNorth", "URL": "https://www.trafigura.com/news-and-insights/press-releases/2025/trafigura-announces-strategic-alliance-with-maritime-technology-provider-zeronorth/", "Source_date": "2025-07-09", "Source_type": "Operator press release with named partner and attributed Global Head of Shipping", "Locator": "Trafigura press release, 9 July 2025, paragraphs 2-5", "Verification": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-009", "Company": "SM Energy", "Title": "SM Energy Company — Fireside Chat Stephens Annual Investment Conference (internally generated transcript; Rule 425 filing)", "URL": "https://www.sec.gov/Archives/edgar/data/893538/000089353825000160/stephensannualinvestorco.htm", "Source_date": "2025-11-21", "Source_type": "Company conference transcript filed with SEC under Rule 425", "Locator": "Transcript p. 7, COO Beth McDonald; conference 20 November, filing 21 November 2025", "Verification": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-010", "Company": "Winn Resources", "Title": "Winn Resources Partners with Collide to Turn Operational Data Into a Competitive Advantage", "URL": "https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html", "Source_date": "2026-03-05", "Source_type": "Vendor-issued press release with named customer executive quotation", "Locator": "PRNewswire release sourced to Collide, 5 March 2026; 'Challenge', 'Solution', and 'Looking Ahead' sections. Specific case figures and approval process in paragraphs equivalent to 50-well cycle, 2-4 hour prior process, missing test flags, human review, and planned revenue extraction.", "Verification": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "M29-011", "Company": "PETRONAS Carigali", "Title": "PETRONAS Advances TriCipta AI through the third Joint Development Agreement with IBM and Tridiagonal.AI", "URL": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "Source_date": "2026-07-08", "Source_type": "Operator media release", "Locator": "Lines 129–143; new domain and scope lines 135–140; prior deployed AI.SEEK and GEB lines 142–143", "Verification": "Full issuer media release reviewed.", "Notes": "29 September 2026 gap review; reported performance is not independently validated.", "Accessed": "2026-09-29"}, {"Source_ID": "SW2-001", "Company": "Koch Energy Services", "Title": "ipushpull launches LLM-enabled agent to automate physical gas trade booking via chat", "URL": "https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat", "Source_date": "2025-09-18", "Source_type": "Vendor announcement with named customer quote", "Locator": "Opening dateline; Darnell Bortz, Director of Natural Gas Trading quote", "Verification": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29"}, {"Source_ID": "SW2-002", "Company": "Koch Energy Services", "Title": "Koch Energy Services and ipushpull automate physical gas trade booking through chat", "URL": "https://ipushpull.com/case-studies/koch-energy-services-and-ipushpull-automate-physical-gas-trade-booking-through-chat", "Source_date": "Undated", "Source_type": "Vendor customer case study", "Locator": "Reimagining trade capture; Human-in-the-loop", "Verification": "Indexed full source reviewed; used for workflow controls, not dated outcome update.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29"}, {"Source_ID": "SW2-003", "Company": "Occidental", "Title": "Occidental Q2 2025 Earnings Call Transcript", "URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy2q25transcript.pdf", "Source_date": "2025-08-07", "Source_type": "Issuer-hosted earnings-call transcript", "Locator": "Page 3, Vicki Hollub prepared remarks", "Verification": "Full 17-page company-hosted transcript reviewed.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29"}, {"Source_ID": "SW2-004", "Company": "Occidental", "Title": "2026 Oxy Proxy Statement", "URL": "https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf", "Source_date": "2026-03-19", "Source_type": "Issuer proxy statement", "Locator": "Compensation Discussion and Analysis, footnote (4), printed page 43 (PDF index 42)", "Verification": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29"}, {"Source_ID": "SW2-005", "Company": "ADNOC Gas", "Title": "ADNOC Gas Integrated Report 2025", "URL": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "Source_date": "2026", "Source_type": "Issuer integrated report", "Locator": "Printed pages 10–11, AI Highlights; PDF physical page 7", "Verification": "Document accessible on 29 September 2026. New review: Printed pages 10–11, AI Highlights; PDF physical page 7. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29", "Followup_source_review": {"date": "2026-09-29", "status": "Latest access review: original document available; Printed pages 10–11, AI Highlights; PDF physical page 7", "note": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected."}}, {"Source_ID": "SW2-006", "Company": "Petrobras", "Title": "Petrobras Business Plan 2026–2030 investor presentation", "URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Source_date": "2025-11", "Source_type": "Issuer investor presentation", "Locator": "Slide 158, Accelerate value generation through digital technologies", "Verification": "Document accessible on 29 September 2026. New review: Slide 158, Accelerate value generation through digital technologies. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-29", "Followup_source_review": {"date": "2026-09-29", "status": "Latest access review: original document available; Slide 158, Accelerate value generation through digital technologies", "note": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping."}}, {"Source_ID": "SW3-001", "Company": "Enbridge", "Title": "Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization", "URL": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "Source_date": "2024-10-07", "Source_type": "Operator press release", "Locator": "AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "Verification": "Document accessible on 30 September 2026. New review: AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-30", "Followup_source_review": {"date": "2026-09-30", "status": "Latest access review: original document available; AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "note": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics."}}, {"Source_ID": "SW3-002", "Company": "Williams", "Title": "Williams tops 90% AI adoption as approved tools become part of everyday work", "URL": "https://www.williams.com/2026/06/01/williams-tops-90-ai-adoption-as-approved-tools-become-part-of-everyday-work/", "Source_date": "2026-06-01", "Source_type": "Operator article", "Locator": "Key takeaways and companywide adoption paragraphs", "Verification": "Original operator article read in full; date from article URL; adoption definition limitation retained.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-30"}, {"Source_ID": "SW3-003", "Company": "Oil and Natural Gas Corporation Limited (ONGC)", "Title": "ONGC Integrated Annual Report 2024-25", "URL": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "Source_date": "2025-08-07", "Source_type": "integrated annual report / statutory filing", "Locator": "Annexures to the Board's Report, printed p. 131, item 30; PDF p. 157", "Verification": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-30"}, {"Source_ID": "SW3-004", "Company": "ADNOC", "Title": "ADNOC and SLB Deploy AI Platform Across Over 120 Drilling Rigs to Strengthen Upstream Performance", "URL": "https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/", "Source_date": "2026-08-04", "Source_type": "Operator press release", "Locator": "Release body: paragraphs on RTOC deployment, workflow, operating scope and reported metrics", "Verification": "Full primary release opened and read; release body reviewed in full.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-30"}, {"Source_ID": "SW3-005", "Company": "Devon Energy", "Title": "Devon Energy Selects Harvey's Legal AI Platform", "URL": "https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform", "Source_date": "2025-08-14", "Source_type": "Vendor announcement with named customer executive quote", "Locator": "Byline and announcement text; Devon Senior Counsel Matt Tilly quote", "Verification": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-09-30"}, {"Source_ID": "SW3-006", "Company": "Vital Energy", "Title": "Unsung Hero: Artificial Lift Boosts Production Without Breaking the Bank", "URL": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted", "Source_date": "2024-10-01", "Source_type": "JPT / SPE technical reporting of operator case", "Locator": "Vital Energy ESP recommendations and nine-well full self-pumping trial", "Verification": "Document accessible on 1 October 2026. New review: Article date at lines 592–599; ExxonMobil SPE 219553 workflow and claims at lines 625–646. Earlier claim-level findings remain scoped to their original review.", "Notes": "Operator account reported by JPT/SPE, not an independently validated trial. Forecast comparison is not a controlled counterfactual. Reliability improvements were expressly not fully attributed to AI. No implementation cost or payback disclosed; trial dates and control safeguards remain unspecified.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Article date at lines 592–599; ExxonMobil SPE 219553 workflow and claims at lines 625–646", "note": "Full relevant article section opened and read, including the two 3-month control stages, operator quotations, workflow description, and caveats. Reporting summarizes a paper and interview; it is not independent validation."}}, {"Source_ID": "SW4-001", "Company": "Moeve (formerly Cepsa)", "Title": "Moeve Consolidated Management Report 2025", "URL": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "Source_date": "2026-02-26", "Source_type": "Operator annual management report", "Locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "Verification": "Document accessible on 1 October 2026. New review: p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "note": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication."}}, {"Source_ID": "SW4-002", "Company": "Moeve (formerly Cepsa)", "Title": "Moeve and Mistral AI join forces to spearhead generative AI solutions in the energy sector", "URL": "https://www.moeveglobal.com/en/press/moeve-mistral-ai-alliance-ai-digital-assistant-industrial-efficiency", "Source_date": "2025-12-04", "Source_type": "Operator press release", "Locator": "Pilot digital assistant paragraphs; press index spells out 4 December 2025", "Verification": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW4-003", "Company": "Presidio Production Company (NYSE: PSD)", "Title": "Presidio Production Company Investor Presentation, dated June 2026", "URL": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026069875/ea029480701_ex99-1.htm", "Source_date": "2026-06", "Source_type": "issuer investor presentation filed with SEC", "Locator": "Slides 15–18 (SEC document lines 122–146); AI deployment, input data and live-versus-target distinctions; slides 13 and 26 for selected acquisition case studies and pump-by-exception routing", "Verification": "Full SEC-hosted HTML presentation opened and reviewed, including legal disclaimer and AI slides.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW4-004", "Company": "SLB", "Title": "SLB enhances productivity with Power Platform and AI", "URL": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "Source_date": "2025-08-08", "Source_type": "Microsoft Learn customer case study with named SLB architect", "Locator": "Flight Airline Automation Bot and tax process: page sections 148–174; page last-updated stamp at line 287; date is page last update, not go-live", "Verification": "Document accessible on 1 October 2026. New review: AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live", "note": "Full official Microsoft Learn page opened and reviewed."}}, {"Source_ID": "SW4-005", "Company": "ExxonMobil / XTO Energy", "Title": "Gas Lift Optimization in the Permian Uses Machine Learning, Artificial Intelligence", "URL": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "Source_date": "2025-10-01", "Source_type": "SPE Journal of Petroleum Technology paper highlight", "Locator": "Accessible abstract and first paragraph; lines 592–620", "Verification": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW4-006", "Company": "ExxonMobil / XTO Energy", "Title": "Gas Lift Optimization in the Permian Using Machine Learning and Artificial Intelligence (SPE 219553-MS)", "URL": "https://www.researchgate.net/publication/383230634_Gas_Lift_Optimization_in_the_Permian_Using_Machine_Learning_and_Artificial_Intelligence", "Source_date": "2024-08", "Source_type": "Operator-authored SPE conference paper abstract", "Locator": "Complete abstract, paragraphs 1–4", "Verification": "Full accessible abstract read. Page explicitly says the full text is unavailable and can be requested from the authors; do not characterize the complete paper as reviewed.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW4-007", "Company": "Dagang Oilfield (PetroChina/CNPC)", "Title": "Machine-learning production forecasting across Dagang Oilfield blocks", "URL": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml", "Source_date": "2026-02-04", "Source_type": "CNPC News Center / China Petroleum News Center operator-group news article", "Locator": "Article section “让科技生产力‘看得见’” (“Make technological productivity visible”), paragraph beginning “人工智能的应用体现出务实的导向” (“AI applications reflect a practical orientation”); published 2026-02-04 10:17.", "Verification": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-001", "Company": "Repsol", "Title": "Harvey Power Users: “People Call Me Harvey”", "URL": "https://www.harvey.ai/blog/harvey-power-users-people-call-me-harvey", "Source_date": "2026-03-30", "Source_type": "Harvey customer interview with Repsol Legal Counsel", "Locator": "Byline/date and rollout scope lines 62–77; legal-team size, rollout lead, training and power users.", "Verification": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-002", "Company": "Repsol", "Title": "Repsol becomes the first Spanish in-house legal team to use Harvey", "URL": "https://www.repsol.com/en/press-room/press-releases/2024/repsol-becomes-the-first-spanish-in-house-legal-team-to-use-harvey/index.cshtml", "Source_date": "2024-04-04 (pre-window context only)", "Source_type": "Repsol issuer press release", "Locator": "Press release date and scope lines 675–689: legal document analysis, contract/regulation comparison, clause suggestions, translation, legal queries, summaries; human validation; estimated up to three hours/week saved per lawyer.", "Verification": "Full issuer page opened and read. Pre-window agreement and estimated time savings used only as history; not counted as an in-window deployment or measured outcome.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-003", "Company": "Spearpoint Resources", "Title": "Three entities, 27 purchaser formats, and 20 to 30 hours a week back", "URL": "https://www.joltly.io/blog/spear-point-revenue-accounting-case-study", "Source_date": "2026-09-05", "Source_type": "Joltly customer case study quoting Spearpoint controller", "Locator": "Date lines 9–14; workflow/architecture lines 24–47 and 130–154; May–August 2026 volumes lines 49–78; outcome quote lines 156–170.", "Verification": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Publisher": "Joltly"}, {"Source_ID": "SW5-004", "Company": "INPEX Corporation", "Title": "INPEX Integrated Report 2024", "URL": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf", "Source_date": "2025-07-04", "Source_type": "Operator integrated report", "Locator": "Printed p. 41 (PDF physical page 42), “Thorough Utilization of Digital Technology,” Case example 2: “Accelerating CCS simulation using machine learning”; official integrated-report index dates publication 4 July 2025.", "Verification": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-005", "Company": "INPEX Corporation", "Title": "INPEX Integrated Report 2025 — Growth Strategy section", "URL": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "Source_date": "2026-07-07", "Source_type": "Operator integrated report (separate Growth Strategy PDF)", "Locator": "Printed pp. 39–42 (PDF pp. 40–43), Technical Strategies and Digital Strategy sections; CCS surrogate was not discussed in the section reviewed.", "Verification": "Document accessible on 1 October 2026. New review: Printed p. 40 (PDF physical page 17 of the 21-page Growth Strategy extract), “Technical Strategies,” Ichthys LNG O&M AI case: development and deployment, report retrieval and maintenance-record structuring; official report index confirms report publication 7 July 2026.. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Printed p. 40 (PDF physical page 17 of the 21-page Growth Strategy extract), “Technical Strategies,” Ichthys LNG O&M AI case: development and deployment, report retrieval and maintenance-record structuring; official report index confirms report publication 7 July 2026.", "note": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date."}}, {"Source_ID": "SW5-006", "Company": "INPEX Corporation", "Title": "INPEX Integrated Report index — publication date confirmation", "URL": "https://www.inpex.com/english/ir/library/integrated_report/", "Source_date": "Undated live index; accessed 2026-10-01", "Source_type": "Issuer report index", "Locator": "Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.", "Verification": "Document accessible on 1 October 2026. New review: Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.", "note": "Full issuer index page opened and publication dates read."}}, {"Source_ID": "SW5-007", "Company": "Saipem", "Title": "Saipem publishes 2025 sustainability results", "URL": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "Source_date": "2026-05-20", "Source_type": "Issuer disclosure", "Locator": "Dated announcement and safety technology discussion", "Verification": "Document accessible on 1 October 2026. New review: Dated announcement and safety technology discussion. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Dated announcement and safety technology discussion", "note": "Original document opened and relevant full sections read; issuer claims are not independently validated."}}, {"Source_ID": "SW5-008", "Company": "Saipem", "Title": "Saipem Smart Cameras privacy notice", "URL": "https://www.saipem.com/sites/default/files/2026-01/75_Privacy_Notice_Smart_Cameras.pdf", "Source_date": "2026-01-28", "Source_type": "Issuer operational notice", "Locator": "pp. 1–2: operating model, human verification, retraining and retention; date printed on all pages", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-009", "Company": "Saipem", "Title": "Saipem MISR Smart Cameras notice — Saipem 10000", "URL": "https://www.saipem.com/sites/default/files/2025-04/Information%20notice%20Smart%20Cameras_Saipem%20MISR%20for%20Petroleum%20Services%20%28S.A.E.%29_0.pdf", "Source_date": "2025-03-31", "Source_type": "Issuer operational notice", "Locator": "pp. 1–3: named vessel, validation and responsible safety personnel", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-010", "Company": "Saipem", "Title": "Saipem Sustainability Overview 2025", "URL": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf", "Source_date": "2026", "Source_type": "Issuer sustainability report", "Locator": "Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.", "Verification": "Document accessible on 1 October 2026. New review: Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.", "note": "Original document opened and relevant full sections read; issuer claims are not independently validated."}}, {"Source_ID": "SW5-011", "Company": "Saipem", "Title": "Saipem Consolidated Sustainability Statement 2025", "URL": "https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf", "Source_date": "2026", "Source_type": "Issuer annual-report extract", "Locator": "Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests", "Verification": "Document accessible on 1 October 2026. New review: Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests", "note": "Original document opened and relevant full sections read; issuer claims are not independently validated."}}, {"Source_ID": "SW5-012", "Company": "MODEC", "Title": "MODEC, Shell Brasil, Shape Digital and Unicamp safety-barrier AI project", "URL": "https://www.modec.com/news/2025/20250813_pr_shape-digital.html", "Source_date": "2025-08-13", "Source_type": "Issuer project announcement", "Locator": "Announcement paragraphs on 36-month project, Phase II and existing Shape Reef", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-013", "Company": "SBM Offshore", "Title": "SBM Offshore Q3 2025 trading update", "URL": "https://www.sbmoffshore.com/newsroom/sbm-offshore-third-quarter-2025-trading-update/", "Source_date": "2025-11-13", "Source_type": "Issuer disclosure", "Locator": "Digitalization and AI section; separates platform rollout from broader alliance", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-014", "Company": "SBM Offshore", "Title": "SBM Offshore and Cognite accelerate AI and data collaboration", "URL": "https://www.sbmoffshore.com/newsroom/sbm-offshore-and-cognite-partner-to-accelerate-ai-and-data-collaboration-to-optimize-floating-production-operations/", "Source_date": "2025-10-02", "Source_type": "Issuer disclosure", "Locator": "Dated release: Brazilian fleet scope, 18-month foundation work and prospective maintenance uses", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-015", "Company": "SBM Offshore", "Title": "SLB and SBM Offshore digital alliance", "URL": "https://www.sbmoffshore.com/newsroom/slb-and-sbm-offshore-agree-digital-alliance-to-transform-fpso-production-performance/", "Source_date": "2025-10-02", "Source_type": "Issuer disclosure", "Locator": "OptiSite, Cognite Data Fusion and Lumi; planned full-field integration", "Verification": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW5-016", "Company": "Sentinel Peak Resources / Quantum Energy Partners", "Title": "Environment — Sentinel Peak Resources", "URL": "https://sentinelpeakresources.com/sustainability/environment/", "Source_date": "undated; accessed 2026-10-01", "Source_type": "Operator environment disclosure", "Locator": "Step 1: Reduced Steam, lines 15–26", "Verification": "Full original HTML page opened and read. It names Sentinel Peak's Intelligent Oilfield as an ML/AI-powered data system for real-time steam-management optimization and estimates 200,000 MT CO2e/year enabled through intelligent oilfield technologies. The page gives no publication date, calculation, baseline, or AI-only attribution split.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW6-001", "Company": "PTTEP", "Title": "PTTEP 56-1 One Report 2025 — internal audit", "URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Source_date": "2026; referenced in invitation dated 27 February 2026", "Source_type": "Operator annual report", "Locator": "Printed p. 223; physical PDF page 212 of downloaded 557-page file, Oversight of Internal Audit Activities.", "Verification": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW6-002", "Company": "PTTEP", "Title": "PTTEP 2026 AGM invitation", "URL": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-invitation-en.pdf", "Source_date": "2026-02-27", "Source_type": "Issuer meeting invitation", "Locator": "Physical PDF page 2: dated invitation lists the 2025 One Report as attachment.", "Verification": "Original PDF opened; date and report attachment read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW6-003", "Company": "MOL Group", "Title": "MOL 2025 AGM documents — Consumer Services", "URL": "https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf", "Source_date": "2025 AGM material; meeting 24 April 2025", "Source_type": "Issuer AGM management report", "Locator": "Printed p. 27; physical PDF page 29, MOL MOVE paragraph.", "Verification": "Original PDF opened; full relevant paragraph and surrounding pages read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW6-004", "Company": "Ecopetrol", "Title": "Ecopetrol 2025 Integrated Management Report — EVO talent map", "URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Source_date": "2026 AGM report; covers 2025", "Source_type": "Operator annual report", "Locator": "Printed p. 229; physical PDF page 230, TEC para el talento humano / EVO.", "Verification": "Document accessible on 1 October 2026. New review: Printed p. 89; physical PDF page 90, Centro Autónomo de Despacho Energético.. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; Printed p. 89; physical PDF page 90, Centro Autónomo de Despacho Energético.", "note": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked."}}, {"Source_ID": "SW7-001", "Company": "Continental Resources", "Title": "Continental Resources Streamlines Operations with Automation: From fragmented emails to orchestrated operations at Continental Resources", "URL": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "Source_date": "No displayed publication/update date; accessed 2026-10-01", "Source_type": "Vendor customer case", "Locator": "Full page opened/read. Workflow and transition from human validation to automated handling at lines 241-255; quoted Sr. IT Manager Justin Cornell at 257-259; results at 260-265; 200+ emails/40+ processes/~5 minutes at 228-237.", "Verification": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-002", "Company": "Continental Resources", "Title": "Honoring the 2025 UiPath AI25 Winners", "URL": "https://www.uipath.com/ai/ai25-awards/winners", "Source_date": "Undated page; references 2025 award cohort; accessed 2026-10-01", "Source_type": "Vendor customer case", "Locator": "Full page opened/read; Continental listing at lines 294-302, reports WellView and Databricks integration and high-90% accuracy vs ~70%, near-equal manual-intervention reduction.", "Verification": "Full page read. No publication/update or metric-measurement date displayed.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-003", "Company": "Continental Resources", "Title": "UiPath Announces 2025 AI25 Awards Winners", "URL": "https://www.uipath.com/newsroom/uipath-announces-2025-ai25-awards-winners", "Source_date": "2025-09-29", "Source_type": "Vendor customer case", "Locator": "Full press-release content reviewed from issuer's newsroom/search result. Names Continental among 2025 AI25 award winners. Does not disclose workflow details or performance metrics.", "Verification": "Dated primary UiPath press release reviewed. Used solely to date the award announcement, not the operational metrics.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-004", "Company": "Continental Resources", "Title": "UiPath + Databricks: Data Intelligence Meets Agentic Automation", "URL": "https://www.uipath.com/platform/agentic-automation/ai-ecosystem/databricks", "Source_date": "Undated page; accessed 2026-10-01", "Source_type": "Vendor partner page", "Locator": "Continental customer panel: Databricks Genie Agents and UiPath Agents for email processing.", "Verification": "Original page reviewed; no date or incremental result.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-005", "Company": "Chord Energy", "Title": "Chord Energy FY2025 results release", "URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "Source_date": "2026-02-25", "Source_type": "issuer earnings release", "Locator": "Operations Update, lines 210–216; specifically line 216 (AI deployment, run time metric, workover rig days).", "Verification": "Full source opened and read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-006", "Company": "Chord Energy", "Title": "Chord Energy 1Q26 Earnings Presentation", "URL": "https://www.chordenergy.com/wp-content/uploads/2026/05/CHRD-Q1-2026-Earnings-Presentation.pdf", "Source_date": "2026-05-05", "Source_type": "issuer investor presentation", "Locator": "Slide 16; full extracted slide lines 457–505.", "Verification": "Full 28-page PDF opened; slide 16 read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-007", "Company": "Chord Energy", "Title": "Chord Energy 2Q26 results release", "URL": "https://ir.chordenergy.com/2026-08-05-Chord-Energy-Reports-Second-Quarter-2026-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Updates-2026-Outlook", "Source_date": "2026-08-05", "Source_type": "issuer earnings release", "Locator": "Operations Update, lines 156–161; continued AI artificial-lift optimization, with no fresh AI-specific result.", "Verification": "Full source opened and read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-008", "Company": "Chord Energy", "Title": "Chord Energy 2Q25 Investor Presentation", "URL": "https://www.chordenergy.com/wp-content/uploads/2025/08/CHRD-2Q25-Earnings-Presentation.pdf", "Source_date": "2025-08-06", "Source_type": "issuer investor presentation", "Locator": "Slide 11 / parsed PDF page 10, lines 273–306.", "Verification": "Full 24-page PDF opened; lists AI/ML scheduling and artificial-lift optimization, gas-lift algorithms, predictive maintenance, and remote monitoring as continuous-improvement initiatives; no distinct deployed AI result or attribution.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-009", "Company": "Devon Energy", "Title": "Devon Energy 2025 Sustainability Report", "URL": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "Source_date": "2025 (report; covers calendar 2024 unless noted)", "Source_type": "Issuer disclosure", "Locator": "printed p. 15 / PDF p. 15, lines 1228-1261; full 90-page PDF read", "Verification": "Document accessible on 1 October 2026. New review: printed p. 9 / PDF p. 9, lines 713-738; full 90-page PDF read. Earlier claim-level findings remain scoped to their original review.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01", "Followup_source_review": {"date": "2026-10-01", "status": "Latest access review: original document available; printed p. 9 / PDF p. 9, lines 713-738; full 90-page PDF read", "note": "Original source opened; relevant passage read."}}, {"Source_ID": "SW7-010", "Company": "Diamondback Energy", "Title": "Diamondback Energy 2025 Corporate Sustainability Report", "URL": "https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418", "Source_date": "2025 (2024 performance reporting)", "Source_type": "Issuer disclosure", "Locator": "printed p. 20 / PDF p. 20, lines 1422-1438; full 51-page PDF read", "Verification": "Original source opened; relevant passage read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-011", "Company": "Diamondback Energy", "Title": "Diamondback Energy Letter to Stockholders (Q3 2025)", "URL": "https://ir.diamondbackenergy.com/node/17156/pdf", "Source_date": "2025-11-03", "Source_type": "Issuer disclosure", "Locator": "PDF p. 1, lines 61-64; full 3-page PDF read", "Verification": "Original source opened; relevant passage read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-012", "Company": "Diamondback Energy", "Title": "Diamondback Energy Letter to Stockholders (Q1 2026)", "URL": "https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11", "Source_date": "2026-05-04", "Source_type": "Issuer disclosure", "Locator": "lines 72-77, especially line 74; full release read", "Verification": "Original source opened; relevant passage read.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-01"}, {"Source_ID": "SW7-013", "Company": "Expand Energy", "Title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "URL": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "Source_date": "No publication/update date displayed; accessed 2026-10-01", "Source_type": "Vendor customer case", "Locator": "Digging deeper into operational efficiency; Energizing the future", "Verification": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "Notes": "Source-attributed evidence. Deployment, reporting period and economics distinguished.", "Accessed": "2026-10-06"}, {"Source_ID": "SW7-014", "Company": "ExxonMobil", "Title": "ExxonMobil Guyana: drilling automation in deepwater operations", "URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "Source_date": "Undated (2026 URL)", "Source_type": "Vendor customer case", "Publisher": "SLB", "Locator": "Results: Autonomous drilling and rig automation drive record-setting performance in Guyana", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained."}}, {"Company": "Devon Energy", "Title": "Devon Energy — Snowflake + Whitson", "URL": "https://www.snowflake.com/en/customers/all-customers/case-study/devon-energy/", "Source_date": "Undated supplemental source; see review for dated corroboration", "Source_type": "Vendor/customer case or talk", "Locator": "Platform comparison; linked full source", "Verification": "Undated story. $50,000 sensor comparison is not $250m of realized savings. Geoscience AI/ML appears as work in progress.", "Notes": "Operating physics workflow; AI not established", "Accessed": "2026-10-01", "Source_ID": "SW7-015"}, {"Company": "bp", "Title": "bp — Databricks Genie / Unity Catalog", "URL": "https://answers.databricks.com/accelerating-maintenance-conversational-ai-bp-uses-genie-databricks-WsDSWkNLmZs", "Source_date": "Undated supplemental source; see review for dated corroboration", "Source_type": "Vendor/customer case or talk", "Locator": "Platform comparison; linked full source", "Verification": "Mixed upstream/downstream. Demo used mock data. No exact publication date, maintenance-outcome measure or project cost verified.", "Notes": "Customer-presenter account; date unresolved", "Accessed": "2026-10-01", "Source_ID": "SW7-016"}, {"Company": "Continental Resources", "Title": "Owner-relations agent", "URL": "https://start.uipath.com/How-Built-a-UiPath-Agent-to-Automate-Owner-Relations.html", "Source_date": "Undated landing page; webinar promoted for 2026-07-14", "Source_type": "Vendor webinar promotion", "Locator": "Landing page reviewed; recording not reviewed", "Verification": "Landing page undated, recording not reviewed, no quantified outcome. Land, lease, suspense-release and JIB extensions remain opportunities, not confirmed deployments.", "Notes": "UiPath promotes a deployed royalty-workflow agent with human approvals. A July 14, 2026 webinar is advertised; it does not establish go-live or each listed task as deployed.", "Accessed": "2026-10-01", "Source_ID": "SW7-017"}, {"Source_ID": "SW8-001", "Source_date": "2026-08-05", "Company": "Devon Energy", "Source_type": "Operator", "Title": "Devon Energy — deployment evidence", "URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Locator": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-002", "Source_date": "2026-05-28 (issuer release announcing 2025 report); IR report index separately lists June 17, 2026", "Company": "Diversified Energy Company", "Source_type": "Issuer report", "Title": "Diversified Energy Company Demonstrates Proven Sustainability Leadership (2025 Sustainability Report release)", "URL": "https://ir.div.energy/news-events/us-press-releases/detail/225/diversified-energy-company-demonstrates-proven-sustainability-leadership", "Locator": "Official issuer release confirms report publication; report content fully opened/read in linked PDF. OneDEC/Snowflake/ServiceNow architecture: report pp. 62-63; data and controls, HITL, audit, AI governance: p. 63; statement automation, historian surveillance and pricing/settlements: p. 63. Report data period is calendar 2025 (p. 67).", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-003", "Source_date": "2025-06-24", "Company": "EQT", "Source_type": "Issuer / partner disclosure", "Title": "EQT Releases 2024 ESG Report, 'Promises Made, Promises Delivered'", "URL": "https://ir.eqt.com/newsroom/news-releases/news-release-details/2025/EQT-Releases-2024-ESG-Report-Promises-Made-Promises-Delivered/default.aspx", "Locator": "Issuer release dates the report; complete official report PDF reviewed at https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf, p. 34: completed 2024 production-data integration, ongoing framework, 2025 midstream and 2026 offset/verification milestones; report boundary/external assurance caveats p. 2 and p. 34.", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-004", "Source_date": "2025-06-24", "Company": "EQT", "Source_type": "Issuer / partner disclosure", "Title": "EQT 2024 ESG Report", "URL": "https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf", "Locator": "Printed p. 34 / PDF page 33; boundaries on p. 2", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-005", "Source_date": "2025-04-04", "Company": "EQT", "Source_type": "Issuer / partner disclosure", "Title": "EQT Achieves Net Zero Scope 1 & 2 Emissions", "URL": "https://contextlabs.com/eqt-net-zero/", "Locator": "Full vendor customer page reviewed; lines 55-79 identifies the Context Labs platform and claims its role as data backbone for EQT Net Zero Now+; use only to describe vendor-stated product role, not to date specific sub-workflows or attribute emissions outcome to AI.", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-006", "Source_date": "2026-08-04", "Company": "ADNOC / SLB", "Source_type": "Source review", "Title": "ADNOC / SLB — engineering effort", "URL": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "Locator": "", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-007", "Source_date": "2024-08-19", "Company": "ExxonMobil / XTO Energy", "Source_type": "Source review", "Title": "ExxonMobil / XTO Energy — oil production uplift", "URL": "https://onepetro.org/SPEALCE/proceedings-abstract/24ALCE/24ALCE/D031S008R001/550379", "Locator": "", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "SW8-008", "Source_date": "2026-02-25", "Company": "Chord Energy", "Source_type": "Source review", "Title": "Chord Energy — rod pump runtime improvement", "URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "Locator": "", "Short_excerpt": "", "Verification": "Relevant source passages reviewed", "Notes": "Source establishes attributed disclosure; not independent outcome validation.", "Accessed": "2026-10-04"}, {"Source_ID": "OCT5-S001", "Source_date": "2025-04-17", "Company": "Imperial Oil", "Source_type": "Investor day presentation (Form 8-K Exhibit 99.1)", "Title": "Upstream ML well dashboard and AI shovel-truck analytics in oil sands", "URL": "https://www.sec.gov/Archives/edgar/data/49938/000004993825000022/a2025irday-finalforexter.htm", "Locator": null, "Short_excerpt": "Machine learning dashboard for > 4,500 wells", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S002", "Source_date": "2025-05-12", "Company": "ENEOS Holdings", "Source_type": "Company disclosure (medium-term management plan, TSE)", "Title": "Group AI organization and AI plant-operation expansion in medium-term plan", "URL": "https://www.hd.eneos.co.jp/english/news/release_information/upload_pdf/upload/20250512_05_04_2009288.pdf", "Locator": null, "Short_excerpt": "Launch a dedicated organization to promote the utilization of AI and pursue the potential of AI utilization across all operations, including the supply chain", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S003", "Source_date": "2026-09-16", "Company": "SK Innovation", "Source_type": "News article (company statements at WAVE2026 expo)", "Title": "Dual Brain AI anomaly detection across Ulsan refinery process units", "URL": "https://en.sedaily.com/finance/2026/09/16/sk-innovation-turns-to-ai-to-lift-refinery-output-by-100", "Locator": null, "Short_excerpt": "AI detects abnormal signs in processes and equipment and proposes responses, and field engineers then review the situation comprehensively before making the final decision.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S004", "Source_date": "2026-07-16", "Company": "GS Caltex", "Source_type": "News article", "Title": "AI surveillance cameras for safety during Yeosu refinery turnaround", "URL": "https://www.koreaherald.com/article/10811139", "Locator": null, "Short_excerpt": "The cameras detect various risk factors in real time, including whether staff members are wearing protective gear and when they approach heavy equipment.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S005", "Source_date": "2025-12-22", "Company": "Sonatrach", "Source_type": "News article (CEO site visit)", "Title": "ENTP EYES AI-assisted real-time remote drilling-rig monitoring center", "URL": "https://www.jeune-independant.net/daoudi-a-lentp-de-hassi-messaoud-lia-pour-optimiser-les-operations-de-forage/", "Locator": null, "Short_excerpt": "un système de télé-monitoring en temps réel, basé sur une solution intégrant l'intelligence artificielle", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S006", "Source_date": "2026-07-30", "Company": "Tourmaline Oil", "Source_type": "Earnings call highlights (GuruFocus via Yahoo Finance)", "Title": "Machine learning informs Montney well landing and completion design", "URL": "https://finance.yahoo.com/energy/articles/tourmaline-oil-corp-trmlf-q2-230031565.html", "Locator": null, "Short_excerpt": "learnings from machine learning on landing and completion optimization", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S007", "Source_date": "2026-06-18", "Company": "TotalEnergies", "Source_type": "operator press release", "Title": "MethaneLive methane monitoring centre; agentic AI to target top emitters", "URL": "https://www.sec.gov/Archives/edgar/data/0000879764/000110465926079516/tm2619405d1_ex99-4.htm", "Locator": null, "Short_excerpt": "This foundation now enables the deployment of agentic AI solutions to go further, by more effectively targeting the highest-emitting equipment", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S008", "Source_date": "2026-05-13", "Company": "Shell", "Source_type": "annual report", "Title": "T-Pulse AI CCTV monitoring of unsafe acts, conditions and leaks", "URL": "https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section/promo.multi.stream/1779352356739/36306d968747b6079d3f800b2b1552a033856b5d/shell-integrated-annual-and-sustainability-report.pdf", "Locator": null, "Short_excerpt": "We continue to expand the deployment of T-Pulse, an AI-automated safety monitoring solution developed by Detect Technologies.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S009", "Source_date": "2026-05-13", "Company": "Shell", "Source_type": "annual report", "Title": "Shell plc Annual Report and Accounts 2025 (Innovation and Technology – 'Pearl GTL in Qatar recognised for innovation', p.73 of PDF)", "URL": "https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section/promo.multi.stream/1779352356739/36306d968747b6079d3f800b2b1552a033856b5d/shell-integrated-annual-and-sustainability-report.pdf", "Locator": null, "Short_excerpt": "The plant employs machine-vision systems to detect hazards in real time and alert supervisors, helping prevent safety incidents.", "Verification": "full text reviewed", "Notes": "Supplemental source, 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S010", "Source_date": "2026-03-09", "Company": "bp", "Source_type": "sustainability report", "Title": "AI safety-observation pilot at Whiting refinery; LENS AI assistant", "URL": "https://www.bp.com/api/files/6cqieuqhq4no/master/6ISrW0tRFKSHbaBT7yeVNf/18b0bc82f1dcf6b0b4d1df187477f8c8/bp-sustainability-report-2025.pdf", "Locator": null, "Short_excerpt": "At our Whiting refinery in the US, we launched a pilot study of an AI-driven safety observation tool to monitor situations and activities", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S011", "Source_date": "2026-07-20", "Company": "EQT", "Source_type": "sustainability report", "Title": "AI-based ergonomic assessments to reduce musculoskeletal risk", "URL": "https://cdn.prod.website-files.com/6751b0d7add5edb1f0141aef/6ab53525a5a0e486b918d352_EQT-2025-Sustainability-Report.pdf", "Locator": "Printed/PDF p.59, EHS Management System and Risk Identification; use began during 2025. Independently rechecked 9 October 2026.", "Short_excerpt": "In 2025, our team began using AI technologies to perform ergonomic assessments, helping evaluate body positioning during common tasks", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review 9 October 2026 primary recheck confirms ergonomic use in 2025; no AI-attributed injury outcome.", "Accessed": "2026-10-05", "Correction_note": "9 October 2026 primary recheck confirms ergonomic use in 2025; no AI-attributed injury outcome."}, {"Source_ID": "OCT5-S012", "Source_date": "2026-06-03", "Company": "Eni", "Source_type": "annual report", "Title": "Safety Presence AI predicting recurring hazards from safety-data weak signals", "URL": "https://report.eni.com/annual-report-2025/en/_assets/downloads/sustainability-eni-ar25.pdf", "Locator": null, "Short_excerpt": "use of an artificial intelligence tool (Safety Presence), solution capable of predicting recurring hazardous situations based on weak signals, recorded in safety databases.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S013", "Source_date": "2025-04-07", "Company": "Woodside", "Source_type": "regulatory filing", "Title": "HaloGuard machine-vision red-zone system on contracted drilling rig", "URL": "https://www.woodside.com/docs/default-source/asx-announcements/2025/021-sustainability-briefing-2025-transcript.pdf?sfvrsn=91b6b681_3", "Locator": null, "Short_excerpt": "The HaloGuard system combines a wearable alarm, real time location transmitter and machine vision system to track the position of personnel and equipment", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S014", "Source_date": "2025-09-23", "Company": "SLB", "Source_type": "regulatory filing", "Title": "EPA-approved SOOFIE continuous methane monitoring with AI event detection", "URL": "https://www.epa.gov/system/files/documents/2025-09/slb-soofie-approval-letter-and-justification-memo.pdf", "Locator": null, "Short_excerpt": "all methane concentration measurements collected across the SOOFIE sensor network are processed through a proprietary AI-based event detection algorithm.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S015", "Source_date": "2025-05-13", "Company": "Clean Connect AI", "Source_type": "regulatory filing", "Title": "EPA-approved LeakFinder autonomous OGI with computer-vision methane detection", "URL": "https://www.epa.gov/system/files/documents/2025-05/clean-connect-alternative-test-method-signed.pdf", "Locator": null, "Short_excerpt": "autonomously detects any methane and/or VOCs emission events imaged by the OGI through proprietary computer vision models, as they occur", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S016", "Source_date": "2025-06-16", "Company": "Percepto", "Source_type": "vendor press release", "Title": "AI Emission Detector analysing autonomous-drone OGI video for methane", "URL": "https://percepto.com/percepto-launches-ai-emission-detector-for-remote-drone-based-methane-surveying/", "Locator": null, "Short_excerpt": "Percepto’s AI analyzes the footage to identify and flag potential methane emissions.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S017", "Source_date": "2025-10-15", "Company": "Chevron", "Source_type": "Vendor press release with named customer quote", "Title": "Arkestro predictive AI procurement expanded across global business units", "URL": "https://arkestro.com/press-releases/arkestro-announces-customer-expansion-with-chevron-driving-predictive-procurement-transformation-across-global-markets/", "Locator": null, "Short_excerpt": "Arkestro's platform provides the predictive capabilities and visibility we need to navigate market complexity and unlock bottom-line value.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S018", "Source_date": "2025-09-19", "Company": "Equinor", "Source_type": "Trade-press article based on vendor customer story", "Title": "AI anomaly detection on robot inspection data at Sture and Mongstad", "URL": "https://www.technologyrecord.com/article/cognite-and-equinor-leverage-robotics-and-ai-to-improve-field-operations", "Locator": null, "Short_excerpt": "The robots collect and stream data to Cognite Data Fusion, where AI analyses it to detect anomalies.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S019", "Source_date": "2026-03-23", "Company": "Aker BP", "Source_type": "Trade press (World Oil)", "Title": "Cognite–NVIDIA NV-Tesseract anomaly detection across wells and equipment", "URL": "https://worldoil.com/news/2026/3/23/ai-anomaly-detection-gains-traction-offshore-as-aker-bp-deploys-cognite-nvidia-solution/", "Locator": null, "Short_excerpt": "Cognite has integrated NVIDIA's NV-Tesseract time-series AI models into its industrial data platform, with Aker BP deploying the combined solution", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S020", "Source_date": "2026-04-24", "Company": "ONGC", "Source_type": "Vendor earnings release (SEC 8-K exhibit)", "Title": "SLB Lumi data and AI platform across all ONGC assets", "URL": "https://www.sec.gov/Archives/edgar/data/0000087347/000119312526174940/d111579dex99.htm", "Locator": null, "Short_excerpt": "SLB's Lumi data and AI platform will be deployed as an enterprise solution across all ONGC assets", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S021", "Source_date": "2026-06-09", "Company": "Pampa Energía", "Source_type": "Vendor press release", "Title": "Halliburton DecisionSpace 365 AI-enabled digital transformation in Vaca Muerta", "URL": "https://www.halliburton.com/en/about-us/press-release/pampa-energia-selects-halliburton-to-support-enterprise-digital-transformation", "Locator": null, "Short_excerpt": "connecting trusted data, proven science, and industrial grade AI in the workflows teams use", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S022", "Source_date": "2025-05-18", "Company": "Aramco", "Source_type": "Trade press (JPT) summarising operator announcement", "Title": "NVIDIA MoU for industrial AI computing, AI hub and robotics", "URL": "https://jpt.spe.org/aramco-signs-mous-worth-up-to-90-billion-for-lng-ai-emissions-reduction-and-asset-management-initiatives", "Locator": null, "Short_excerpt": "Aramco and NVIDIA to collaborate on developing advanced industrial AI computing infrastructure.", "Verification": "full text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S023", "Source_date": "2025-05-18", "Company": "Aramco", "Source_type": "Trade press (JPT) summarising operator announcement", "Title": "Aramco Signs MOUs Worth Up to $90 Billion for LNG, AI, Emissions Reduction, and Asset Management Initiatives", "URL": "https://jpt.spe.org/aramco-signs-mous-worth-up-to-90-billion-for-lng-ai-emissions-reduction-and-asset-management-initiatives", "Locator": null, "Short_excerpt": "Aramco Digital and Qualcomm have signed an MOU to explore a strategic collaboration focused on digital transformation.", "Verification": "full text reviewed", "Notes": "Supplemental source, 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S024", "Source_date": "2026-08-12", "Company": "Presidio Production", "Source_type": "SEC-filed operator earnings-call transcript", "Title": "Presidio Production Company Q2 2026 Earnings Call Transcript", "URL": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "Locator": "Chris Hammack remarks, printed page 3: weekend coverage, 24 plunger boxes, human judgment; Will Ulrich page 2 for aggregate uplift/revenue", "Short_excerpt": "None of these replace field judgment", "Verification": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "Notes": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage.", "Accessed": "2026-10-18", "Publisher": "Presidio Production", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original text read; targeted workflow/metric check", "independent_validation": "Not established", "note": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage."}}, {"Source_ID": "OCT5-S025", "Source_date": "2025-11-03", "Company": "ADNOC", "Source_type": "SPE technical paper authored by ADNOC Onshore and AIQ staff", "Title": "Autonomous gas-lift control across 200 onshore producer wells", "URL": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/D021S043R003/793441", "Locator": "SPE-230130-MS abstract, deployment results and challenges paragraphs", "Short_excerpt": null, "Verification": "Original OnePetro abstract reviewed through search-indexed text; full paper not accessed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S026", "Source_date": "2026-02-25", "Company": "Northern Oil and Gas (NOG)", "Source_type": "Operator investor presentation", "Title": "Drakkar AI and data platform for investment and portfolio decisions", "URL": "https://s203.q4cdn.com/492547254/files/doc_financials/2025/q4/NOG-Earnings-Presentation-Q4-YE-2025-FINAL.pdf", "Locator": "Slide 21, Drakkar; cover date February 25, 2026", "Short_excerpt": null, "Verification": "Issuer-hosted presentation text reviewed", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S027", "Source_date": "2026-05-28", "Company": "Enbridge", "Source_type": "Operator sustainability report", "Title": "Pipeline deformation prioritization with machine learning", "URL": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "Locator": "Report printed page 28, LP: Using machine learning to prioritize activities", "Short_excerpt": null, "Verification": "Operator report page 28 reviewed through original publisher search-indexed text", "Notes": "Added 5 October 2026 research review", "Accessed": "2026-10-05"}, {"Source_ID": "OCT5-S028", "Source_date": "2026-04-30", "Company": "Quantum Capital Group", "Source_type": "Microsoft customer story", "Title": "Land-title research at an unnamed Quantum portfolio company", "URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Locator": "From the field to the back office and beyond — land-title research paragraphs", "Short_excerpt": null, "Verification": "Original land-title passage reopened 10 October 2026; prior full-document review 6 October 2026. Results remain vendor-reported.", "Notes": "Prior locator referred to field planning on the same page. Corrected to the land-title section after a fresh passage check; no outcome classification changed.", "Accessed": "2026-10-10"}, {"Source_ID": "OCT5-S029", "Source_date": "2025-04-15", "Company": "PRIO", "Source_type": "Microsoft customer story", "Title": "Copilot-assisted HR job descriptions and competency mapping", "URL": "https://news.microsoft.com/pt-br/prio-utiliza-inteligencia-artificial-no-rh-e-economiza-mais-de-70-horas-por-setor/", "Locator": "PRIO HR workflow and statements by Marcelo Ludwig", "Short_excerpt": null, "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S030", "Source_date": "2025-04-07", "Company": "Kiewit", "Source_type": "Company-owned pilot account", "Title": "T-Pulse safety-camera pilot at an offshore fabrication yard", "URL": "https://www.kiewit.com/newsroom/a-different-kind-of-pilot/", "Locator": "KOS Ingleside pilot; accuracy and customization paragraphs", "Short_excerpt": null, "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S031", "Source_date": "2025-11-07", "Company": "Canadian Natural Resources", "Source_type": "Operator investor presentation", "Title": "AI-assisted oil-sands asset-integrity monitoring", "URL": "https://www.cnrl.com/content/uploads/2025/11/W_CNQ-Investor-Open-House_Nov-7-2025-DL.pdf", "Locator": "Investor Open House, printed slide 51 (PDF page 29, upper half)", "Short_excerpt": null, "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S032", "Source_date": "2026-04-30", "Company": "Quantum Capital Group", "Source_type": "Microsoft customer story", "Title": "Geospatial field planning for an unnamed Quantum portfolio operator", "URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Locator": "Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph", "Short_excerpt": "algorithmic improvements have reduced solve times by 50 to 65%", "Verification": "Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-21", "Publisher": "Microsoft", "Current_review": {"last_checked": "2026-10-21", "source_read": "Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation.", "independent_validation": "Not established", "note": "Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model."}}, {"Source_ID": "OCT5-S033", "Source_date": "2026 annual report covering 2025; exact publication day not confirmed", "Company": "EOG Resources", "Source_type": "Operator annual report", "Title": "Proprietary generative AI for collaboration and operational data", "URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "Short_excerpt": null, "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S034", "Source_date": "2026 annual report covering 2025; exact publication day not confirmed", "Company": "EOG Resources", "Source_type": "Operator annual report", "Title": "Proprietary machine-learning production optimizer", "URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S035", "Source_date": "2026-02-23", "Company": "Ovintiv", "Source_type": "SEC-filed operator annual report", "Title": "Proprietary AI for drilling and centralized operations", "URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "Short_excerpt": null, "Verification": "Full original document reviewed on 6 October 2026", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S036", "Source_date": "2026-02-23", "Company": "Ovintiv", "Source_type": "SEC-filed operator annual report", "Title": "Real-time AI optimization of fracturing", "URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S037", "Source_date": "Undated source; see chronology", "Company": "TC Energy", "Source_type": "AWS customer story", "Title": "Pipeline IQ engineering-document verification agents", "URL": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "Locator": "PIPER, SCOUT and outcome sections", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S038", "Source_date": "Undated case; referenced 2025-07-17", "Company": "Baker Hughes", "Source_type": "EPAM delivery-partner client case study", "Title": "GenAI technical-support assistant for artificial-lift field engineers", "URL": "https://www.epam.com/services/client-work/leveraging-genai-to-increase-productivity", "Locator": "Artificial Lift Systems assistant; Results; supporting EPAM newsroom item dated 17 July 2025", "Short_excerpt": null, "Verification": "Full partner case and dated related media summary reviewed in three Luna rounds and editor check, 6 October 2026.", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S039", "Source_date": "2025-07-17", "Company": "Baker Hughes", "Source_type": "EPAM-hosted related media summary", "Title": "How AI Accelerates Oil Production and Saves Thousands of Hours of Support", "URL": "https://www.epam.com/about/newsroom/in-the-news/2025/how-ai-accelerates-oil-production-and-saves-thousands-of-hours-of-support", "Locator": "Dated heading and artificial-lift assistant paragraphs; link to client case", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S040", "Source_date": "2026-05", "Company": "Petrobras", "Source_type": "Issuer management report", "Title": "Petrobras Relatório de Gestão 2025", "URL": "https://transparencia.petrobras.com.br/documents/41711919/42223019/Relatorio%2Bde%2BGestao%2B-%2B2025.pdf/9ef35500-1f38-0cb6-539d-072b0569514f?download=true&t=1779977455000&version=1.0", "Locator": "2025 innovation highlights, physical/printed p. 133", "Short_excerpt": null, "Verification": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S041", "Source_date": "Undated; reviewed 2026-10-06", "Company": "Expand Energy", "Source_type": "Snowflake customer case study", "Title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "URL": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "Locator": "Digging deeper into operational efficiency; Energizing the future", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S042", "Source_date": "2025 (2024 report; exact publication day not stated)", "Company": "ADNOC Gas", "Source_type": "Issuer integrated annual report", "Title": "ADNOC Gas Integrated Report 2024", "URL": "https://www.adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas-integrated-report-2024-english.ashx", "Locator": "Printed pp. 26–27, PDF physical p. 15; CPAD deployment and roadmap", "Short_excerpt": null, "Verification": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06"}, {"Source_ID": "OCT5-S043", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Operator annual report", "Title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics", "Short_excerpt": null, "Verification": "Relevant original issuer source independently reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Relevant original issuer annual-report pages independently reviewed", "publisher": "Aramco", "locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics"}, "Publisher": "Aramco", "Current_review": {"source_read": "Technology Realized Value definition and AI subtotal reviewed", "current_access": "Original annual-report section reopened", "last_checked": "2026-10-10", "independent_validation": "Issuer describes third-party verification of aggregate TRV; no assurance opinion reviewed", "note": "Aramco describes its aggregate TRV metric as third-party verified and incremental to a business baseline. No independent assurance opinion was reviewed here. The $2.6bn AI subtotal is company-reported; AI-only allocation by project or benefit type and full costs are not disclosed. TRV includes capital or operating expenditure avoided and revenue added. Capital avoidance is not incremental cash income, and aggregate TRV is not project-level net cash ROI."}}, {"Source_ID": "OCT5-S044", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Operator results presentation", "Title": "Saudi Aramco FY 2025 Results presentation", "URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "Locator": "“Pioneering AI & technology: progressing value-creation opportunities,” slide 10 / PDF p. 10; portfolio TRV value-type mix (not allocated to AI)", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Official operator FY2025 results presentation independently reviewed", "publisher": "Aramco", "locator": "“Pioneering AI & technology: progressing value-creation opportunities,” slide 10 / PDF p. 10; portfolio TRV value-type mix (not allocated to AI)"}, "Publisher": "Aramco"}, {"Source_ID": "OCT5-S045", "Source_date": "2026-09-15", "Company": "TotalEnergies", "Source_type": "Operator press release", "Title": "TotalEnergies and Mistral announce a three-year frontier subsurface AI program", "URL": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "Locator": "Release dated 2026-09-15; >€100m / 3-year program; “petaflops of data” wording flagged", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Original issuer release independently reviewed", "publisher": "TotalEnergies", "locator": "Release dated 2026-09-15; >€100m / 3-year program; “petaflops of data” wording flagged"}, "Publisher": "TotalEnergies"}, {"Source_ID": "OCT5-S046", "Source_date": "2026-02-11", "Company": "TotalEnergies", "Source_type": "Operator results-call transcript", "Title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "URL": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf", "Locator": "2026-02-11 transcript, PDF pp. 9–10 / lines 358–385", "Short_excerpt": null, "Verification": "Relevant original primary source reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Relevant original issuer transcript pages independently reviewed", "publisher": "TotalEnergies", "locator": "2026-02-11 transcript, PDF pp. 9–10 / lines 358–385"}, "Publisher": "TotalEnergies"}, {"Source_ID": "OCT5-S047", "Source_date": "2026-02-09", "Company": "ADNOC Gas", "Source_type": "Operator results presentation", "Title": "ADNOC Gas Q4/FY2025 Results Presentation", "URL": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "Locator": "2026-02-09, slide 7 / PDF p. 6; disclaimer PDF p. 2", "Short_excerpt": null, "Verification": "Relevant original primary source reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Official operator presentation text and slide layout independently reviewed", "publisher": "ADNOC Gas", "locator": "2026-02-09, slide 7 / PDF p. 6; disclaimer PDF p. 2"}, "Publisher": "ADNOC Gas"}, {"Source_ID": "OCT5-S048", "Source_date": "2025-11-02", "Company": "ADNOC Gas", "Source_type": "Operator press release", "Title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "URL": "https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq", "Locator": "2025-11-02, paragraphs 107–125", "Short_excerpt": null, "Verification": "Relevant original primary source reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Original operator release independently reviewed", "publisher": "ADNOC Gas", "locator": "2025-11-02, paragraphs 107–125"}, "Publisher": "ADNOC Gas"}, {"Source_ID": "OCT5-S049", "Source_date": "2025-10-10", "Company": "ExxonMobil", "Source_type": "Operator article", "Title": "Introducing Vantage: A new lens on upstream operations", "URL": "https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations", "Locator": "2025-10-10, “Big numbers, bigger impact,” lines 54–64", "Short_excerpt": null, "Verification": "Relevant original primary source reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Original operator article independently reviewed", "publisher": "ExxonMobil", "locator": "2025-10-10, “Big numbers, bigger impact,” lines 54–64"}, "Publisher": "ExxonMobil"}, {"Source_ID": "OCT5-S050", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Operator annual report", "Title": "Aramco Intelligent Live Earth Model for AI-assisted geosteering", "URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics", "Short_excerpt": null, "Verification": "Relevant original annual-report pages reviewed on 2026-10-06.", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Relevant original issuer annual-report pages independently reviewed", "publisher": "Aramco", "locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics"}, "Publisher": "Aramco"}, {"Source_ID": "OCT5-S051", "Source_date": "2026-03-10", "Company": "Aramco", "Source_type": "Operator annual report", "Title": "Aramco PE.AI advanced petrophysical prediction and synthetic well logs", "URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics", "Short_excerpt": null, "Verification": "Relevant original annual-report pages reviewed on 2026-10-06.", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Relevant original issuer annual-report pages independently reviewed", "publisher": "Aramco", "locator": "Annual Report 2025, printed pp. 64–65 / PDF pp. 37–38; TRV definition and AI/non-AI split; p. 64 named AI workflow metrics"}, "Publisher": "Aramco"}, {"Source_ID": "OCT5-S052", "Source_date": "2026-05-06", "Company": "TotalEnergies", "Source_type": "Operator press release", "Title": "Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D", "URL": "https://totalenergies.com/newsroom/totalenergies-develops-pangea-5-a-next-generation-supercomputer-469512/?lang=eng", "Locator": "2026-05-06, paragraphs 38–50", "Short_excerpt": null, "Verification": "Relevant original issuer source independently reviewed on 2026-10-06.", "Notes": "Added 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "Original operator release independently reviewed", "publisher": "TotalEnergies", "locator": "2026-05-06, paragraphs 38–50"}, "Publisher": "TotalEnergies"}, {"Source_ID": "OCT5-S053", "Source_date": "Undated; accessed 2026-10-06", "Company": "Spearpoint Resources", "Source_type": "First-party vendor offer page", "Title": "Joltly Pricing — Operators & Non-Op / Royalty Owners", "URL": "https://www.joltly.io/pricing", "Locator": "Undated page accessed 2026-10-06; pricing cards and FAQ", "Short_excerpt": null, "Verification": "full text reviewed", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "First-party vendor offer page reviewed", "publisher": "Joltly", "locator": "Undated page accessed 2026-10-06; pricing cards and FAQ"}, "Publisher": "Joltly"}, {"Source_ID": "OCT5-S054", "Source_date": "2026-08-08", "Company": "Spearpoint Resources", "Source_type": "First-party vendor workflow article", "Title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "URL": "https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas", "Locator": "2026-08-08; JIB/revenue-statement descriptions", "Short_excerpt": null, "Verification": "Full original vendor pages reviewed on 2026-10-06.", "Notes": "Supplemental source, 6 October 2026 research review", "Accessed": "2026-10-06", "Executive_source_review": {"reviewed_on": "2026-10-06", "status": "First-party vendor workflow article reviewed", "publisher": "Joltly", "locator": "2026-08-08; JIB/revenue-statement descriptions"}, "Publisher": "Joltly"}, {"Source_ID": "OCT9-QCG-2026", "Source_date": "2026", "Company": "Quantum Capital Group", "Source_type": "Sponsor stakeholder report", "Title": "Quantum 2026 Stakeholder Report", "URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Locator": "Pages 11–15; page 68 disclaimer", "Verification": "Relevant primary sections reviewed", "Notes": "Report information defaults to 31 December 2025 unless otherwise stated. Exact publication day not stated. Current naming and platform transition; performance remains sponsor-reported.", "Accessed": "2026-10-09"}, {"Source_ID": "OCT10-PRESIDIO-0817", "Source_date": "2026-08-17", "Company": "Presidio Production", "Source_type": "SEC-filed investor presentation", "Title": "Presidio August 2026 investor presentation", "URL": "https://ir.bypresidio.com/sec-filings/all-sec-filings/content/0001213900-26-090532/ea030229001ex99-1.htm", "Locator": "Slides 20–23; current surveillance versus 2026 deployment column on slide 21; GOAL slide 23", "Verification": "Relevant issuer-filed presentation text reviewed on 10 October 2026; column interpretation also recorded in the October 9 review.", "Accessed": "2026-10-10", "Notes": "Disclosure date is the SEC filing date. Roughly 2,300 platform wells appears on the GOAL slide; do not infer completed fleet-wide autonomous control.", "Current_review": {"source_read": "Relevant original presentation text reviewed", "current_access": "Issuer SEC-filings page reopened", "last_checked": "2026-10-10", "independent_validation": "Not established", "note": "Source review establishes issuer statements. Production targets, illustrative revenue and roadmap tools are distinct from achieved results."}}, {"Source_ID": "OCT10-AI-284-SWPSC", "Company": "BPX Energy", "Title": "Daily lift recommendations across a 441-well Permian pilot", "URL": "https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Publisher": "Southwestern Petroleum Short Course", "Locator": "SWPSC 2026051 abstract: method, 441-well asset and reported results", "Authors": "Xecta Digital Labs and BPX Energy; BPX coauthors Timothy Credeur and Kevin McNeilly", "Accessed": "2026-10-10", "Verification": "Relevant original source text reviewed; performance remains author-reported, not independently validated.", "Notes": "2026 is year precision only; exact publication day is not established.", "Current_review": {"last_checked": "2026-10-10", "source_read": "Original source reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT10-AI-285-SWPSC", "Company": "Burk Royalty", "Title": "Pump-by-exception monitoring and control at a West Texas lease", "URL": "https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Publisher": "Southwestern Petroleum Short Course", "Locator": "SWPSC 2026052 abstract; linked full paper PDF pp. 3, 7–9 (site configuration, eight-month results, economics and projections)", "Authors": "Mario Campos, Amplified Industries; Guy Tippy, Burk Royalty Co.", "Accessed": "2026-10-10", "Verification": "Relevant original source text reviewed; performance remains author-reported, not independently validated.", "Notes": "2026 is year precision only; exact publication day is not established.", "Current_review": {"last_checked": "2026-10-10", "source_read": "Original source reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT10-BURK-PDF", "Company": "Burk Royalty", "Title": "Field Optimization Reimagined: full paper", "URL": "https://www.swpshortcourse.org/sites/default/files/papers/2026052%20-%20Field%20Optimization%20Reimagined%20Data%20At%20The%20Core%20Exceptions%20At%20The%20Center%2C%20People%20In%20Control.pdf", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Publisher": "Southwestern Petroleum Short Course", "Locator": "PDF pp. 3, 7–9; figure on p. 8", "Authors": "Mario Campos, Amplified Industries; Guy Tippy, Burk Royalty Co.", "Accessed": "2026-10-10", "Verification": "Relevant original source text reviewed; performance remains author-reported, not independently validated.", "Notes": "2026 is year precision only; exact publication day is not established.", "Current_review": {"last_checked": "2026-10-10", "source_read": "Original source reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT10-AI-286-SWPSC", "Company": "Expand Energy", "Title": "Autonomous liquid unloading on nine Haynesville gas wells", "URL": "https://www.swpshortcourse.org/paper/2026001-autonomous-edge-based-optimization-liquid-loading-intermittent-gas-wells", "Source_date": "2026", "Source_type": "Operator/vendor-coauthored technical paper", "Publisher": "Southwestern Petroleum Short Course", "Locator": "SWPSC 2026001 abstract; JPT 1 July 2026, Haynesville Field Results", "Authors": "SLB and Expand Energy, including Expand coauthor E. Gies; related JPT account authored by SLB", "Accessed": "2026-10-10", "Verification": "Relevant original source text reviewed; performance remains author-reported, not independently validated.", "Notes": "2026 is year precision only; exact publication day is not established.", "Current_review": {"last_checked": "2026-10-10", "source_read": "Original source reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT10-EXPAND-JPT", "Company": "Expand Energy", "Title": "Edge Computing and IIoT Enable Autonomous Artificial Lift and Well Optimization", "URL": "https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments", "Source_date": "2026-07-01", "Source_type": "Supplier-authored technical account", "Publisher": "Journal of Petroleum Technology / SPE", "Locator": "Field Results: Haynesville Basin", "Authors": "SLB and Expand Energy, including Expand coauthor E. Gies; related JPT account authored by SLB", "Accessed": "2026-10-10", "Verification": "Relevant original source text reviewed; performance remains author-reported, not independently validated.", "Notes": "", "Current_review": {"last_checked": "2026-10-10", "source_read": "Original source reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT10-EOG-UTICA", "Company": "EOG Resources", "Title": "EOG August 2026 investor presentation", "URL": "https://investors.eogresources.com/image/EOG_0826.pdf", "Source_date": "2026-08-04", "Source_type": "Operator investor presentation", "Publisher": "EOG Resources", "Locator": "Printed slide 14: Outstanding Utica Performance", "Accessed": "2026-10-10", "Verification": "Original issuer deck reviewed; attachment indexed 4 August 2026 ahead of 5 August call.", "Notes": "The deck does not mention AI or machine learning. The 5% Utica base-performance result is related operating-technology context, not an AI-attributed result."}, {"Source_ID": "OCT12-FLY-PLATFORM", "Company": "Flywheel Energy", "Title": "Flywheel AI Platform Engineer", "URL": "https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750", "Source_date": "2026", "Source_type": "Employer recruitment disclosure", "Publisher": "Flywheel Energy / LinkedIn", "Locator": "Job Summary and Key Responsibilities", "Accessed": "2026-10-12", "Verification": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "Notes": "Year precision only; exact posting date and deployment start are not established.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "independent_validation": "Not established"}}, {"Source_ID": "OCT12-FLY-ACCOUNTING", "Company": "Flywheel Energy", "Title": "Flywheel AI Analyst — accounting automation", "URL": "https://www.linkedin.com/jobs/view/ai-analyst-at-flywheel-energy-llc-4472009913", "Source_date": "2026", "Source_type": "Employer recruitment disclosure", "Publisher": "Flywheel Energy / LinkedIn", "Locator": "Role Summary; Automation; Data Management; Required Qualifications", "Accessed": "2026-10-12", "Verification": "Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments.", "Notes": "Year precision only; exact posting date and deployment start are not established.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments.", "independent_validation": "Not established"}}, {"Source_ID": "OCT12-FLY-FEED", "Company": "Flywheel Energy", "Title": "Flywheel recruitment update: land, accounting and AI platform", "URL": "https://www.linkedin.com/company/flywheel-energy", "Source_date": "2026", "Source_type": "Employer recruitment disclosure", "Publisher": "Flywheel Energy / LinkedIn", "Locator": "AI recruitment update naming AI Analyst–Land, AI Analyst–Accounting and AI Platform Engineer", "Accessed": "2026-10-12", "Verification": "Company-authored LinkedIn feed update reviewed 12 October 2026; relative date only. Feed position may change.", "Notes": "Year precision only; exact posting date and deployment start are not established.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Company-authored LinkedIn feed update reviewed 12 October 2026; relative date only. Feed position may change.", "independent_validation": "Not established"}}, {"Source_ID": "OCT12-FLY-QUORUM", "Company": "Flywheel Energy", "Title": "Quorum On Demand Land — Flywheel testimonial", "URL": "https://www.quorumsoftware.com/solutions/upstream-on-demand/land/", "Source_date": "Undated", "Source_type": "Vendor customer testimonial", "Publisher": "Quorum Software", "Locator": "Proven for High-Growth Land Organizations; Flywheel customer testimonial", "Accessed": "2026-10-12", "Verification": "Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel.", "Notes": "Undated page; does not establish an in-window rollout.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel.", "independent_validation": "Not established"}}, {"Source_ID": "OCT12-FLY-INTERN", "Company": "Flywheel Energy", "Title": "Flywheel AI Intern — Summer 2027", "URL": "https://www.linkedin.com/jobs/view/ai-intern-summer-2027-at-flywheel-energy-llc-4461575049", "Source_date": "2026", "Source_type": "Employer recruitment disclosure", "Publisher": "Flywheel Energy / LinkedIn", "Locator": "Indexed employer description: About the AI Internship and What You Will Do", "Accessed": "2026-10-12", "Verification": "Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source.", "Notes": "Year precision only; exact posting date and deployment start are not established.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source.", "independent_validation": "Not established"}}, {"Source_ID": "OCT12-FLY-COMPANY", "Company": "Flywheel Energy", "Title": "Flywheel Energy — company and operating partner context", "URL": "https://www.flywheelenergy.com/", "Source_date": "Undated", "Source_type": "Company website", "Publisher": "Flywheel Energy", "Locator": "Private E&P description; Stone Ridge footer", "Accessed": "2026-10-12", "Verification": "Official company page reviewed 12 October 2026. Identifies Flywheel as an operating partner of Stone Ridge Energy; does not establish a precise legal ownership schedule or internal AI adoption.", "Notes": "Undated page; does not establish an in-window rollout.", "Current_review": {"last_checked": "2026-10-12", "source_read": "Official company page reviewed 12 October 2026. Identifies Flywheel as an operating partner of Stone Ridge Energy; does not establish a precise legal ownership schedule or internal AI adoption.", "independent_validation": "Not established"}}, {"Source_ID": "OCT14-CNX-SAFETY", "Company": "CNX Resources", "Title": "CNX Workplace Health and Safety", "URL": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "Source_date": "Undated", "Source_type": "Company website", "Publisher": "CNX Resources", "Locator": "Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections", "Accessed": "2026-10-14", "Verification": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Notes": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original text reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT14-AKER-ABB", "Company": "Aker BP", "Title": "Cognite and ABB collaborate on industrial agentic applications", "URL": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "Source_date": "2026-05-21", "Source_type": "Vendor announcement", "Publisher": "Cognite", "Locator": "Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits", "Accessed": "2026-10-14", "Verification": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Notes": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original text reviewed", "independent_validation": "Not established"}}, {"Source_ID": "OCT14B-CNPC-XINJIANG", "Company": "PetroChina", "Title": "CNPC: industrial large model for rod-pumped-well production optimization", "URL": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "Source_date": "2026-04-15", "Source_type": "Company news article", "Publisher": "CNPC / China Petroleum News", "Locator": "Article dated 15 April 2026; first four body paragraphs", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain."}}, {"Source_ID": "OCT14B-CNX-METHANE", "Company": "CNX Resources", "Title": "CNX Commitment to Methane Reduction", "URL": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "Source_date": "Undated", "Source_type": "Company website", "Publisher": "CNX Resources", "Locator": "2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority."}}, {"Source_ID": "OCT14B-HG-INVOICES", "Company": "HG Oil Holdings", "Title": "HG Oil Holdings: AI invoice processing", "URL": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "Source_date": "Undated", "Source_type": "Vendor customer case", "Publisher": "Preisser Solutions", "Locator": "Before; What we built; Operational handoffs; Outcomes", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata."}}, {"Source_ID": "OCT14B-HG-IDENTITY", "Company": "HG Oil Holdings", "Title": "HG Oil Holdings About", "URL": "https://www.hgoilholdings.com/about", "Source_date": "Undated", "Source_type": "Company website", "Publisher": "HG Oil Holdings", "Locator": "Company description; current company-wide wells are not AI scope", "Accessed": "2026-10-14", "Verification": "Original page reviewed 14 October 2026; adjacent context, not independent outcome validation.", "Notes": "Identity and general control context do not establish invoice deployment dates or measured return."}, {"Source_ID": "OCT14B-HG-CONTROLS", "Company": "HG Oil Holdings", "Title": "Preisser AI invoice processing for small businesses", "URL": "https://preissersolutions.com/use-cases/ai-invoice-processing-small-business", "Source_date": "Undated", "Source_type": "Vendor product page", "Publisher": "Preisser Solutions", "Locator": "Human payment approval; names HG as reference; general product controls", "Accessed": "2026-10-14", "Verification": "Original page reviewed 14 October 2026; adjacent context, not independent outcome validation.", "Notes": "Identity and general control context do not establish invoice deployment dates or measured return."}, {"Source_ID": "OCT14C-ARAMCO-CORE", "Company": "Aramco", "Title": "Understanding the subsurface: inside Aramco’s Core Laboratories Center", "URL": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "Source_date": "2026-08-25", "Source_type": "Company magazine article", "Publisher": "Aramco", "Locator": "25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established."}}, {"Source_ID": "OCT14C-ECOPETROL-CADE", "Company": "Ecopetrol", "Title": "Ecopetrol Integrated Management Report 2025", "URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Source_date": "2026 (2025 annual report)", "Source_type": "Operator annual report", "Publisher": "Ecopetrol", "Locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given."}}, {"Source_ID": "OCT14C-PETRONAS-TRICIPTA", "Company": "PETRONAS", "Title": "PETRONAS advances TriCipta AI through the third joint development agreement", "URL": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "Source_date": "2026-07-08", "Source_type": "Company release", "Publisher": "PETRONAS", "Locator": "8 July 2026 release: third JDA scope and final two paragraphs on earlier tools", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed."}}, {"Source_ID": "OCT14C-CHORD-LIFT", "Company": "Chord Energy", "Title": "Chord fourth-quarter and full-year 2025 results", "URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "Source_date": "2026-02-25", "Source_type": "Operator earnings release", "Publisher": "Chord Energy", "Locator": "Operations Update / Production and LOE; two consecutive artificial-lift paragraphs", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted."}}, {"Source_ID": "WATCH14PM-INPEX", "Company": "INPEX", "Title": "INPEX transforms business processes with Microsoft 365 Copilot and AI agents", "URL": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "Source_date": "2026-09-01", "Source_type": "Vendor customer story", "Publisher": "Microsoft / named INPEX speakers", "Locator": "Research tool; Building a foundation for change; Looking beyond estimated annual value", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method."}}, {"Source_ID": "WATCH14PM-WILLIAMS", "Company": "Williams", "Title": "Williams transforms frontline work with Copilot, gaining 90% adoption", "URL": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "Source_date": "2026-09-09", "Source_type": "Vendor customer story", "Publisher": "Microsoft / named Williams speakers", "Locator": "Bringing AI to the frontline first; Accelerating operations with Copilot", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share."}}, {"Source_ID": "WATCH14PM-NOV", "Company": "NOV", "Title": "NOV’s Drilling Beliefs & Analytics advances digital operations in Egypt", "URL": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "Source_date": "2026-03-11", "Source_type": "Vendor deployment article", "Publisher": "NOV", "Locator": "11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example", "Accessed": "2026-10-14", "Verification": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Notes": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.", "Current_review": {"last_checked": "2026-10-14", "source_read": "Relevant original source reviewed", "current_access": "Original text retrieved", "independent_validation": "Not established", "note": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established."}}, {"Source_ID": "WATCH15AM-CNOOC", "Company": "CNOOC", "Title": "2025 Environmental, Social and Governance Report", "URL": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "Source_date": "2026-04-09", "Source_type": "Company publication", "Publisher": "CNOOC Limited", "Locator": "Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026", "Accessed": "2026-10-15", "Verification": "Original PDF and relevant page reviewed by editor 15 October 2026; claims remain company-reported.", "Notes": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Relevant original issuer PDF passage read; CNOOC and Petrobras pages also rendered", "current_access": "Original PDF retrieved", "independent_validation": "Not established", "note": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved."}}, {"Source_ID": "WATCH15AM-SUNCOR", "Company": "Suncor", "Title": "Second Quarter 2026 Financial Results Call", "URL": "https://www.suncor.com/-/media/project/suncor/files/investor-centre/transcripts/2026-q2-suncor-energy-financial-results-call-transcript-en.pdf?created=20260818150055&modified=20260828195420", "Source_date": "2026-08-05", "Source_type": "Official earnings-call transcript", "Publisher": "Suncor Energy", "Locator": "PDF page 10 (page index 9), Q&A lines 331–358, Dave Oldreive and Rich Kruger; call dated 5 August 2026", "Accessed": "2026-10-15", "Verification": "Original PDF and relevant page reviewed by editor 15 October 2026; claims remain company-reported.", "Notes": "No architecture beyond linear programming, counterfactual, model-specific benefit, implementation cost or detailed control boundary. Keep it outside new AI-adoption and result counts. Suncor haulage and truck dispatch remain separate existing records.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Relevant original issuer PDF passage read; CNOOC and Petrobras pages also rendered", "current_access": "Original PDF retrieved", "independent_validation": "Not established", "note": "No architecture beyond linear programming, counterfactual, model-specific benefit, implementation cost or detailed control boundary. Keep it outside new AI-adoption and result counts. Suncor haulage and truck dispatch remain separate existing records."}}, {"Source_ID": "WATCH15AM-PETROBRAS", "Company": "Petrobras", "Title": "2026–2030 Business Plan", "URL": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "Source_date": "2025-11-27", "Source_type": "Investor presentation", "Publisher": "Petrobras", "Locator": "Printed slide 158 / PDF page 158, purchase-request automation box; separate from the AI and digital-twin boxes", "Accessed": "2026-10-15", "Verification": "Original PDF and relevant page reviewed by editor 15 October 2026; claims remain company-reported.", "Notes": "Company-reported values without a disclosed baseline, cost method, implementation cost or independent validation. The slide does not establish whether the estimate is realized or forecast. Do not count this as AI adoption or add it to the separate AI benefits.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Relevant original issuer PDF passage read; CNOOC and Petrobras pages also rendered", "current_access": "Original PDF retrieved", "independent_validation": "Not established", "note": "Company-reported values without a disclosed baseline, cost method, implementation cost or independent validation. The slide does not establish whether the estimate is realized or forecast. Do not count this as AI adoption or add it to the separate AI benefits."}}, {"Source_ID": "WATCH15X-GEORGE", "Company": "Chevron", "Title": "Maximizing Every Barrel: Artificial Lift Embraces Automation", "URL": "https://jpt.spe.org/maximizing-every-barrel-artificial-lift-embraces-automation", "Source_date": "2026-09-04", "Source_type": "Original conference reporting of named operator remarks", "Publisher": "SPE / Journal of Petroleum Technology", "Locator": "Chevron section: Peter Nesom remarks following the image caption; 25 August 2026 plenary, reported 4 September", "Short_excerpt": "used for over 5,000 wells", "Accessed": "2026-10-15", "Verification": "Original SPE/JPT report read by editor 15 October 2026; named operator remarks reported by the publisher, not issuer technical validation.", "Notes": "Original conference reporting of named operator remarks, not an operator-authored deployment paper. The event program corroborates speaker and session only. Keep as reported use awaiting technical corroboration, outside additional adoption/result counts. Do not infer an LLM architecture, autonomous field control or a company-wide knowledge system.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Original SPE/JPT report read by editor 15 October 2026; named operator remarks reported by the publisher, not issuer technical validation.", "current_access": "Publisher text or indexed abstract available; full conference paper unavailable", "independent_validation": "Not established", "note": "Original conference reporting of named operator remarks, not an operator-authored deployment paper. The event program corroborates speaker and session only. Keep as reported use awaiting technical corroboration, outside additional adoption/result counts. Do not infer an LLM architecture, autonomous field control or a company-wide knowledge system."}}, {"Source_ID": "WATCH15X-PETROBRAS", "Company": "Petrobras", "Title": "Virtual Assistant Using Generative Pre-Trained Transformers on Reservoir Management Context", "URL": "https://onepetro.org/OTCBRASIL/proceedings-abstract/25OTCB/25OTCB/D011S005R007/792383", "Source_date": "2025-10-21", "Source_type": "Operator-coauthored conference abstract", "Publisher": "Petrobras and Tecgraf / PUC-Rio authors; Offshore Technology Conference", "Locator": "OTC-36185-MS; publisher-indexed author list and abstract, test-scenario paragraphs", "Short_excerpt": "this work proposes an intelligent virtual assistant specialized in Reservoir Management systems.", "Accessed": "2026-10-15", "Verification": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved.", "Notes": "Publisher-indexed original abstract reviewed; full paper not retrieved. Retained as research, outside operating adoption and outcome counts. No evidence of permission to change reservoir plans or execute downstream actions.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved.", "current_access": "Publisher text or indexed abstract available; full conference paper unavailable", "independent_validation": "Not established", "note": "Publisher-indexed original abstract reviewed; full paper not retrieved. Retained as research, outside operating adoption and outcome counts. No evidence of permission to change reservoir plans or execute downstream actions."}}, {"Source_ID": "WATCH15X-EXXON-ESP", "Company": "ExxonMobil", "Title": "Closed-Loop ESP Optimization to Maximize Asset Value", "URL": "https://onepetro.org/SPEESP/proceedings-abstract/25ESP/25ESP/D031S006R001/788064", "Source_date": "2025-08-18", "Source_type": "Operator-authored conference abstract", "Publisher": "ExxonMobil authors / Society of Petroleum Engineers", "Locator": "SPE-225261-MS; abstract, system checks and 26-well pilot paragraphs", "Short_excerpt": "The closed-loop optimization system was piloted on 26 ESP wells in our Midland assets.", "Accessed": "2026-10-15", "Verification": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved.", "Notes": "Original publisher-indexed abstract reviewed; full paper not retrieved. It establishes physics-based control and a validation gate, not an LLM controlling pumps. This is separate from AI-235’s gas-lift ML and its oil-uplift figures.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved.", "current_access": "Publisher text or indexed abstract available; full conference paper unavailable", "independent_validation": "Not established", "note": "Original publisher-indexed abstract reviewed; full paper not retrieved. It establishes physics-based control and a validation gate, not an LLM controlling pumps. This is separate from AI-235’s gas-lift ML and its oil-uplift figures."}}, {"Source_ID": "WATCH15X-AETHON", "Company": "Aethon Energy", "Title": "Dallas CIO ORBIE special advertising supplement", "URL": "https://media.bizj.us/view/img/13143357/low-res-dallasorbiesciospecial-section2026-8951090d-b07f-4f22-bfba-90c5a075a999.pdf", "Source_date": "2026-02-20 to 2026-02-26", "Source_type": "Named executive profile in advertising supplement", "Publisher": "Dallas Business Journal / DallasCIO advertising supplement", "Locator": "PDF page 10 / printed page 10, Kang Chen profile; supplement dated 20–26 February 2026", "Short_excerpt": "fully migrating all IT systems to the cloud.", "Accessed": "2026-10-15", "Verification": "Original advertising-supplement PDF text read by editor 15 October 2026; named executive self-report, not independent validation.", "Notes": "Named executive self-report in an advertising supplement, not independent technical validation. Cloud migration and upskilling do not prove AI readiness or deployment. Keep as company transformation context, outside AI adoption and outcome counts.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Original advertising-supplement PDF text read by editor 15 October 2026; named executive self-report, not independent validation.", "current_access": "Publisher text or indexed abstract available; full conference paper unavailable", "independent_validation": "Not established", "note": "Named executive self-report in an advertising supplement, not independent technical validation. Cloud migration and upskilling do not prove AI readiness or deployment. Keep as company transformation context, outside AI adoption and outcome counts."}}, {"Source_ID": "WATCH15PM-REPSOL", "Company": "Repsol", "Title": "2025 Group Annual Financial Report", "URL": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/ref/2026/ref19022026-2025-group-annual-financial-report.pdf", "Source_date": "2026-02-19", "Source_type": "Original issuer annual report", "Publisher": "Repsol", "Locator": "PDF page 195 / printed Management Report page 45, section 5.5: AI-based autonomous optimizers", "Short_excerpt": "end of the initial phase of the first Autonomous Optimizer with Control (OAC)", "Accessed": "2026-10-15", "Verification": "Editor retrieved the original PDF, read the relevant text and inspected the rendered page on 15 October 2026.", "Notes": "Completing an initial phase does not establish routine autonomous operation or a plant-wide rollout. Keep separate from Repsol’s multiagent, workforce and legal programs.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Editor retrieved the original PDF, read the relevant text and inspected the rendered page on 15 October 2026.", "current_access": "Editor retrieved the original PDF, read the relevant text and inspected the rendered page on 15 October 2026.", "independent_validation": "Not established", "note": "Completing an initial phase does not establish routine autonomous operation or a plant-wide rollout. Keep separate from Repsol’s multiagent, workforce and legal programs."}}, {"Source_ID": "WATCH15PM-CORVA-OLIA", "Company": "Corva / Oliasoft", "Title": "Corva and Oliasoft Create a Continuous Learning Loop Between Well Design and Drilling", "URL": "https://www.corva.ai/blog/corva-and-oliasoft-create-a-continuous-learning-loop-between-well-design-and-drilling", "Source_date": "2026-10-12", "Source_type": "Joint vendor announcement", "Publisher": "Corva and Oliasoft", "Locator": "Joint release: engineering-to-operations, operations-to-engineering, and customer-pilot paragraphs", "Short_excerpt": "preparing customer pilots", "Accessed": "2026-10-15", "Verification": "Editor read both vendors’ original releases on 15 October 2026; neither supplies customer-pilot outcomes.", "Notes": "Model feedback is not evidence of automatic rig control. Two vendor versions of the same release do not provide independent validation; action permissions are unspecified.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Editor read both vendors’ original releases on 15 October 2026; neither supplies customer-pilot outcomes.", "current_access": "Editor read both vendors’ original releases on 15 October 2026; neither supplies customer-pilot outcomes.", "independent_validation": "Not established", "note": "Model feedback is not evidence of automatic rig control. Two vendor versions of the same release do not provide independent validation; action permissions are unspecified."}}, {"Source_ID": "WATCH15PM-FORECAST", "Company": "UNICAMP / UNISIM", "Title": "Data-driven production forecast for oil and gas reservoirs - Advantages and limitations", "URL": "https://www.sciencedirect.com/science/article/pii/S2949891026002137", "Source_date": "2026-10 (issue month; exact online date unknown)", "Source_type": "Technical journal article", "Publisher": "Geoenergy Science and Engineering / Elsevier", "Locator": "Volume 265, article 214569; DOI 10.1016/j.geoen.2026.214569; publisher abstract and indexed sections, no PDF page locator", "Short_excerpt": "History-only inputs tend to fail to capture operational regime changes", "Accessed": "2026-10-15", "Verification": "Editor independently read publisher-indexed abstract and section text; direct retrieval failed, and no full paper or numerical tables were obtained.", "Notes": "Publisher abstract and indexed section text only; full paper unavailable. This is a synthetic benchmark, not a named failed field deployment or an economic result.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Editor independently read publisher-indexed abstract and section text; direct retrieval failed, and no full paper or numerical tables were obtained.", "current_access": "Editor independently read publisher-indexed abstract and section text; direct retrieval failed, and no full paper or numerical tables were obtained.", "independent_validation": "Not established", "note": "Publisher abstract and indexed section text only; full paper unavailable. This is a synthetic benchmark, not a named failed field deployment or an economic result."}}, {"Source_ID": "WATCH15PM-OFGEM", "Company": "Ofgem AI Reg Lab", "Title": "AI Reg Lab: High-level findings for licensees and stakeholders — February 2026", "URL": "https://www.ofgem.gov.uk/sites/default/files/2026-04/AI-Reg-Lab-Februray-2026.pdf", "Source_date": "2026-04-29", "Source_type": "Official regulator lab summary", "Publisher": "Ofgem", "Locator": "PDF page 1: publication date; page 2: pilot and lab scope; page 3: themes; page 4: disclaimer", "Short_excerpt": "clear identification and distinguishability between authoritative data and AI-generated inferences", "Accessed": "2026-10-15", "Verification": "Editor read the original four-page PDF on 15 October 2026. © Crown copyright 2026; document reused under the Open Government Licence.", "Notes": "An anonymized discussion, not a system audit, observed failure, regulatory endorsement or compliance advice. Energy-infrastructure context; no petroleum operator is identified. Findings reflect participants’ views, not Ofgem policy.", "Current_review": {"last_checked": "2026-10-15", "source_read": "Editor read the original four-page PDF on 15 October 2026. © Crown copyright 2026; document reused under the Open Government Licence.", "current_access": "Editor read the original four-page PDF on 15 October 2026. © Crown copyright 2026; document reused under the Open Government Licence.", "independent_validation": "Not established", "note": "An anonymized discussion, not a system audit, observed failure, regulatory endorsement or compliance advice. Energy-infrastructure context; no petroleum operator is identified. Findings reflect participants’ views, not Ofgem policy."}}, {"Source_ID": "WATCH16AM-TOTAL-DRUM", "Company": "TotalEnergies", "Title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "URL": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "Source_date": "2026-06-11", "Source_type": "Original conference reporting of named operator remarks", "Publisher": "Control Global / Honeywell Users Group coverage", "Locator": "Drum cycling prediction section; Thomas Klein remarks at Honeywell Users Group Americas 2026", "Short_excerpt": "so this use case wasn’t retained", "Accessed": "2026-10-16", "Verification": "Editor read the original conference report and both partners’ earlier releases. The decision to drop this use case rests on reported named remarks, not a company-authored outcome report.", "Notes": "Original conference reporting, labeled Sponsored. We could not retrieve the presentation or a later company statement. The earlier company release confirms the broader pilot, not the decision to drop this use case. Keep these figures separate from its five-event result.", "Current_review": {"last_checked": "2026-10-16", "source_read": "Editor read the original conference report and both partners’ earlier releases. The decision to drop this use case rests on reported named remarks, not a company-authored outcome report.", "current_access": "Original publication read", "independent_validation": "Not established", "note": "Original conference reporting, labeled Sponsored. We could not retrieve the presentation or a later company statement. The earlier company release confirms the broader pilot, not the decision to drop this use case. Keep these figures separate from its five-event result."}}, {"Source_ID": "WATCH16AM-HP", "Company": "Helmerich & Payne", "Title": "H&P & ExxonMobil Expand Deployment of FlexRobotics Technology", "URL": "https://ir.hpinc.com/news/press-releases/news-details/2026/HP--ExxonMobil-Expand-Deployment-of-FlexRobotics-Technology/default.aspx", "Source_date": "2026-10-07", "Source_type": "Original company announcement", "Publisher": "Helmerich & Payne", "Locator": "Opening release paragraphs: first two systems, seven additional units over next 12 months; ExxonMobil quotation", "Short_excerpt": "expects to deploy seven additional FlexRobotics systems", "Accessed": "2026-10-16", "Verification": "Original H&P release read by editor; customer quotation corroborates intent, not quantified results.", "Notes": "An ExxonMobil executive supports expansion intent; the joint account is not independent outcome validation. It does not establish installation of the separate AI camera add-on in AI-116. Keep as robotics context, outside AI-adoption counts.", "Current_review": {"last_checked": "2026-10-16", "source_read": "Original H&P release read by editor; customer quotation corroborates intent, not quantified results.", "current_access": "Original publication read", "independent_validation": "Not established", "note": "An ExxonMobil executive supports expansion intent; the joint account is not independent outcome validation. It does not establish installation of the separate AI camera add-on in AI-116. Keep as robotics context, outside AI-adoption counts."}}, {"Source_ID": "WATCH16AM-RETINA", "Company": "SLB", "Title": "Retina At-Bit Imaging System Surpasses 50-Well Milestone in First Year of Deployment", "URL": "https://www.slb.com/newsroom/features/2026/2026-1015-retina", "Source_date": "2026-10-15", "Source_type": "Original company announcement", "Publisher": "SLB", "Locator": "Published date; Global adoption and Global deployment sections", "Short_excerpt": "more than 50 wells across 18 locations worldwide", "Accessed": "2026-10-16", "Verification": "Editor read the original SLB feature; performance and scope remain vendor-reported.", "Notes": "No AI/ML mechanism or automatic control is identified. Keep as instrumentation context; do not transfer OQ Oman’s separate drilling-tool results to Retina.", "Current_review": {"last_checked": "2026-10-16", "source_read": "Editor read the original SLB feature; performance and scope remain vendor-reported.", "current_access": "Original publication read", "independent_validation": "Not established", "note": "No AI/ML mechanism or automatic control is identified. Keep as instrumentation context; do not transfer OQ Oman’s separate drilling-tool results to Retina."}}, {"Source_ID": "WATCH16PM-PTTEP", "Company": "PTTEP", "Title": "Minutes of the 2026 Annual General Shareholders’ Meeting", "URL": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-minutes-en.pdf", "Source_date": "2026-04-10", "Source_type": "Original company meeting minutes", "Publisher": "PTTEP", "Locator": "PDF page 13 / printed page 12 of 25, response to question 1.7; cover letter on PDF page 1 confirms publication date", "Short_excerpt": "continuously deployed AI and automation technologies across its core business processes", "Accessed": "2026-10-16", "Verification": "Editor read the official translated minutes and inspected the rendered passage. Company self-report, not independent outcome validation.", "Notes": "No fields, vendors, well count, model performance or approval/control boundaries are disclosed. The separate USD 328m new-business budget is not AI-only. This broad account does not establish the scope or success of the later offshore advisory pilot in AI-067.", "Current_review": {"last_checked": "2026-10-16", "source_read": "Editor read the official translated minutes and inspected the rendered passage. Company self-report, not independent outcome validation.", "current_access": "Original PDF and source page available", "independent_validation": "Not established", "note": "No fields, vendors, well count, model performance or approval/control boundaries are disclosed. The separate USD 328m new-business budget is not AI-only. This broad account does not establish the scope or success of the later offshore advisory pilot in AI-067."}}, {"Source_ID": "WATCH16PM-IEA", "Company": "International Energy Agency", "Title": "Key Questions on Energy and AI", "URL": "https://www.iea.org/reports/key-questions-on-energy-and-ai", "Source_date": "2026-04-16", "Source_type": "Original industry research report", "Publisher": "International Energy Agency", "Locator": "Official report page: Published 16 April 2026; PDF page 10: demand outlook; PDF/printed page 84: Figure 9.4; pages 80–85: adoption barriers", "Short_excerpt": "the lack of digital skills is the single largest barrier", "Accessed": "2026-10-16", "Verification": "Editor read the official page and relevant full-PDF passages, including the rendered Figure 9.4. IEA report text is CC BY 4.0; Annex A has a separate noncommercial share-alike licence.", "Notes": "Broad energy-sector evidence; the reviewed chapter supplies no survey respondent count or upstream-only subgroup. Projections are not realized benefits or ROI. This is a distinct follow-up to the 2025 report INS001, not a replacement.", "Current_review": {"last_checked": "2026-10-16", "source_read": "Editor read the official page and relevant full-PDF passages, including the rendered Figure 9.4. IEA report text is CC BY 4.0; Annex A has a separate noncommercial share-alike licence.", "current_access": "Original PDF and source page available", "independent_validation": "Not established", "note": "Broad energy-sector evidence; the reviewed chapter supplies no survey respondent count or upstream-only subgroup. Projections are not realized benefits or ROI. This is a distinct follow-up to the 2025 report INS001, not a replacement."}}, {"Source_ID": "WATCH17AM-CNRL", "Company": "Canadian Natural Resources", "Title": "Investor Open House · 7 November 2025", "URL": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "Source_date": "2025-11-07", "Source_type": "Operator investor presentation", "Publisher": "Canadian Natural Resources", "Locator": "PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention", "Short_excerpt": "Optimization of hydrogen through use of AI — Saving ~$8 million per year.", "Accessed": "2026-10-17", "Verification": "Original source read by editor 17 October 2026; claims remain issuer-reported.", "Notes": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.", "Current_review": {"last_checked": "2026-10-17", "source_read": "Original source read", "current_access": "Original PDF or SEC/issuer release available", "independent_validation": "Not established", "note": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate."}}, {"Source_ID": "WATCH17AM-KANTORI", "Company": "Baker Hughes", "Title": "Kantori drilling advisory case · UAE", "URL": "https://dam.bakerhughes.com/m/6c9e1d4e04184778/original/i-Trak-service-maximized-ROP-saved-22hrs-UAE-cs.pdf#page=1", "Source_date": "2026-04", "Source_type": "Vendor case sheet", "Publisher": "Baker Hughes", "Locator": "Single-page case sheet; Solution, Results and footer revision 04/2026", "Short_excerpt": "Deployed Kantori autonomous directional drilling in advisory mode", "Accessed": "2026-10-17", "Verification": "Original source read by editor 17 October 2026; claims remain issuer-reported.", "Notes": "The customer and well are unnamed. A prior-well comparison does not isolate AI from the wider automation service, crew or geology. Costs, accepted recommendations and repeatability are undisclosed; no ROI is established. The deployment date is unknown; the PDF revision does not establish a 2026 deployment.", "Current_review": {"last_checked": "2026-10-17", "source_read": "Original source read", "current_access": "Original PDF or SEC/issuer release available", "independent_validation": "Not established", "note": "The customer and well are unnamed. A prior-well comparison does not isolate AI from the wider automation service, crew or geology. Costs, accepted recommendations and repeatability are undisclosed; no ROI is established. The deployment date is unknown; the PDF revision does not establish a 2026 deployment."}}, {"Source_ID": "WATCH17AM-PDO", "Company": "Petroleum Development Oman", "Title": "Kongsberg Digital and PDO digital-twin agreement", "URL": "https://kommunikasjon.ntb.no/pressemelding/18645626/kongsberg-digital-and-petroleum-development-oman-sign-agreement-for-deployment-of-ai-powered-digital-twin?lang=en&publisherId=17849150", "Source_date": "2025-09-08", "Source_type": "Issuer press release", "Publisher": "KONGSBERG / NTB", "Locator": "Release timestamp 8 September 2025; scope and deployment-ambition paragraphs", "Short_excerpt": "Kognitwin unifies and contextualises operational (OT) and enterprise (IT) data", "Accessed": "2026-10-17", "Verification": "Original source read by editor 17 October 2026; claims remain issuer-reported.", "Notes": "No fields, facilities, users or workflow counts are identified. The company-description paragraph about gas operations does not define deployment scope. A passed target date alone does not establish delay or failure.", "Current_review": {"last_checked": "2026-10-17", "source_read": "Original source read", "current_access": "Original PDF or SEC/issuer release available", "independent_validation": "Not established", "note": "No fields, facilities, users or workflow counts are identified. The company-description paragraph about gas operations does not define deployment scope. A passed target date alone does not establish delay or failure."}}, {"Source_ID": "WATCH17PM-INPEX", "Company": "INPEX", "Title": "INPEX commences operations of SAP Cloud ERP", "URL": "https://www.inpex.com/english/news/upload/20260624.pdf#page=2", "Source_date": "2026-06-24", "Source_type": "Joint company announcement", "Publisher": "INPEX / SAP Japan / ABeam Consulting", "Locator": "PDF pages 1–2; Japan go-live and Fit-to-Standard approach on page 2", "Short_excerpt": "full-scale operation to commence in Japan in January 2026", "Accessed": "2026-10-17", "Verification": "Editor read the official three-page PDF and inspected page 2. ERP use is established; AI operation and Copilot integration are not.", "Notes": "AI appears as compatibility and a future efficiency objective, not a deployed function. Users, departments, costs and approval rules are unspecified. No connection to the separate Copilot/management-meeting agents in AI-294 is established.", "Current_review": {"last_checked": "2026-10-17", "source_read": "Editor read the official three-page PDF and inspected page 2. ERP use is established; AI operation and Copilot integration are not.", "current_access": "Original PDF available", "independent_validation": "Not established", "note": "AI appears as compatibility and a future efficiency objective, not a deployed function. Users, departments, costs and approval rules are unspecified. No connection to the separate Copilot/management-meeting agents in AI-294 is established."}}, {"Source_ID": "WATCH17PM-FDP", "Company": "OriGen.AI / SPE", "Title": "Field Development Plan Optimization Under Uncertainty Using Physics-Informed Neural Networks and Generative AI: A Real-Field Case Study in North America", "URL": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/792786", "Source_date": "2025-11-03", "Source_type": "Supplier-authored conference modeling study", "Publisher": "SPE / OnePetro; authors affiliated with OriGen.AI", "Locator": "SPE-229333-MS, abstract; publisher metadata: Published November 03 2025; conference 3–6 November 2025", "Short_excerpt": "expected net present value", "Accessed": "2026-10-17", "Verification": "Publisher-indexed metadata and complete abstract inspected by editor. Direct OnePetro page failed; full paper, figures and validation protocol not reviewed.", "Notes": "Editor checked the publisher-indexed abstract; direct page access failed and the full paper was unavailable. Error definition, test split, economic baseline, costs, human review and operator identity remain unverified. One field does not establish repeatability.", "Current_review": {"last_checked": "2026-10-17", "source_read": "Publisher-indexed metadata and complete abstract inspected by editor. Direct OnePetro page failed; full paper, figures and validation protocol not reviewed.", "current_access": "Publisher-indexed abstract available; full paper unavailable", "independent_validation": "Not established", "note": "Editor checked the publisher-indexed abstract; direct page access failed and the full paper was unavailable. Error definition, test split, economic baseline, costs, human review and operator identity remain unverified. One field does not establish repeatability."}}, {"Source_ID": "WATCH18AM-KOC-DIGEST", "Company": "Kuwait Oil Company", "Title": "SANAD AI: The Generative AI Drilling Assistant", "URL": "https://www.kockw.com/sites/EN/Periodicals/Digest%20April%20-%20June%202026.pdf#page=28", "Source_date": "2026-Q2", "Source_type": "Operator implementation account and attributed engineer interview", "Publisher": "Kuwait Oil Company", "Locator": "PDF pages 28–31 / printed 26–29; 14 SOPs on PDF 30; validation, ownership and roadmap on PDF 31", "Short_excerpt": "Every output was validated", "Accessed": "2026-10-18", "Verification": "Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.", "Notes": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.", "current_access": "Original source read", "independent_validation": "Not established", "note": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085."}}, {"Source_ID": "WATCH18AM-KOC-INTERVIEW", "Company": "Kuwait Oil Company", "Title": "KOC’s Sanad uses AI to unlock drilling knowledge", "URL": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "Source_date": "2026-09-29", "Source_type": "Operator implementation account and attributed engineer interview", "Publisher": "KUNA interview, republished by Kuwait Times", "Locator": "Timestamp; lower-Burgan field test, SOP-related savings, RAG description and human responsibility", "Short_excerpt": "Engineering judgment and responsibility remain with the specialist", "Accessed": "2026-10-18", "Verification": "Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.", "Notes": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.", "current_access": "Original source read", "independent_validation": "Not established", "note": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085."}}, {"Source_ID": "WATCH18AM-SPGLOBAL", "Company": "S&P Global Energy", "Title": "From Data to Dialogue: S&P Global Energy’s Genie and MCP architecture", "URL": "https://www.databricks.com/blog/data-dialogue-how-sp-global-energy-made-its-structured-data-estate-conversational-databricks", "Source_date": "2026-09-25", "Source_type": "Vendor-published technical customer account", "Publisher": "Databricks / S&P Global Energy", "Locator": "Date; per-dataset curation; Unity Catalog permissions; FastMCP composition; customer quote and benchmark loop", "Short_excerpt": "now takes days", "Accessed": "2026-10-18", "Verification": "Editor read the original vendor customer account. Architecture and benefits remain vendor/customer-reported.", "Notes": "This is a vendor-published customer account, not an independent quality or security assessment. Benchmarks are described without scores. Access to research data does not establish authority to trade, approve investment or control field equipment.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Editor read the original vendor customer account. Architecture and benefits remain vendor/customer-reported.", "current_access": "Original source read", "independent_validation": "Not established", "note": "This is a vendor-published customer account, not an independent quality or security assessment. Benchmarks are described without scores. Access to research data does not establish authority to trade, approve investment or control field equipment."}}, {"Source_ID": "WATCH18AM-PETROREC", "Company": "PetroReconcavo", "Title": "Relatório de Sustentabilidade 2024", "URL": "https://www.petroreconcavo.com.br/Download.aspx?Arquivo=OXSl%2FB7SEmrx%2BHoNzBeFlQ%3D%3D#page=78", "Source_date": "2025-07-14", "Source_type": "Operator sustainability report and publication notice", "Publisher": "PetroReconcavo", "Locator": "PDF / printed page 78: SAP, Projeto Waves and data lake", "Short_excerpt": "A nova rede possibilita monitoramento em tempo real", "Accessed": "2026-10-18", "Verification": "Editor retrieved the original 100-page Portuguese PDF after encoding the plus sign in its download parameter; page 78 read and rendered.", "Notes": "No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Editor retrieved the original 100-page Portuguese PDF after encoding the plus sign in its download parameter; page 78 read and rendered.", "current_access": "Original source read", "independent_validation": "Not established", "note": "No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology."}}, {"Source_ID": "WATCH18AM-PETROREC-DATE", "Company": "PetroReconcavo", "Title": "2024 Sustainability Report · Notice to the Market", "URL": "https://ri.petroreconcavo.com.br/Download.aspx?Arquivo=6hzxLJkhGOrPzcEDtUKx3A%3D%3D&IdCanal=ckeQVNzVtM%2FZZ4T5tqRLLw%3D%3D", "Source_date": "2025-07-14", "Source_type": "Operator sustainability report and publication notice", "Publisher": "PetroReconcavo", "Locator": "PDF pages 1–2; publication announcement and English-edition timing", "Short_excerpt": "published, on this date, its 2024 Sustainability Report", "Accessed": "2026-10-18", "Verification": "Original issuer notice read; Portuguese publication 14 July, English edition separately planned for 25 July 2025.", "Notes": "No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original issuer notice read; Portuguese publication 14 July, English edition separately planned for 25 July 2025.", "current_access": "Original source read", "independent_validation": "Not established", "note": "No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology."}}, {"Source_ID": "WATCH18PM-OMV", "Company": "OMV", "Title": "How AI is powering better decisions in the race to net zero", "URL": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "Source_type": "Operator technical account", "Publisher": "OMV", "Locator": "Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker", "Short_excerpt": "not generative AI, but analytical, agent-based AI", "Accessed": "2026-10-18", "Verification": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "Notes": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132).", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original text read; targeted workflow/metric check", "independent_validation": "Not established", "note": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132)."}, "Source_date": "2026-04-14"}, {"Source_ID": "WATCH18PM-NOV", "Company": "NOV", "Title": "NOV Reports Second Quarter 2026 Earnings", "URL": "https://investors.nov.com/news-releases/news-release-details/nov-reports-second-quarter-2026-earnings", "Source_type": "Supplier earnings release", "Publisher": "NOV", "Locator": "Significant Achievements: Norwegian sulphate-removal unit; separate remote service-rig monitoring paragraph", "Short_excerpt": "recommend maintenance actions using the Max Platform", "Accessed": "2026-10-18", "Verification": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "Notes": "Supplier-only account; operator, unit and user population unnamed. Approval, override and safety rules undisclosed. Separate service-rig monitoring in the same release is not evidence of AI.", "Current_review": {"last_checked": "2026-10-18", "source_read": "Original text read; targeted workflow/metric check", "independent_validation": "Not established", "note": "Supplier-only account; operator, unit and user population unnamed. Approval, override and safety rules undisclosed. Separate service-rig monitoring in the same release is not evidence of AI."}, "Source_date": "2026-07-28"}, {"Source_ID": "WATCH19AM-HAVTIL", "Company": "DNV / Havtil", "Title": "Can artificial intelligence help improve learning from incidents?", "URL": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2026/can-artificial-intelligence-help-improve-learning-from-incidents/", "Source_date": "2026-01-07", "Source_type": "Regulator workshop account and technical slides", "Publisher": "Norwegian Ocean Industry Authority (Havtil)", "Locator": "Workshop date/attendance; How AI can help; Quality and trust; Obstacles; The way forward", "Short_excerpt": "No specific further activities were identified", "Accessed": "2026-10-19", "Verification": "Original regulator account read by editor 19 October 2026. Workshop experience is not a representative survey or a tool benchmark.", "Notes": "Forty representatives are not forty companies or a survey sample. No named failure or agreed follow-up project is identified. Listed framework user-test participants do not establish AI deployments. This is distinct from the 2024 responsible-AI overview (INS004).", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original regulator account read by editor 19 October 2026. Workshop experience is not a representative survey or a tool benchmark.", "independent_validation": "Not established", "note": "Forty representatives are not forty companies or a survey sample. No named failure or agreed follow-up project is identified. Listed framework user-test participants do not establish AI deployments. This is distinct from the 2024 responsible-AI overview (INS004)."}}, {"Source_ID": "WATCH19AM-HAVTIL-SLIDES", "Company": "DNV / Havtil", "Title": "Bruk av kunstig intelligens til læring etter hendelser", "URL": "https://www.havtil.no/contentassets/2152e5ff08734a219a61315d4a50464c/havtil-arbeidsmote-20251028-ki-for-laring-av-hendelser.pdf#page=15", "Source_date": "2025-10-28 workshop; posting date unknown", "Source_type": "Regulator workshop account and technical slides", "Publisher": "Havtil / DNV", "Locator": "36-page deck: physical/printed p. 15 workflow; p. 20 contributing-factor framework and high-severity root-cause-analysis boundary", "Short_excerpt": "Prat med data", "Accessed": "2026-10-19", "Verification": "Original PDF text read; pages 15 and 20 rendered and inspected. Web screenshot failed; local PyMuPDF rendering succeeded. Event date is not a verified posting date.", "Notes": "Forty representatives are not forty companies or a survey sample. No named failure or agreed follow-up project is identified. Listed framework user-test participants do not establish AI deployments. This is distinct from the 2024 responsible-AI overview (INS004).", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original PDF text read; pages 15 and 20 rendered and inspected. Web screenshot failed; local PyMuPDF rendering succeeded. Event date is not a verified posting date.", "independent_validation": "Not established", "note": "Forty representatives are not forty companies or a survey sample. No named failure or agreed follow-up project is identified. Listed framework user-test participants do not establish AI deployments. This is distinct from the 2024 responsible-AI overview (INS004)."}}, {"Source_ID": "WATCH19AM-TOTAL-AWS", "Company": "TotalEnergies", "Title": "TotalEnergies puts Gen AI to work on its trading operations", "URL": "https://aws.amazon.com/solutions/case-studies/totalenergies/", "Source_date": "Undated", "Source_type": "Vendor-published customer account", "Publisher": "AWS / TotalEnergies Trading & Shipping", "Locator": "Overview; Olivier Collon customer quotation; AWS Services Used", "Short_excerpt": "matching structured and unstructured data within transaction flows", "Accessed": "2026-10-19", "Verification": "Original AWS customer account read by editor 19 October 2026. No visible publication date; a secondary 2024-10-28 clue could not be confirmed from an original dated source.", "Notes": "Vendor-published customer testimony, not independent validation. No costs or net ROI. Matching records does not establish authority to place trades, release payments or change books. No specific model provider can be inferred from the Bedrock service description.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original AWS customer account read by editor 19 October 2026. No visible publication date; a secondary 2024-10-28 clue could not be confirmed from an original dated source.", "independent_validation": "Not established", "note": "Vendor-published customer testimony, not independent validation. No costs or net ROI. Matching records does not establish authority to place trades, release payments or change books. No specific model provider can be inferred from the Bedrock service description."}}, {"Source_ID": "WATCH19EX-MURPHY", "Company": "Murphy Oil", "Title": "Framework for AI- and ML-Based Predictive Maintenance for Offshore Rotating Equipment", "URL": "https://jpt.spe.org/framework-for-ai-and-ml-based-predictive-maintenance-for-offshore-rotating-equipment", "Source_date": "2025-01-20", "Source_type": "Technical project account", "Publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "Locator": "Methodology; Observations; Discussion: Project Planning and Management, Data Prerequisites, Domain Knowledge, Model Development, Performance, and User Experience; Figs. 4–6", "Short_excerpt": "the first predictive model was deployed 6 months after the project kickoff", "Accessed": "2026-10-19", "Verification": "Original technical account read by editor 19 October 2026. Text and qualitative figure discussion reviewed; numeric chart values not transcribed. Not an independent outcome audit.", "Notes": "The reviewed 2023 alerts are older context; current operating status is unknown. The authors cite insufficient supplier equipment expertise and reported mitigation—not abandonment. Platform names, supplier and precision/recall are undisclosed; no autonomous maintenance established.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original technical account read by editor 19 October 2026. Text and qualitative figure discussion reviewed; numeric chart values not transcribed. Not an independent outcome audit.", "independent_validation": "Not established", "note": "The reviewed 2023 alerts are older context; current operating status is unknown. The authors cite insufficient supplier equipment expertise and reported mitigation—not abandonment. Platform names, supplier and precision/recall are undisclosed; no autonomous maintenance established."}}, {"Source_ID": "WATCH19EX-NOV-AURA", "Company": "NOV", "Title": "Aura moves remote crane operations from concept to reality", "URL": "https://www.nov.com/news/aura-moves-remote-crane-operations-from-concept-to-reality", "Source_date": "2026-08-20", "Source_type": "Supplier original announcement", "Publisher": "NOV", "Locator": "Remote-access paragraph; Kristiansand/Søgne test; first delivery; two planned deliveries; phased customer adoption", "Short_excerpt": "One fully remote-operated crane has already been delivered", "Accessed": "2026-10-19", "Verification": "Original supplier account read by editor 19 October 2026. Supplier demonstration and delivery are distinguished from customer operating use.", "Notes": "No AI/ML identified. Delivery does not establish shore-controlled customer operations or autonomous lifting. Customer, vessel, operational hours, interlocks and fallback arrangements are undisclosed.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original supplier account read by editor 19 October 2026. Supplier demonstration and delivery are distinguished from customer operating use.", "independent_validation": "Not established", "note": "No AI/ML identified. Delivery does not establish shore-controlled customer operations or autonomous lifting. Customer, vessel, operational hours, interlocks and fallback arrangements are undisclosed."}}, {"Source_ID": "WATCH19EX-OCEANEERING", "Company": "Oceaneering", "Title": "Oceaneering Secures Multi-Discipline Cross-Regional Offshore Survey Contracts for Ocean Intervention II", "URL": "https://www.oceaneering.com/oceaneering-secures-multi-discipline-cross-regional-offshore-survey-contracts-for-ocean-intervention-ii/", "Source_date": "2026-06-30", "Source_type": "Supplier original announcement", "Publisher": "Oceaneering", "Locator": "June 30 release: opening contract/program paragraphs; Trinidad/U.S. Gulf scope; simultaneous operations; supplier quotation; 2025 vessel upgrades", "Short_excerpt": "The program commenced in April 2026", "Accessed": "2026-10-19", "Verification": "Original supplier release read by editor 19 October 2026. Begun campaign is distinguished from planned scopes, durations and unverified completion.", "Notes": "Autonomous survey is not proof of ML or LLM use. The vessel still has a crew. Operator identity, human supervision, intervention rules, achieved campaign length and comparative performance are undisclosed.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original supplier release read by editor 19 October 2026. Begun campaign is distinguished from planned scopes, durations and unverified completion.", "independent_validation": "Not established", "note": "Autonomous survey is not proof of ML or LLM use. The vessel still has a crew. Operator identity, human supervision, intervention rules, achieved campaign length and comparative performance are undisclosed."}}, {"Source_ID": "WATCH19PM-METERING", "Company": "SPE / JPT", "Title": "From Measurement to Commercial Confidence: An Operator’s View of Custody Transfer—and a Responsible Role for AI", "URL": "https://jpt.spe.org/from-measurement-to-commercial-confidence-an-operators-view-of-custody-transfer-and-a-responsible-role-for-ai", "Source_date": "2026-10-12", "Source_type": "Technical-section editorial", "Publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "Locator": "Meter Accuracy Is Not System Credibility; Custody Transfer is an Organizational System; Automation and AI Can Assist, but Accountability Cannot Be Automated; author biographies", "Short_excerpt": "an unvalidated output cannot independently change a billed quantity", "Accessed": "2026-10-19", "Verification": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "Notes": "Author affiliations include operators and suppliers, but this is not an ExxonMobil or service-company deployment announcement. Shared inputs can make checks agree while retaining the same bias. No evidence that AI changed billed quantities.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "independent_validation": "Not established", "note": "Author affiliations include operators and suppliers, but this is not an ExxonMobil or service-company deployment announcement. Shared inputs can make checks agree while retaining the same bias. No evidence that AI changed billed quantities."}}, {"Source_ID": "WATCH19PM-STIMULATION", "Company": "Discover Artificial Intelligence", "Title": "A systematic and critical review of artificial intelligence applications for enhancing well stimulation operations and sustainable reservoir management", "URL": "https://link.springer.com/article/10.1007/s44163-026-02455-x", "Source_date": "2026-10-15", "Source_type": "Academic literature review", "Publisher": "Discover Artificial Intelligence / Springer Nature", "Locator": "Abstract; §§2.2–2.3 review selection/quality; §5.8 heterogeneity; §§6.1.1–6.1.5 validation, integration, readiness and economics; §7.4 denominators", "Short_excerpt": "Field-scale and cross-field validation are limited", "Accessed": "2026-10-19", "Verification": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "Notes": "Review conclusions depend on heterogeneous studies and incomplete reporting. We have not re-audited every underlying experiment. No named abandonment or industry failure rate; performance averages cannot rank companies or predict another field’s results.", "Current_review": {"last_checked": "2026-10-19", "source_read": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "independent_validation": "Not established", "note": "Review conclusions depend on heterogeneous studies and incomplete reporting. We have not re-audited every underlying experiment. No named abandonment or industry failure rate; performance averages cannot rank companies or predict another field’s results."}}, {"Source_ID": "WATCH20AM-OILINDIA", "Company": "Oil India", "Title": "Digitalization in OIL — DRIVE 2.0 programme", "URL": "https://www.oil-india.com/digitalfootprint", "Source_date": "Undated", "Source_type": "Issuer programme overview; undated", "Publisher": "Oil India Limited", "Locator": "Digitalization in OIL; DRIVE 2.0 §§1–10; §2 380-well scope; §4 approvals; §7 notifications; header and footer update strings", "Short_excerpt": "10 digital initiatives are being taken up under Project DRIVE 2.0", "Accessed": "2026-10-20", "Verification": "Original issuer page read by editor 20 October 2026. Public /digitalfootprint path accessible; earlier /index.php/digitalfootprint attempt was blocked. Header and footer update strings are not verified publication or launch dates.", "Notes": "One undated issuer plan, now accessible after an earlier blocked route. No completion dates, live user population or model providers established. Ten workstreams are not ten operating AI deployments. Publication and DRIVE 2.0 start dates unknown. Header: “11/09/2026 09:35 AM”; footer: “20/10/2026 11:28 AM”. These update strings do not date deployment. The 2019 start belongs to earlier DRIVE.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Original issuer page read by editor 20 October 2026. Public /digitalfootprint path accessible; earlier /index.php/digitalfootprint attempt was blocked. Header and footer update strings are not verified publication or launch dates.", "independent_validation": "Not established", "note": "One undated issuer plan, now accessible after an earlier blocked route. No completion dates, live user population or model providers established. Ten workstreams are not ten operating AI deployments."}}, {"Source_ID": "WATCH20UP-KODIAK", "Company": "Atlas Energy Solutions", "Title": "Kodiak Driver Sets New Uptime Record By Running 64 Consecutive Days on an Unsupervised Driverless Truck", "URL": "https://kodiak.ai/news/kodiak-driver-uptime-record", "Source_date": "2026-10-20", "Source_type": "Supplier announcement", "Publisher": "Kodiak AI", "Locator": "Opening uptime definition and dates; Atlas fleet as-of paragraph; future fleet and highway plans", "Short_excerpt": "running and available for service", "Accessed": "2026-10-20", "Verification": "Full original supplier release retrieved (HTTP 200) and read. Supplier-reported availability and fleet figures; no independent reliability, safety or financial audit.", "Notes": "Available does not mean continuously driving or hauling. No mileage, dispatch utilization, collision/exposure comparison or cost denominator disclosed. Kodiak AI is separate from Kodiak Gas Services (AI-193). Announced October 20, 2026. The 64-day streak ran June 30–September 1. The fleet and paid-hours figures are as of June 30; 100 trucks by mid-2027 and a year-end highway launch remain plans.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Full original supplier release retrieved (HTTP 200) and read. Supplier-reported availability and fleet figures; no independent reliability, safety or financial audit.", "independent_validation": "Not established", "note": "Available does not mean continuously driving or hauling. No mileage, dispatch utilization, collision/exposure comparison or cost denominator disclosed. Kodiak AI is separate from Kodiak Gas Services (AI-193)."}}, {"Source_ID": "WATCH20UP-SPE", "Company": "SPE / OnePetro", "Title": "SPE extends i2k Connect information-classification collaboration", "URL": "https://wrex.marketminute.com/article/bizwire-2026-10-20-society-of-petroleum-engineers-spe-partners-with-i2k-connect-for-information-classification", "Source_date": "2026-10-20", "Source_type": "Issuer announcement and original publisher account", "Publisher": "i2k Connect / Business Wire; syndicated by MarketMinute", "Locator": "Opening announcement; classification paragraph; SPE CEO statement", "Short_excerpt": "The classification is fully automated", "Accessed": "2026-10-20", "Verification": "Complete issuer-authored Business Wire release read on an accessible syndication page. Distributor original timed out; identical release also retrieved from AOL. No independent outcome validation. Distributor URL: https://www.businesswire.com/news/home/20261020770911/en/", "Notes": "The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model. Classification release: October 20, 2026. JPT article: September 10 publication header, with an undated following-conference licensing update. Neither dates completed classification rollout.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Complete issuer-authored Business Wire release read on an accessible syndication page. Distributor original timed out; identical release also retrieved from AOL. No independent outcome validation. Distributor URL: https://www.businesswire.com/news/home/20261020770911/en/", "independent_validation": "Not established", "note": "The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model."}}, {"Source_ID": "WATCH20UP-SPE-JPT", "Company": "SPE / OnePetro", "Title": "Industry Executives Offer Practical Advice on How To Use Modern AI Models Wisely", "URL": "https://jpt.spe.org/industry-executives-offer-practical-advice-on-how-to-use-modern-ai-models-wisely", "Source_date": "2026-09-10", "Source_type": "Issuer announcement and original publisher account", "Publisher": "Society of Petroleum Engineers / JPT", "Locator": "SPE’s AI Journey; licensing-demand discussion; following-conference Landmark paragraph; cost and metadata discussion", "Short_excerpt": "EnRG-LLM remains available for licensing", "Accessed": "2026-10-20", "Verification": "Full original publisher article retrieved and read 20 October. September 10 publication header; later following-conference paragraph is undated. Direct current text takes precedence over stale search extracts.", "Notes": "The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model. Classification release: October 20, 2026. JPT article: September 10 publication header, with an undated following-conference licensing update. Neither dates completed classification rollout.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Full original publisher article retrieved and read 20 October. September 10 publication header; later following-conference paragraph is undated. Direct current text takes precedence over stale search extracts.", "independent_validation": "Not established", "note": "The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model."}}, {"Source_ID": "WATCH20UP-BP", "Company": "bp", "Title": "bp awards Halliburton integrated contract for Bumerangue field appraisal in Brazil", "URL": "https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil", "Source_date": "2026-08-25", "Source_type": "Supplier announcement", "Publisher": "Halliburton", "Locator": "Opening and contract-scope paragraphs; LOGIX sentence", "Short_excerpt": "LOGIX™ automation and remote operations will be deployed", "Accessed": "2026-10-20", "Verification": "Full original supplier release retrieved (HTTP 200) and read. Future-tense scope; no completion or measured result established.", "Notes": "A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform. Original announcement August 25, 2026; checked October 20. This is older evidence newly added to this review, not a launch announced today.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Full original supplier release retrieved (HTTP 200) and read. Future-tense scope; no completion or measured result established.", "independent_validation": "Not established", "note": "A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform."}}, {"Source_ID": "WATCH20PM-USEDC", "Company": "U.S. Energy Development Corporation", "Title": "The Bottleneck Isn’t the Model. It’s the Execution.", "URL": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "Source_date": "2026-07", "Source_type": "Operator/vendor-coauthored field account", "Publisher": "The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors", "Locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "Short_excerpt": "We activated the system across 12 Permian wells beginning in December 2025.", "Accessed": "2026-10-20", "Verification": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.", "Notes": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case. Started December 2025. July 2026 article; exact publication day unknown. Six-month results, not a year or the operator’s full fleet. This older case was missed in earlier research.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.", "independent_validation": "Not established", "note": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case."}}, {"Source_ID": "WATCH20PM-CADE", "Company": "Ecopetrol", "Title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "URL": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "Source_date": "2026-05-04", "Source_type": "Operator announcement", "Publisher": "Ecopetrol", "Locator": "May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs", "Short_excerpt": "En el campo Rubiales, desde marzo de 2026, se gestionan 173 MW de energía", "Accessed": "2026-10-20", "Verification": "Full original operator page retrieved (HTTP200) and read by editor. Coverage and plans remain company-reported; no independent cost or performance validation.", "Notes": "No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population. Issuer announcement May 4, 2026; Rubiales coverage reported from March 2026. Older evidence newly checked, not a new October launch.", "Current_review": {"last_checked": "2026-10-20", "source_read": "Full original operator page retrieved (HTTP200) and read by editor. Coverage and plans remain company-reported; no independent cost or performance validation.", "independent_validation": "Not established", "note": "No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population."}}, {"Source_ID": "WATCH21AM-PTTEP", "Company": "PTTEP", "Title": "PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy", "URL": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "Source_date": "2025-11-03", "Source_type": "Operator technology feature", "Publisher": "PTTEP", "Locator": "Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples", "Short_excerpt": "PTTEP has launched its in-house Enterprise AI Engine, “X.brain,”", "Accessed": "2026-10-21", "Verification": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.", "Notes": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.", "independent_validation": "Not established", "note": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons."}}, {"Source_ID": "WATCH21AM-MNVR-CASE", "Company": "Energy Solutions / MNVR AI", "Title": "Five Failures Across Five Systems", "URL": "https://www.mnvrai.ai/s/MNVR-Customer-Story-Energy-Solutions-5-Examples-09222026.pdf", "Source_date": "Unknown", "Source_type": "Vendor/maintenance-partner customer case", "Publisher": "MNVR AI / Energy Solutions", "Locator": "PDF pp. 1–3: seven generators; five fault cases; estimated costs; six-week period and retained diagnoses", "Short_excerpt": "Over six weeks, ODIE identified five separate developing failures", "Accessed": "2026-10-21", "Verification": "Original three-page vendor/partner PDF retrieved HTTP200, full text read and pages 1–2 visually inspected. No publication date; filename is not date evidence.", "Notes": "Selected equipment faults are not failed AI deployments. No total alert population, missed-fault count, false-positive rate, control group or full system cost. Some repairs were scheduled or could be planned; not all five were completed. Roughly 5,200 signals means data signals, not assets. Publication date unknown; checked 21 October 2026. Six-week calendar window unknown. PDF filename is not a verified date. Adjacent gas-to-power operations, not an upstream adoption count.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original three-page vendor/partner PDF retrieved HTTP200, full text read and pages 1–2 visually inspected. No publication date; filename is not date evidence.", "independent_validation": "Not established", "note": "Selected equipment faults are not failed AI deployments. No total alert population, missed-fault count, false-positive rate, control group or full system cost. Some repairs were scheduled or could be planned; not all five were completed. Roughly 5,200 signals means data signals, not assets."}}, {"Source_ID": "WATCH21AM-MNVR-PRODUCT", "Company": "Energy Solutions / MNVR AI", "Title": "ODIE — Operations Diagnostic Intelligence Engine", "URL": "https://www.mnvrai.ai/odie", "Source_date": "Unknown", "Source_type": "Vendor product description", "Publisher": "MNVR AI", "Locator": "Human in the loop / operators make decisions section", "Short_excerpt": "ODIE informs decisions. Operators make them.", "Accessed": "2026-10-21", "Verification": "Original product page retrieved HTTP200 and read by editor. Undated product description; not a field performance validation.", "Notes": "Selected equipment faults are not failed AI deployments. No total alert population, missed-fault count, false-positive rate, control group or full system cost. Some repairs were scheduled or could be planned; not all five were completed. Roughly 5,200 signals means data signals, not assets. Publication date unknown; checked 21 October 2026. Six-week calendar window unknown. PDF filename is not a verified date. Adjacent gas-to-power operations, not an upstream adoption count.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original product page retrieved HTTP200 and read by editor. Undated product description; not a field performance validation.", "independent_validation": "Not established", "note": "Selected equipment faults are not failed AI deployments. No total alert population, missed-fault count, false-positive rate, control group or full system cost. Some repairs were scheduled or could be planned; not all five were completed. Roughly 5,200 signals means data signals, not assets."}}, {"Source_ID": "WATCH21EX-SLB", "Company": "SLB", "Title": "SLB Announces Second-Quarter 2026 Results", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "Source_date": "2026-07-24", "Source_type": "Issuer earnings release", "Publisher": "SLB", "Locator": "Core / Technology Highlights; automated cementing paragraph and Shell Velox-1 follow-up", "Short_excerpt": "surpassed 100 remote and automated cementing jobs", "Accessed": "2026-10-21", "Verification": "Original issuer release opened and relevant technology-highlight paragraphs read by editor. Remote and automated categories are not broken out; no AI method specified.", "Notes": "Automation and remote operation do not establish machine learning or LLM use. Human supervision, control envelopes, fallback and override rules are not specified. The same release’s AI Production Studio is a pre-field demonstration environment, not proof that those capabilities ran on Shell’s well.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original issuer release opened and relevant technology-highlight paragraphs read by editor. Remote and automated categories are not broken out; no AI method specified.", "independent_validation": "Not established", "note": "Automation and remote operation do not establish machine learning or LLM use. Human supervision, control envelopes, fallback and override rules are not specified. The same release’s AI Production Studio is a pre-field demonstration environment, not proof that those capabilities ran on Shell’s well."}}, {"Source_ID": "WATCH21PM-SHELL", "Company": "Shell", "Title": "The Data Detectives — Solving the 10-year-old problem at Perdido", "URL": "https://www.shell.com/what-we-do/digitalisation/artificial-intelligence.html", "Source_date": "Unknown", "Source_type": "Operator video transcript", "Publisher": "Shell", "Locator": "Artificial intelligence examples / Perdido video transcript; initial blank, plateau, operator input and refined precursor signature", "Short_excerpt": "the refined algorithm identifies the chemical fingerprint ahead of 70 percent of pump disruptions", "Accessed": "2026-10-21", "Verification": "Original operator webpage and relevant video transcript read by editor. Page/video publication and operating dates are not established; historical or undated context only.", "Notes": "The transcript does not establish a live alert or control workflow, prevented disruptions or realized savings. The two-day example is not an average warning time. Initial stalled analysis is not proof of an abandoned or failed field deployment. Undated webpage and video; operating and evaluation dates unknown. Already mentioned in the coverage log, now structured as a readable case. It is not a new 2026 deployment.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original operator webpage and relevant video transcript read by editor. Page/video publication and operating dates are not established; historical or undated context only.", "independent_validation": "Not established", "note": "The transcript does not establish a live alert or control workflow, prevented disruptions or realized savings. The two-day example is not an average warning time. Initial stalled analysis is not proof of an abandoned or failed field deployment."}}, {"Source_ID": "WATCH21PM-DBX", "Company": "Quantum Capital Group", "Title": "Transforming energy deal evaluation with Lakebase", "URL": "https://www.databricks.com/customers/quantum-capital-group", "Source_date": "Unknown", "Source_type": "Vendor customer story", "Publisher": "Databricks", "Locator": "Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs", "Short_excerpt": "The firm now runs over 11,000 unique data queries per month", "Accessed": "2026-10-21", "Verification": "Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.", "Notes": "No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution. Undated source; adoption and evaluation periods unknown. Extends existing sponsor platform AI-052, separately from QB Energy’s field-planning workflow AI-054.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.", "independent_validation": "Not established", "note": "No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution."}}, {"Source_ID": "WATCH21PM-IEA", "Company": "International Energy Agency", "Title": "Powering Artificial Intelligence: Strategies for integrating data centres into electricity systems", "URL": "https://www.iea.org/reports/powering-artificial-intelligence", "Source_date": "2026-10-14", "Source_type": "Institutional policy and project-pipeline report", "Publisher": "International Energy Agency", "Locator": "Dated landing page; executive summary; PDF pp.6–10, 36–47 and 58–62; gas-generation pipeline on p.60", "Short_excerpt": "Connection queues contain duplicative or speculative projects", "Accessed": "2026-10-21", "Verification": "Original IEA landing page, executive summary and relevant report passages read. Date verified on landing page. Policy and project-pipeline analysis, not operator AI adoption or contracted gas demand.", "Notes": "Project announcements do not establish completion, utilization, fuel burn or firm offtake. The report’s abstract covers 34 policy jurisdictions and about 90% of global data-centre capacity; those are policy-study scope measures, not an oil-and-gas AI-adoption sample. Published 14 October 2026. Gas-project figure uses IEA analysis of Global Energy Monitor data from August 2026. A distinct October edition, not a new announcement today or an amount to add to the April IEA report.", "Current_review": {"last_checked": "2026-10-21", "source_read": "Original IEA landing page, executive summary and relevant report passages read. Date verified on landing page. Policy and project-pipeline analysis, not operator AI adoption or contracted gas demand.", "independent_validation": "Not established", "note": "Project announcements do not establish completion, utilization, fuel burn or firm offtake. The report’s abstract covers 34 policy jurisdictions and about 90% of global data-centre capacity; those are policy-study scope measures, not an oil-and-gas AI-adoption sample."}}, {"Source_ID": "WATCH22AM-FAST", "Company": "TC Energy", "Title": "Practical AI in Action: TC Energy’s Path to a Unified, Scalable Asset Reliability Platform", "URL": "https://cdn.mediavalet.com/eunl/content/W8TklEW5w0msg_eHg0KrTA/gz5XgM6QhE6XBVDXhH9-pA/Original/Practical%20AI%20in%20Action%3A%20TC%20Energy%27s%20Path%20to%20a%20Unified%2C%20Scalable%20Asset%20Reliability%20Platform.pdf", "Source_date": "Unknown", "Source_type": "Joint customer conference presentation; vendor opinion", "Publisher": "TC Energy and Accenture / AVEVA World 2026", "Locator": "PDF pp.14–16: advisory workflow, upcoming FAST and release roadmap; p.17: architecture; p.20: forecast; p.22: future AI work; p.23: release wording", "Short_excerpt": "FAST is TC Energy’s upcoming Facility Asset Surveillance Technology", "Accessed": "2026-10-22", "Verification": "Original 27-page presentation downloaded and read by editor; pp.14,16,22 visually inspected. Deck internally dated 29 May 2026; exact publication date unknown.", "Notes": "The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features. Undated AVEVA World 2026 presentation, internally dated 29 May. Page 16 schedules Canada for Q3 2026 and US/Mexico for Q4; page 23 describes a first release in under seven months. The 16 October opinion does not resolve completion.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Original 27-page presentation downloaded and read by editor; pp.14,16,22 visually inspected. Deck internally dated 29 May 2026; exact publication date unknown.", "independent_validation": "Not established", "note": "The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features."}}, {"Source_ID": "WATCH22AM-AVEVA", "Company": "TC Energy", "Title": "How the oil and gas industry is putting data first", "URL": "https://www.aveva.com/en/our-industrial-life/type/opinion/how-the-oil-and-gas-industry-is-putting-data-first/", "Source_date": "2026-10-16", "Source_type": "Joint customer conference presentation; vendor opinion", "Publisher": "AVEVA / Cindy Crow", "Locator": "Posted date; Improved reliability through AI monitoring section", "Short_excerpt": "the company saved $100 million of value in failure avoidance per year", "Accessed": "2026-10-22", "Verification": "Original vendor opinion read by editor. Named-customer program claims are not independent outcome validation or release confirmation.", "Notes": "The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features. Undated AVEVA World 2026 presentation, internally dated 29 May. Page 16 schedules Canada for Q3 2026 and US/Mexico for Q4; page 23 describes a first release in under seven months. The 16 October opinion does not resolve completion.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Original vendor opinion read by editor. Named-customer program claims are not independent outcome validation or release confirmation.", "independent_validation": "Not established", "note": "The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features."}}, {"Source_ID": "WATCH22EX-FANG", "Company": "Diamondback Energy", "Title": "AI Lead – Value and Economics · R100885", "URL": "https://diamondbackenergy.wd12.myworkdayjobs.com/DBE/job/Midland-TX/AI-Lead---Value-and-Economics_R100885", "Source_date": "Unknown", "Source_type": "Official hiring description", "Publisher": "Diamondback Energy / official Workday careers", "Locator": "Job Responsibilities: valuation framework, baselines/comparison groups, benefits attribution, continue/scale/pivot/stop recommendations. Official public jobPostingInfo payload retrieved.", "Short_excerpt": "recommend whether investments should continue, scale, pivot, or stop", "Accessed": "2026-10-22", "Verification": "Editor retrieved and read the full official public Workday job payload. Human-facing page returns no extracted text; search copy corroborates responsibilities. Publication date not verified; metadata startDate is not a verified hire or operating date.", "Notes": "A job description establishes intended responsibilities. It does not show that these controls exist today or that any project was stopped. Existing Diamondback operating initiatives retain their own evidence. The posting was available when checked on 22 October 2026. Its publication date and whether the role has been filled are unverified.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor retrieved and read the full official public Workday job payload. Human-facing page returns no extracted text; search copy corroborates responsibilities. Publication date not verified; metadata startDate is not a verified hire or operating date.", "independent_validation": "Not established", "note": "A job description establishes intended responsibilities. It does not show that these controls exist today or that any project was stopped. Existing Diamondback operating initiatives retain their own evidence."}}, {"Source_ID": "WATCH22EX-ITALGAS", "Company": "Italgas", "Title": "Building Production Voice Agents with Databricks and AI Gateway", "URL": "https://answers.databricks.com/building-production-voice-agents-databricks-ai-gateway-w8XJClcgjx4", "Source_date": "Unknown", "Source_type": "Vendor-hosted customer conference transcript", "Publisher": "Databricks / named Italgas and Reply speakers", "Locator": "Full transcript: 02:04 and09:38 baseline conflict;10:10–16:57 architecture and CRM;19:28–20:15 limited beta, planned September rollout and claimed changes;21:05 human operators.", "Short_excerpt": "we are in a kind of beta testing moment", "Accessed": "2026-10-22", "Verification": "Editor read vendor-hosted named-speaker transcript, not just its summary/FAQs. Original video link fetch failed. Transcript baseline numbers conflict; exact talk/publication date and full rollout remain unverified.", "Notes": "The transcript’s earlier baseline figures do not reconcile. The summary omits the beta boundary. Do not reconstruct absolute outcomes, infer full rollout or treat these claims as measured upstream results. Undated talk. Full call coverage was planned for September, with no verified year or completion. It is not confirmed as a fresh announcement.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor read vendor-hosted named-speaker transcript, not just its summary/FAQs. Original video link fetch failed. Transcript baseline numbers conflict; exact talk/publication date and full rollout remain unverified.", "independent_validation": "Not established", "note": "The transcript’s earlier baseline figures do not reconcile. The summary omits the beta boundary. Do not reconstruct absolute outcomes, infer full rollout or treat these claims as measured upstream results."}}, {"Source_ID": "WATCH22FAIL-SPE", "Company": "SPE pressure-prediction study", "Title": "Evaluation of Influential Parameters for Prediction of Flowing Bottom Hole Pressure in Dry Gas Wells Using Machine Learning", "URL": "https://onepetro.org/SPENAIC/proceedings-abstract/26NAIC/26NAIC/SPE-234889-MS/802231", "Source_date": "2026-08-10", "Source_type": "Publisher-indexed technical abstract", "Publisher": "Society of Petroleum Engineers / OnePetro", "Locator": "SPE-234889-MS; publisher-indexed abstract: dataset/split and Random Forest/ANN comparison paragraphs. Published 10 August 2026; conference 3–5 August 2026.", "Short_excerpt": "ANN model performed poorly", "Accessed": "2026-10-22", "Verification": "Editor independently read the search-indexed original publisher abstract and metadata. Direct OnePetro/DOI opens failed; local page fetch returned 403. Full paper, raw data and complete methods not reviewed. No claim of field deployment or independent replication.", "Notes": "Only the publisher-indexed abstract was reviewed; direct access failed. No raw data, full validation protocol, deployed system, stop decision or later outcome verified. This does not establish a failure rate for industry AI projects. Published 10 August 2026. The ProBHP data are attributed to sources from 1975 and 1986; collection dates, well count, geography and test chronology are unspecified.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor independently read the search-indexed original publisher abstract and metadata. Direct OnePetro/DOI opens failed; local page fetch returned 403. Full paper, raw data and complete methods not reviewed. No claim of field deployment or independent replication.", "independent_validation": "Not established", "note": "Only the publisher-indexed abstract was reviewed; direct access failed. No raw data, full validation protocol, deployed system, stop decision or later outcome verified. This does not establish a failure rate for industry AI projects."}}, {"Source_ID": "WATCH22ROI-EY", "Company": "Oil, gas and chemicals survey", "Title": "Technology Adoption and Workforce Readiness in Oil & Gas and Chemicals — EY Future of Energy Survey", "URL": "https://www.ey.com/content/dam/ey-unified-site/ey-com/en-us/campaigns/energy-resources/documents/ey-ogc-future-of-energy-external-report.pdf", "Source_date": "2025-03", "Source_type": "Original executive survey", "Publisher": "EY", "Locator": "Printed pp. 7–8: expectations and scaling; pp. 27–29: methodology and profile (PDF indices 6–7, 26–28).", "Short_excerpt": "just 27% say emerging technologies such as AI are delivering on their promise", "Accessed": "2026-10-22", "Verification": "Editor read the publisher-hosted PDF, including findings and methodology. Exact item bases and respondent microdata are not supplied.", "Notes": "Existing October 5 research, newly indexed in the central library. Pooled oil, gas and chemicals; no realized cash ROI or project failure rate."}, {"Source_ID": "WATCH22PM-VISTA", "Company": "Vista Energy", "Title": "Vista Investor Day 2025", "URL": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "Source_date": "2025-11-12", "Source_type": "Company investor presentation", "Publisher": "Vista Energy", "Locator": "Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.", "Short_excerpt": "Implemented since March 2025", "Accessed": "2026-10-22", "Verification": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "Notes": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "independent_validation": "Not established", "note": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted."}}, {"Source_ID": "WATCH22PM-VISTA-SEC", "Company": "Vista Energy", "Title": "Vista Investor Day 2025 — SEC exhibit", "URL": "https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm", "Source_date": "2025-11-12", "Source_type": "Company investor presentation", "Publisher": "Vista Energy / SEC EDGAR", "Locator": "Exhibit heading: November 12, 2025; same slide 19 presentation.", "Short_excerpt": "Vista Investor Day", "Accessed": "2026-10-22", "Verification": "Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes.", "Notes": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes.", "independent_validation": "Not established", "note": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted."}}, {"Source_ID": "WATCH22PM-SAIPEM", "Company": "Saipem", "Title": "Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000", "URL": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "Source_date": "2026-02-23", "Source_type": "Company announcement", "Publisher": "Saipem", "Locator": "Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.", "Short_excerpt": "This pilot project", "Accessed": "2026-10-22", "Verification": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "Notes": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed. Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "independent_validation": "Not established", "note": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed."}}, {"Source_ID": "WATCH22PM-WORLEY", "Company": "Worley", "Title": "FY2025 full-year results presentation", "URL": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "Source_date": "2025-08-27", "Source_type": "Company investor presentation", "Publisher": "Worley", "Locator": "Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.", "Short_excerpt": "Recently deployed solutions through our Advanced Development Lab", "Accessed": "2026-10-22", "Verification": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "Notes": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established. FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "independent_validation": "Not established", "note": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established."}}, {"Source_ID": "WATCH22PM-WHITECAP", "Company": "Whitecap Resources", "Title": "Whitecap Resources customer story", "URL": "https://www.shakudo.io/customers/whitecap-resources", "Source_date": "2026-04-23", "Source_type": "Vendor customer story", "Publisher": "Shakudo", "Locator": "Key Results; govern-first platform, AI operating model, human review and endpoint being built toward. Page is undated; date comes from vendor customer index.", "Short_excerpt": "The endpoint being built toward", "Accessed": "2026-10-22", "Verification": "Editor read the original vendor customer story. Named customer executive is quoted; operator-side corroboration and outcome methods not verified.", "Notes": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved. Vendor index dated 23 April 2026; story itself undated. Platform use described, but project start and evaluation dates are unknown.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor read the original vendor customer story. Named customer executive is quoted; operator-side corroboration and outcome methods not verified.", "independent_validation": "Not established", "note": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved."}}, {"Source_ID": "WATCH22PM-WHITECAP-DATE", "Company": "Whitecap Resources", "Title": "Shakudo customer stories — Whitecap entry", "URL": "https://www.shakudo.io/customers", "Source_date": "2026-04-23", "Source_type": "Vendor customer story", "Publisher": "Shakudo", "Locator": "Whitecap customer-story listing dated April 23, 2026.", "Short_excerpt": "Whitecap Resources", "Accessed": "2026-10-22", "Verification": "Original vendor customer index read; verifies listed publication date, not a deployment or benefit.", "Notes": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved. Vendor index dated 23 April 2026; story itself undated. Platform use described, but project start and evaluation dates are unknown.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Original vendor customer index read; verifies listed publication date, not a deployment or benefit.", "independent_validation": "Not established", "note": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved."}}, {"Source_ID": "WATCH22PM-TEN", "Company": "Technip Energies", "Title": "Technip Energies 2025 Annual Report", "URL": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "Source_date": "2026 (exact publication day unverified)", "Source_type": "Indexed original annual-report excerpt", "Publisher": "Technip Energies", "Locator": "Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.", "Short_excerpt": "more than 55 digital and AI initiatives", "Accessed": "2026-10-22", "Verification": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "Notes": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved. 2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.", "Current_review": {"last_checked": "2026-10-22", "source_read": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "independent_validation": "Not established", "note": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved."}}, {"Source_ID": "WATCH23AM-SLB-Q3", "Company": "SLB", "Title": "SLB third-quarter 2026 results — selected operating and digital highlights", "URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Format": "pdf", "Source_date": "2026-10-23", "Source_type": "Supplier earnings release", "Publisher": "SLB", "Locator": "PDF p. 16: Eni, Azule, ADNOC and BSP Digital highlights; PDF p. 14: production chemistry. HTML release has the same paragraphs.", "Short_excerpt": "SLB Announces Third-Quarter 2026 Results", "Accessed": "2026-10-23", "Verification": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Notes": "Supplier customer claims, not independently validated. Company-wide financial performance and AI data-center demand are separate from operating AI benefits.", "Current_review": {"last_checked": "2026-10-23", "source_read": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "independent_validation": "Not established", "note": "Review covers the specified customer paragraphs; no call transcript reviewed."}}, {"Source_ID": "WATCH23AM-AZULE", "Company": "SLB", "Title": "Azule Energy pioneers digital well planning and rig automation in Africa", "URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa", "Format": "html", "Source_date": "Unknown", "Source_type": "Supplier customer case study", "Publisher": "SLB", "Locator": "Location; Results: Achieving operational excellence; Products Used: DrillOps AI-driven workflows.", "Short_excerpt": "digitally ingested the well plan", "Accessed": "2026-10-23", "Verification": "Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration.", "Notes": "Supplier customer claims, not independently validated. Company-wide financial performance and AI data-center demand are separate from operating AI benefits.", "Current_review": {"last_checked": "2026-10-23", "source_read": "Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration.", "independent_validation": "Not established", "note": "Review covers the specified customer paragraphs; no call transcript reviewed."}}, {"Source_ID": "WATCH23PM-PETROBRAS", "Company": "Petrobras", "Title": "Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice", "URL": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "Source_date": "Unknown", "Source_type": "Company supplier-channel pilot notice", "Publisher": "Petrobras", "Locator": "Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.", "Short_excerpt": "A supervisão e a decisão final permanecem sempre com nossos avaliadores técnicos.", "Accessed": "2026-10-23", "Verification": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "Notes": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.", "Current_review": {"last_checked": "2026-10-23", "source_read": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "independent_validation": "Not established", "note": "Human final decision is explicit. No observed failure or measured outcome disclosed."}}, {"Source_ID": "WATCH24AM-HAL-RELIABILITY", "Company": "Halliburton", "Title": "How predictive intelligence reduces downtime and improves equipment reliability", "URL": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "Source_date": "2026-10-22", "Source_type": "Supplier product account", "Publisher": "Halliburton", "Locator": "A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.", "Short_excerpt": "helped teams avoid an estimated $125,000 to $150,000 in downtime costs", "Accessed": "2026-10-24", "Verification": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "Notes": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "independent_validation": "Not established", "note": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016."}}, {"Source_ID": "WATCH24AM-HAL-PRODUCT", "Company": "Halliburton", "Title": "LOGIX Unit Vitality — cement equipment health monitoring", "URL": "https://www.halliburton.com/en/products/logix-unit-vitality", "Source_date": "Unknown", "Source_type": "Supplier product account", "Publisher": "Halliburton", "Locator": "Automated cement equipment health assessment; AI-driven models; Real-time monitoring closes the loop.", "Short_excerpt": "more than 400 real-time parameters", "Accessed": "2026-10-24", "Verification": "Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments.", "Notes": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments.", "independent_validation": "Not established", "note": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016."}}, {"Source_ID": "WATCH24AM-ONGC-ICE", "Company": "ONGC", "Title": "ONGC Technology — Project ICE", "URL": "https://www.ongcindia.com/web/eng/about-ongc/technology", "Source_date": "Unknown", "Source_type": "Company technology overview", "Publisher": "ONGC", "Locator": "Project ICE, five paragraphs. Site-wide last-updated stamp does not establish when this section or its roadmap was published.", "Short_excerpt": "24,000 users, and 23 SAP modules", "Accessed": "2026-10-24", "Verification": "Editor read the original ONGC Project ICE section. Legacy SAP R/3 rollout is dated 2004; section publication date and S/4HANA transition date are unknown. Website-wide last-updated label is not a section publication date.", "Notes": "Historical non-AI context only. AI-enabled innovation is a roadmap aim; no specific live AI workflow or decision authority established. A shared platform does not by itself prove clean data, adoption quality or AI readiness.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the original ONGC Project ICE section. Legacy SAP R/3 rollout is dated 2004; section publication date and S/4HANA transition date are unknown. Website-wide last-updated label is not a section publication date.", "independent_validation": "Not established", "note": "Historical non-AI context only. AI-enabled innovation is a roadmap aim; no specific live AI workflow or decision authority established. A shared platform does not by itself prove clean data, adoption quality or AI readiness."}}, {"Source_ID": "SLBTH-Q3", "Company": "SLB", "Title": "SLB third-quarter 2026 results — Digital economics and category definitions", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "Format": "html", "Source_date": "2026-10-23", "Source_type": "Supplier earnings release", "Publisher": "SLB", "Locator": "Digital results; Supplementary Information questions 10–13; Digital customer highlights.", "Short_excerpt": "$693 million decreased 1% sequentially", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-FY25", "Company": "SLB", "Title": "SLB full-year 2025 results — Digital financial baseline", "URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "Format": "html", "Source_date": "2026-01-23", "Source_type": "Supplier earnings release", "Publisher": "SLB", "Locator": "Supplementary Information question 12, Digital FY2025 table.", "Short_excerpt": "Digital revenue increased 9% on a full-year basis", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-MENON", "Company": "SLB", "Title": "Shashi Menon: The Race to Scale Digital Operations and AI", "URL": "https://investorcenter.slb.com/static-files/cf773c3d-348f-479f-aeb0-81f5418844ca", "Format": "pdf", "Source_date": "2026-06-17", "Source_type": "Investor presentation and prepared script", "Publisher": "SLB", "Locator": "PDF pp. 3–5: data access and control; p. 11: seismic training; pp. 14–15: monetization.", "Short_excerpt": "we trained it with public and SLB multiclient seismic data sets", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-BIGUET", "Company": "SLB", "Title": "Stephane Biguet: Proven Growth. Accretive Returns.", "URL": "https://investorcenter.slb.com/static-files/0835f6dc-d2fb-4bc0-b85e-afecadcf36fe", "Format": "pdf", "Source_date": "2026-06-17", "Source_type": "Investor presentation and prepared script", "Publisher": "SLB", "Locator": "PDF p. 4: customer base; pp. 7–10: commercial models; p. 11: 2030 ambitions.", "Short_excerpt": "more than 1,500 customers, including 90 of the world’s top 100 oil and gas producers", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-TERMS", "Company": "SLB", "Title": "SLB Online Services Terms and Conditions", "URL": "https://www.software.slb.com/slb-online-services-terms-and-conditions", "Format": "html", "Source_date": "Unknown", "Source_type": "Published online-service terms", "Publisher": "SLB", "Locator": "Definitions: Customer Data and Interaction Data; sections 8(a)–(b), 9(a), 10(a).", "Short_excerpt": "SLB claims no ownership rights in Customer Data.", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-GENAI", "Company": "SLB", "Title": "SLB Online Services Service Specific Terms", "URL": "https://www.software.slb.com/slb-online-services-service-specific-terms", "Format": "html", "Source_date": "Unknown", "Source_type": "Published online-service terms", "Publisher": "SLB", "Locator": "Generative AI Services: definitions; Connected Gen AI A(i), A(iv); Embedded Gen AI B(iii).", "Short_excerpt": "SLB shall not use Customer Prompts or Output to train models or further develop the Online Services without Customer’s consent.", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-MULLIN", "Company": "SLB", "Title": "Robert Mullin interview: 3.5% of the S&P, 20% of the Cash", "URL": "https://georgenoble.substack.com/p/35-of-the-s-and-p-20-of-the-cash", "Format": "html", "Source_date": "2026-10-20", "Source_type": "Investor opinion / interview", "Publisher": "The Noble Update", "Locator": "Host introduction and linked interview; exact SLB interview timestamp not established.", "Short_excerpt": "The oil service giant hiding a real AI business that could be worth more than the whole company", "Accessed": "2026-10-24", "Verification": "Original host introduction reviewed; confirms the investment framing and interview identity. Full interview passage and exact timestamp not independently checked. An Eisman episode containing the same argument remains unconfirmed.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Original host introduction reviewed; confirms the investment framing and interview identity. Full interview passage and exact timestamp not independently checked. An Eisman episode containing the same argument remains unconfirmed.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "SLBTH-RANNEN", "Company": "SLB", "Title": "Trygve Randen: Our Digital Advantage", "URL": "https://investorcenter.slb.com/static-files/7aa731c6-2f38-47e4-9aa2-473a8d04ec6e", "Format": "pdf", "Source_date": "2026-06-17", "Source_type": "Investor presentation and prepared script", "Publisher": "SLB", "Locator": "PDF pp. 2–3: governed platforms; pp. 5–6: hosting and trust; p. 7: operational footprint.", "Short_excerpt": "more than ninety percent of global production is simulated or modelled using at least one of our digital solutions", "Accessed": "2026-10-24", "Verification": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "Notes": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "independent_validation": "Not established", "note": "Investor opinion, supplier technology claims, segment economics and published standard terms are distinct evidence classes."}}, {"Source_ID": "WATCH24PM-WFRD-Q1", "Company": "Petroleum Development Oman", "Title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "Source_date": "2026-04-21", "Source_type": "Supplier earnings release", "Publisher": "Weatherford", "Locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "Short_excerpt": "moving the technology from pilot to operational use", "Accessed": "2026-10-24", "Verification": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "Notes": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "independent_validation": "Not established", "note": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case."}}, {"Source_ID": "WATCH24PM-WFRD-Q3", "Company": "Weatherford", "Title": "Weatherford Announces Third Quarter 2026 Results — rod-lift energy recovery", "URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18871", "Source_date": "2026-10-20", "Source_type": "Supplier earnings release", "Publisher": "Weatherford", "Locator": "Technology Highlights → Production and Intervention; first U.S. ForeSite Power Regenerative System retrofit paragraph.", "Short_excerpt": "delivering over 50% in energy savings", "Accessed": "2026-10-24", "Verification": "Editor read original issuer HTML and displayed 20 October 2026 publication date on 24 October. One unidentified U.S. well; supplier result, not independent validation.", "Notes": "Anonymous single-installation supplier claim. Do not treat the percentage as company-wide energy savings, production uplift, net cash ROI or an AI result. The word ForeSite links a brand family, not proof this equipment used the PDO predictive model.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original issuer HTML and displayed 20 October 2026 publication date on 24 October. One unidentified U.S. well; supplier result, not independent validation.", "independent_validation": "Not established", "note": "Anonymous single-installation supplier claim. Do not treat the percentage as company-wide energy savings, production uplift, net cash ROI or an AI result. The word ForeSite links a brand family, not proof this equipment used the PDO predictive model."}}, {"Source_ID": "DEEP24-MARS", "Company": "UNEP IMEO / MARS", "Title": "Operational machine learning for remote spectroscopic detection of CH4 point sources", "URL": "https://arxiv.org/html/2511.07719v2", "Source_date": "2026-03-24", "Source_type": "Original technical preprint and publisher abstract", "Publisher": "Růžička and coauthors / arXiv", "Locator": "Version 2, 24 March 2026; sections 3.1, 5.1, 6.2 and Table 2; section 6.4 and Table 5.", "Short_excerpt": "the models remain useful in guiding the reviewers", "Accessed": "2026-10-24", "Verification": "Editor read original full-text methods, Table 2 and operational workflow. Author preprint, listed as in review on arXiv; raw data/code not replicated. First version 11 November 2025.", "Notes": "A successful redesign, not a cancelled operator project. Results span emissions sectors; no new named oil-and-gas operator adoption is established. Raw data and code were not rerun. A separate 17 October conference abstract reports a similar benchmark-to-operation gap; its full chapter was unavailable.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original full-text methods, Table 2 and operational workflow. Author preprint, listed as in review on arXiv; raw data/code not replicated. First version 11 November 2025.", "independent_validation": "Not replicated by editor", "note": "A successful redesign, not a cancelled operator project. Results span emissions sectors; no new named oil-and-gas operator adoption is established. Raw data and code were not rerun. A separate 17 October conference abstract reports a similar benchmark-to-operation gap; its full chapter was unavailable."}}, {"Source_ID": "DEEP24-SPRINGER", "Company": "UNEP IMEO / MARS", "Title": "Reducing False Positives in Operational Remote Sensing: Lessons from Methane Plume Detection", "URL": "https://doi.org/10.1007/978-3-032-41086-3_46", "Source_date": "2026-10-17", "Source_type": "Original technical preprint and publisher abstract", "Publisher": "Springer Nature / Discovery Science 2026", "Locator": "Publisher abstract and First Online metadata; chapter pp. 721–736. Full subscription chapter not reviewed.", "Short_excerpt": "generate substantially more operational false positives", "Accessed": "2026-10-24", "Verification": "Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass.", "Notes": "A successful redesign, not a cancelled operator project. Results span emissions sectors; no new named oil-and-gas operator adoption is established. Raw data and code were not rerun. A separate 17 October conference abstract reports a similar benchmark-to-operation gap; its full chapter was unavailable.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass.", "independent_validation": "Not replicated by editor", "note": "A successful redesign, not a cancelled operator project. Results span emissions sectors; no new named oil-and-gas operator adoption is established. Raw data and code were not rerun. A separate 17 October conference abstract reports a similar benchmark-to-operation gap; its full chapter was unavailable."}}, {"Source_ID": "DEEP24-HAVTIL", "Company": "Havtil / Norwegian petroleum operators", "Title": "Artificial intelligence (AI) in the petroleum industry is in need of better oversight and management", "URL": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/reports-from-projects/2026/artificial-intelligence-ai-in-the-petroleum-industry-is-in-need-of-better-oversight-and-management/", "Source_date": "2026-10-23", "Source_type": "Original regulator survey summary", "Publisher": "Norwegian Ocean Industry Authority (Havtil)", "Locator": "Varying levels of maturity; An overview is essential for effective risk management; Prudent development and use.", "Short_excerpt": "Many companies lack a comprehensive overview", "Accessed": "2026-10-24", "Verification": "Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page.", "Notes": "This is a regulator’s interpretation of company responses, not proof of a failed deployment, an enforcement finding against every operator or measured poor ROI. It does not rank companies.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page.", "independent_validation": "Not replicated by editor", "note": "This is a regulator’s interpretation of company responses, not proof of a failed deployment, an enforcement finding against every operator or measured poor ROI. It does not rank companies."}}, {"Source_ID": "SOL5-WESTBRAE", "Company": "Marathon Oil / West Brae (historical trial)", "Title": "Well Intelligence – Optimised Production", "URL": "https://www.netzerotc.com/case-studies/well-intelligence-optimised-production/", "Source_date": "Unknown", "Source_type": "Original project assessment and supplier case", "Publisher": "Net Zero Technology Centre", "Locator": "Project overview; Industry value; Lessons learned. Undated historical account naming Marathon and West Brae.", "Short_excerpt": "the application does not appear to add significant value to Marathon", "Accessed": "2026-10-24", "Verification": "Editor read original project summary and lessons. Exact trial/publication dates unknown; website copyright is not a date. Project assessment, not independently audited cash returns.", "Notes": "No cancellation, negative net cash return or later scale is established. Intelligent Plant has an undated anonymous account with similar trial details that claims benefits; we cannot confirm identical case identity or reconcile it with NZTC’s named assessment.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original project summary and lessons. Exact trial/publication dates unknown; website copyright is not a date. Project assessment, not independently audited cash returns.", "independent_validation": "Not replicated by editor", "note": "No cancellation, negative net cash return or later scale is established. Intelligent Plant has an undated anonymous account with similar trial details that claims benefits; we cannot confirm identical case identity or reconcile it with NZTC’s named assessment."}}, {"Source_ID": "SOL5-WELLINT-SUPPLIER", "Company": "Marathon Oil / West Brae (historical trial)", "Title": "Case Study: Enhancing Production with Well Intelligence AI Model", "URL": "https://www.intelligentplant.com/case-study-well-intelligence", "Source_date": "Unknown", "Source_type": "Original project assessment and supplier case", "Publisher": "Intelligent Plant", "Locator": "Challenge and Testing; Historical Data Review; Results and Impact. Anonymous client, no publication date.", "Short_excerpt": "This finding alone justified the deployment of Well Intelligence.", "Accessed": "2026-10-24", "Verification": "Editor read original supplier case. Anonymous account shares four-week/~8,000-bbl/day descriptors with NZTC account; same-trial identity not confirmed. No measured financial result or operator corroboration.", "Notes": "No cancellation, negative net cash return or later scale is established. Intelligent Plant has an undated anonymous account with similar trial details that claims benefits; we cannot confirm identical case identity or reconcile it with NZTC’s named assessment.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original supplier case. Anonymous account shares four-week/~8,000-bbl/day descriptors with NZTC account; same-trial identity not confirmed. No measured financial result or operator corroboration.", "independent_validation": "Not replicated by editor", "note": "No cancellation, negative net cash return or later scale is established. Intelligent Plant has an undated anonymous account with similar trial details that claims benefits; we cannot confirm identical case identity or reconcile it with NZTC’s named assessment."}}, {"Source_ID": "SOL5-WELLRECORDS", "Company": "Los Alamos / NETL well-record study", "Title": "Information extraction from historical well records using a large language model", "URL": "https://www.nature.com/articles/s41598-024-81846-5", "Source_date": "2024-12-30", "Source_type": "Original peer-reviewed technical study", "Publisher": "Ma, Santos, Lackey, Viswanathan and O’Malley / Scientific Reports", "Locator": "Methods: LLMs performance evaluation; Results: Prompt engineering, Tables 2–3, chain-of-thought and comparison with other LLMs. DOI 10.1038/s41598-024-81846-5.", "Short_excerpt": "none of the prompts resulted in completely correct extraction for the 10 Pennsylvania well records", "Accessed": "2026-10-24", "Verification": "Editor read original open-access publisher full text, metric definitions and results. Paper independently cross-checked by two SOL tracks; data/code not rerun. Peer-reviewed publication date 30 December 2024; tests use 2024-era models, not current capability.", "Notes": "Unequal 150- and 10-document cohorts; no independent untouched prompt-selection holdout established. Do not attribute all errors to the LLM, or infer failed operator rollout, measured savings or net ROI.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original open-access publisher full text, metric definitions and results. Paper independently cross-checked by two SOL tracks; data/code not rerun. Peer-reviewed publication date 30 December 2024; tests use 2024-era models, not current capability.", "independent_validation": "Not replicated by editor", "note": "Unequal 150- and 10-document cohorts; no independent untouched prompt-selection holdout established. Do not attribute all errors to the LLM, or infer failed operator rollout, measured savings or net ROI."}}, {"Source_ID": "SOL5-RAG", "Company": "Nigeria LNG-affiliated document-retrieval study", "Title": "Evaluating Chunking Strategies For Retrieval-Augmented Generation in Oil and Gas Enterprise Documents", "URL": "https://arxiv.org/pdf/2603.24556", "Source_date": "2026-03-25", "Source_type": "Original technical proceedings author manuscript", "Publisher": "Taiwo and Yusoff / CS&IT proceedings author manuscript", "Locator": "Original 11-page PDF: methods PDF pp. 4–7; Tables 1–3 PDF pp. 8–9, printed pp. 64–65. arXiv first submitted 25 March 2026; exact proceedings publication day unknown.", "Short_excerpt": "all four methods demonstrated limited effectiveness on P&IDs", "Accessed": "2026-10-24", "Verification": "Editor read original author PDF and visually inspected Tables 1–3. arXiv says manuscript matches CCSEIT 2026 proceedings; raw proprietary corpus and relevance labels unavailable. Author affiliation does not establish deployment or corpus ownership.", "Notes": "Small proprietary corpus; diagram-query count and annotator agreement undisclosed. No live-user, hallucination or adoption rate. Employer affiliation does not establish that these tested configurations were deployed at Nigeria LNG.", "Current_review": {"last_checked": "2026-10-24", "source_read": "Editor read original author PDF and visually inspected Tables 1–3. arXiv says manuscript matches CCSEIT 2026 proceedings; raw proprietary corpus and relevance labels unavailable. Author affiliation does not establish deployment or corpus ownership.", "independent_validation": "Not replicated by editor", "note": "Small proprietary corpus; diagram-query count and annotator agreement undisclosed. No live-user, hallucination or adoption rate. Employer affiliation does not establish that these tested configurations were deployed at Nigeria LNG."}}, {"Source_ID": "WATCH25AM-PTTEP-XBOT", "Company": "PTTEP", "Title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Source_date": "Unknown", "Source_type": "Operator annual report", "Publisher": "PTTEP", "Format": "pdf", "Locator": "Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate.", "Short_excerpt": "average processing time by five hours per purchase order", "Accessed": "2026-10-25", "Verification": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "Notes": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "independent_validation": "Not established", "note": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot."}}, {"Source_ID": "WATCH25AM-PTTEP-OPEX", "Company": "PTTEP", "Title": "PTTEP organizes OpEx Day 2025, advancing operational excellence", "URL": "https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence", "Source_date": "2025-07-09", "Source_type": "Operator press release", "Publisher": "PTTEP", "Format": "html", "Locator": "Press release, procurement AI sentence in operational-excellence examples.", "Short_excerpt": "application of AI to streamline procurement processes", "Accessed": "2026-10-25", "Verification": "Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.", "Notes": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.", "independent_validation": "Not established", "note": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot."}}, {"Source_ID": "WATCH25AM-BKV-RUNWAY", "Company": "BKV Corporation", "Title": "BKV 2025 Sustainability Report — Project Runway", "URL": "https://www.bkv.com/2025-sustainability-report-governance", "Source_date": "Unknown", "Source_type": "Operator annual report", "Publisher": "BKV Corporation", "Format": "html", "Locator": "Governance: Project Runway; Informed by field operations; Collaboration beyond implementation.", "Short_excerpt": "the consolidated Barnett platform went live on November 19, 2025", "Accessed": "2026-10-25", "Verification": "Editor read original operator page on 25 October 2026. Publication day not displayed; report landing page CEO letter dated September 2026. Metrics are issuer-reported.", "Notes": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original operator page on 25 October 2026. Publication day not displayed; report landing page CEO letter dated September 2026. Metrics are issuer-reported.", "independent_validation": "Not established", "note": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date."}}, {"Source_ID": "WATCH25AM-BKV-REPORT", "Company": "BKV Corporation", "Title": "BKV 2025 Sustainability Report — letter date", "URL": "https://www.bkv.com/2025-sustainability-report", "Source_date": "Unknown", "Source_type": "Operator report landing page", "Publisher": "BKV Corporation", "Format": "html", "Locator": "Letter from CEO footer: September 2026.", "Short_excerpt": "Letter from the CEO | September 2026", "Accessed": "2026-10-25", "Verification": "Editor read original report landing page; letter month is not exact report publication date.", "Notes": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original report landing page; letter month is not exact report publication date.", "independent_validation": "Not established", "note": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date."}}, {"Source_ID": "WATCH25AM-HARBOUR-COUPA", "Company": "Harbour Energy", "Title": "Harbour Energy Coupa Supplier Portal FAQ, version 4", "URL": "https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf", "Source_date": "2025-06", "Source_type": "Operator supplier guide", "Publisher": "Harbour Energy", "Format": "pdf", "Locator": "PDF p.1 version 4, June 2025 and PO process; pp.2–4 permissions, invoices, receipts and disputes; p.5 refers to ex-Wintershall Dea vendor guides.", "Short_excerpt": "All new purchase orders of Harbour Energy are generated through the Coupa platform.", "Accessed": "2026-10-25", "Verification": "Editor read original 5-page operator PDF. Exact revision day, go-live and legal-entity scope unknown. The linked vendor page separately retains SAP Business Network.", "Notes": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original 5-page operator PDF. Exact revision day, go-live and legal-entity scope unknown. The linked vendor page separately retains SAP Business Network.", "independent_validation": "Not established", "note": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates."}}, {"Source_ID": "WATCH25AM-HARBOUR-VENDORS", "Company": "Harbour Energy", "Title": "Harbour vendor relations — distinct supplier channels", "URL": "https://www.harbourenergy.com/vendor-relations/", "Source_date": "Unknown", "Source_type": "Operator supplier webpage", "Publisher": "Harbour Energy", "Format": "html", "Locator": "Separate sections: vendors using SAP Business Network; ex-Wintershall Dea vendors.", "Short_excerpt": "Applicable to ex-Wintershall Dea vendors", "Accessed": "2026-10-25", "Verification": "Editor read original operator page on 25 October 2026. Undated current channel guidance; SAP start November 2022 is older context, not new AI.", "Notes": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original operator page on 25 October 2026. Undated current channel guidance; SAP start November 2022 is older context, not new AI.", "independent_validation": "Not established", "note": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates."}}, {"Source_ID": "WATCH25AM-COP-GEP", "Company": "ConocoPhillips", "Title": "ConocoPhillips GEP SMART Introduction", "URL": "https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf", "Source_date": "2024-12", "Source_type": "Operator supplier guide", "Publisher": "ConocoPhillips", "Format": "pdf", "Locator": "PDF p.1 updated December 2024; p.3 transition target; p.5 intended benefits; p.6 supplier invoice flow.", "Short_excerpt": "all business units should have transitioned to GEP SMART", "Accessed": "2026-10-25", "Verification": "Editor read original 7-page issuer PDF. January 2025 is a target, not verified unit-by-unit completion. Exact revision day unknown.", "Notes": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original 7-page issuer PDF. January 2025 is a target, not verified unit-by-unit completion. Exact revision day unknown.", "independent_validation": "Not established", "note": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption."}}, {"Source_ID": "WATCH25AM-COP-VENDORS", "Company": "ConocoPhillips", "Title": "ConocoPhillips technical upgrades to supplier systems", "URL": "https://vendors.conocophillips.com/supplier-guide/technical-upgrades-to-supplier-systems/", "Source_date": "Unknown", "Source_type": "Operator supplier webpage", "Publisher": "ConocoPhillips", "Format": "html", "Locator": "GEP portal/EDI and Enverus routes; supplier training and implemented SSIS support.", "Short_excerpt": "Exceptions to the invoicing process are identified", "Accessed": "2026-10-25", "Verification": "Editor read original live supplier page on 25 October 2026. Page retains future-tense January 2025 language, so it cannot verify completion date. No publication date displayed.", "Notes": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption.", "Current_review": {"last_checked": "2026-10-25", "source_read": "Editor read original live supplier page on 25 October 2026. Page retains future-tense January 2025 language, so it cannot verify completion date. No publication date displayed.", "independent_validation": "Not established", "note": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption."}}, {"Source_ID": "WATCH25PM-YPF-EKONS", "Company": "YPF", "Title": "YPF refinery blending — interview with BCG project lead Roberto Rodrigo", "URL": "https://mase.lmneuquen.com/empresas/ia-ypf-la-tecnologia-que-gano-el-premio-digital-del-ano-y-ya-se-aplica-la-refineria-la-plata-n1253008", "Source_date": "Unknown", "Source_type": "Original supplier interview", "Publisher": "Más Energía / LM Neuquén; Martina Mongelluzzo", "Format": "html", "Locator": "Sections Del laboratorio a la operación en tiempo real and Argentina, el primer caso de implementación.", "Short_excerpt": "la solución puede reducir hasta un 50% el tiempo de ejecución de los blending", "Accessed": "2026-10-25", "Verification": "Editor read the original interview. The exact date is not displayed in its retrieved body; a syndicated copy is dated 11 September 2026. Original named supplier interview, not operator confirmation or independent performance validation.", "Notes": "Original interview with the supplier project lead, not an operator filing or independent result. No realized saving, quality/error measure, cost or net ROI. No stream, tank, blend, user or throughput denominator. Further rollout is an ambition. Separate from the earlier YPF cataloging and inspection profile."}, {"Source_ID": "WATCH25PM-DIXSTONE-IFS", "Company": "Dixstone", "Title": "Dixstone selects IFS Cloud for offshore services", "URL": "https://www.ifs.com/pl/insights/news/dixstone-selects-ifs-cloud-to-power-global-offshore-operations", "Source_date": "2025-12-09", "Source_type": "Supplier announcement", "Publisher": "IFS", "Format": "html", "Locator": "Dated release: selection, project controls and fleet maintenance paragraphs; About Dixstone separates affiliate status from Perenco E&P.", "Short_excerpt": "Dixstone has selected IFS Cloud", "Accessed": "2026-10-25", "Verification": "Editor read the original IFS release dated 9 December 2025. A vendor selection announcement with a customer quote; future capabilities and company footprint are not completed rollout or outcomes.", "Notes": "Vendor announcement, not proof of operating AI or delivered savings. No model, asset rollout, approval boundary or net ROI. Dixstone is a Perenco affiliate consolidating offshore services; this is not a Perenco E&P field deployment."}, {"Source_ID": "RESULTS-PERENCO", "Company": "Perenco", "Title": "Perenco enhances drilling efficiency with advanced automation solutions", "URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/perenco-enhances-drilling-efficiency-with-advanced-automation-solutions", "Source_date": "Unknown", "Source_type": "Supplier customer case study", "Publisher": "SLB", "Format": "html", "Locator": "Summary and Results; three-well 2025 campaign.", "Short_excerpt": "achieved a 19% increase in rate of penetration", "Accessed": "2026-10-25", "Verification": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "Notes": "Raw drilling-speed baseline, matched offset population, complete costs and AI-only contribution unknown."}, {"Source_ID": "RESULTS-WAFI", "Company": "Wafi Energy Pakistan", "Title": "From manual accounts payable to AI-powered finance in 12 weeks", "URL": "https://spendconsole.ai/success-stories/from-manual-accounts-payable-to-ai-powered-finance-in-12-weeks/", "Source_date": "Unknown", "Source_type": "Supplier customer case study", "Publisher": "SpendConsole", "Format": "html", "Locator": "Client profile; solution; results; automated-scenario qualification.", "Short_excerpt": "under a minute in automated scenarios", "Accessed": "2026-10-25", "Verification": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "Notes": "Invoice volume, automated-scenario share, precise measurement window and full costs unknown. This is the Pakistan Shell-brand licensee, not Shell-wide adoption."}, {"Source_ID": "RESULTS-APOLLO", "Company": "Apollo Resources / unnamed SENM operator", "Title": "HSV-01 chemical treatment case study", "URL": "https://apolloresourcesllc.net/apollo-resources-hsv-01-oil-treating-chemistry-effectively-eliminates-30000-ppm-of-h2s-in-senm/", "Source_date": "Unknown", "Source_type": "Supplier customer case study", "Publisher": "Apollo Resources", "Format": "html", "Locator": "Solution and Results; 30-day observation.", "Short_excerpt": "Chemical consumption decreased by 65%", "Accessed": "2026-10-25", "Verification": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "Notes": "Population, throughput, cost calculation, installation costs and longer-term durability unknown. New chemistry and equipment prevent attributing the result to software alone."}, {"Source_ID": "RESULTS-WEATHERFORD", "Company": "Weatherford / unnamed trial operators", "Title": "Maximizing chemical efficiency with rigless precision-based capillary injection", "URL": "https://www.weatherford.com/documents/papers/2025/maximizing-chemical-efficiency-and-reducing-environmental-footprint-with-rigless,-precision-based-ca/", "Source_date": "2025-09-16", "Source_type": "Supplier customer case study", "Publisher": "Weatherford / SPE", "Format": "html", "Locator": "Public abstract, SPE-227643-MS; field-trial results.", "Short_excerpt": "a 65% reduction in chemical usage per well", "Accessed": "2026-10-25", "Verification": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "Notes": "Operator, geography, well population, raw baseline, cost perimeter and installation costs unknown. This does not establish all-in ROI or scaled deployment."}, {"Source_ID": "RESULTS-SLB-CHEM", "Company": "SLB / unnamed Permian operator", "Title": "Permian field chemical automation case study", "URL": "https://www.slb.com/resource-library/case-study-with-navigation/production-chemicals/permian-basin-field-automations-cs", "Source_date": "Unknown", "Source_type": "Supplier earnings release / customer account", "Publisher": "SLB", "Format": "html", "Locator": "Results and solution sections; 20-month dosing comparison and first-nine-month savings are different periods.", "Short_excerpt": "USD 4.4 million in net savings in the first 9 months", "Accessed": "2026-10-25", "Verification": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "Notes": "An anonymous supplier account, without a controlled comparison or complete net ROI. The percentage change concerns the share of overtreated wells, not chemical volume or total cost. The case study and Q3 release are the same publisher, not independent corroboration."}, {"Source_ID": "D26-TOPTAL", "Company": "Toptal / unnamed oil-and-gas technology company", "Title": "Mineral-title evaluation and regression testing", "URL": "https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy", "Source_date": "Unknown", "Source_type": "Technical evaluation", "Publisher": "Toptal", "Format": "html", "Locator": "Challenge; Ground-truth AI Evaluation; Outcome. No publication/evaluation date displayed.", "Short_excerpt": "approximately 27% to 2% while maintaining roughly 93% recall", "Accessed": "2026-10-26", "Verification": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "Notes": "Anonymous client, undated provider case. No independently confirmed live error rate, realized labor saving, full cost or ROI. Reduced false positives does not eliminate human title review."}, {"Source_ID": "D26-SLB-DDR", "Company": "SLB", "Title": "AI approach for drilling operation insights — US 20260063025 A 1", "URL": "https://patents.google.com/patent/US20260063025A1/en", "Source_date": "2026-03-05", "Source_type": "Technical evaluation", "Publisher": "Schlumberger Technology Corporation / published US patent application", "Format": "html", "Locator": "Table 15 and section 5, Model Performance Comparative Analysis; published application, not a granted-patent or commercial-deployment claim.", "Short_excerpt": "especially for minority class when the training data is insufficient", "Accessed": "2026-10-26", "Verification": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "Notes": "Applicant-authored technical examples, not an independently validated field outcome. Other examples use differing cohorts and later model configurations. No measured reduction in drilling incidents or complete ROI."}, {"Source_ID": "D26-ESP-PAPER", "Company": "Baker Hughes / unnamed operator", "Title": "ESP failure prediction — SPE-221571-MS", "URL": "https://dam.bakerhughes.com/asset/2145a967-01f1-4ffa-8ee1-37b085e6fff1/spe-221571-ms-bkr-licensed.pdf", "Source_date": "2024-10-29/30", "Source_type": "Supplier-authored conference paper", "Publisher": "Baker Hughes authors / SPE", "Format": "pdf", "Locator": "Physical PDF pp.4–7: population, train/test split, event definitions, sensitivity, missed failures and daily prediction deployment. Conference dates 29–30 October 2024.", "Short_excerpt": "five out of the nine ESPs missed failures", "Accessed": "2026-10-26", "Verification": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "Notes": "Existing technical limitation rechecked; no new named adoption or complete ROI."}], "context": [{"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "Aramco launches Metabrain 250B-parameter industrial LLM", "date": "2024-03", "theme": "Enterprise GenAI rollout", "summary": "Aramco unveiled METABRAIN, a 250-billion-parameter generative AI model trained on about 7 trillion tokens of public and internal data spanning ~90 years of company history. It targets drilling-plan analysis, well recommendations and refined-product price/market forecasting, with a 1-trillion-parameter version planned.", "metrics": "250B parameters; 7T training tokens; ~90 years of data; 80+ AI professionals", "vendors": "Aramco Digital", "takeaway": "Analyst interpretation: Proprietary operational history is the moat; NOCs with deep archives are building domain LLMs rather than relying only on general-purpose models.", "source": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm", "source_name": "Middle East AI News", "id": 1, "ID": "CTX-REF-001", "Reference_ID": "REF-001", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R001", "Audit_review": {"id": "CTX-REF-001", "company": "Saudi Aramco", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Historical March 4, 2024 secondary account; model size and intended tasks are reported claims. The planned larger model is not proof of completion.", "correction": "", "source_url": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Chevron", "segment": "Supermajor", "hq": "USA", "title": "Up to 20% workforce cut, technology and hubs cited", "date": "2025-02", "theme": "Strategy & org", "summary": "Unverified inherited reference: Chevron said it would cut 15-20% of its global workforce (up to ~9,000 roles) by end-2026 to save $2-3B. Levers included leveraging technology to raise productivity and changing where work is done via global capability centers such as Manila, Buenos Aires and its Bengaluru ENGINE center.", "metrics": "Unverified inherited figures; source access unresolved.", "vendors": "", "takeaway": "Analyst interpretation: Majors are pairing technology-enabled productivity with centralized service hubs; AI enablement pitches should target work consolidated into these centers.", "source": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce", "source_name": "Inspectioneering", "id": 14, "ID": "CTX-REF-014", "Reference_ID": "REF-014", "Category": "Context / survey / enabling ecosystem", "Evidence": "Source access unresolved", "Source_ID": "R014", "Audit_review": {"id": "CTX-REF-014", "company": "Chevron", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference.", "correction": "", "source_url": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce"}], "residual_gaps": ["Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference."]}}, {"company": "Shell", "segment": "Supermajor", "hq": "UK", "title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "date": "2024-09-09", "theme": "Safety & emissions", "summary": "Shell and NVIDIA used Fourier Neural Operators in NVIDIA PhysicsNeMo to build surrogate models of CO2 storage that run about 100,000x faster than numerical simulators while keeping R-squared above 0.97. This enables screening tens of thousands of storage configurations for CCS site selection.", "metrics": "~100,000x speedup; R-squared >0.97; 1,000-year simulation horizon", "vendors": "NVIDIA PhysicsNeMo", "takeaway": "Analyst interpretation: Physics-ML surrogates turn slow simulation into fast scenario screening, a strong pattern for subsurface and CCS workflows.", "source": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/", "source_name": "NVIDIA Technical Blog", "id": 18, "ID": "CTX-REF-018", "Reference_ID": "REF-018", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R018", "Audit_review": {"id": "CTX-REF-018", "company": "Shell", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "NVIDIA dates this technical article September 9, 2024, before the reporting window. Computational speedup and accuracy describe model tests, not field-wide economic returns.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/"}], "residual_gaps": []}}, {"company": "BP", "segment": "Supermajor", "hq": "UK", "title": "bp renews Palantir for 5 years with AIP LLMs", "date": "2024-09", "theme": "Partnership & investment", "summary": "Unverified inherited reference: bp renewed its decade-long Palantir relationship with a five-year strategic agreement centered on Palantir's Artificial Intelligence Platform (AIP). AIP brings LLMs into bp's operations on top of digital twins fed by more than 2 million sensors to speed human decision-making.", "metrics": "Unverified inherited figures; source access unresolved.", "vendors": "Palantir (Foundry, AIP)", "takeaway": "Analyst interpretation: Operators with an existing ontology and digital-twin layer can add LLM copilots quickly; the data model is the durable asset.", "source": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/", "source_name": "Palantir IR", "id": 19, "ID": "CTX-REF-019", "Reference_ID": "REF-019", "Category": "Date eligibility needs checking", "Evidence": "Source access unresolved", "Source_ID": "R019", "Audit_review": {"id": "CTX-REF-019", "company": "BP", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here.", "correction": "", "source_url": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/"}], "residual_gaps": ["Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here."]}}, {"company": "BP", "segment": "Supermajor", "hq": "UK", "title": "bp expands job cuts as AI drives efficiency", "date": "2025-08", "theme": "Strategy & org", "summary": "Unverified inherited reference: bp raised expected job losses to about 6,200 (roughly 15% of office-based staff) plus contractor cuts. CEO Murray Auchincloss said technology is helping improve capital productivity and drive cost reductions, with AI part of the overhaul toward a $5B cost-cut target by end-2027.", "metrics": "Unverified inherited figures; source access unresolved.", "vendors": "", "takeaway": "Analyst interpretation: AI is now explicitly linked to headcount plans at majors; change management and redeployment belong in the business case alongside productivity.", "source": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html", "source_name": "Yahoo Finance UK", "id": 20, "ID": "CTX-REF-020", "Reference_ID": "REF-020", "Category": "Context / survey / enabling ecosystem", "Evidence": "Source access unresolved", "Source_ID": "R020", "Audit_review": {"id": "CTX-REF-020", "company": "BP", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI.", "correction": "", "source_url": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html"}], "residual_gaps": ["Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI."]}}, {"company": "TotalEnergies", "segment": "Supermajor", "hq": "France", "title": "TotalEnergies describes supervised AI and prior Copilot adoption", "date": "2025-01-24", "theme": "Enterprise GenAI rollout", "summary": "TotalEnergies said it was one of the first companies to adopt Microsoft's generative AI assistant to boost employee productivity. Its 300-person Digital Factory runs AI for production forecasting, predictive maintenance and renewables siting, with experts keeping the final say.", "metrics": "300 AI/digital experts in Digital Factory", "vendors": "Microsoft Copilot", "takeaway": "Analyst interpretation: Horizontal copilots plus a central digital factory for vertical use cases is the standard two-track model among majors.", "source": "https://totalenergies.com/news/news/ia-expediting-energy-transition", "source_name": "TotalEnergies", "id": 23, "ID": "CTX-REF-023", "Reference_ID": "REF-023", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R023", "Audit_review": {"id": "CTX-REF-023", "company": "TotalEnergies", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://totalenergies.com/news/news/ia-expediting-energy-transition"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The referenced TotalEnergies article is dated January 24, 2025, not February 2024. It discusses a prior Copilot adoption and continuing supervised AI use; it is context alongside existing TotalEnergies dossiers.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://totalenergies.com/news/news/ia-expediting-energy-transition"}], "residual_gaps": []}}, {"company": "Petrobras", "segment": "NOC", "hq": "Brazil", "title": "ChatPetrobras GenAI rolled out to 110,000 workers", "date": "2023-12", "theme": "Enterprise GenAI rollout", "summary": "Petrobras built ChatPetrobras with Microsoft on Azure OpenAI Service (GPT models) and made it available through an internal portal to 110,000 employees and contractors. It supports reports, summaries, presentations and Q&A to cut time and cost in administrative and operational processes.", "metrics": "110,000 workers (employees + contractors)", "vendors": "Microsoft, Azure OpenAI", "takeaway": "Analyst interpretation: A secure internal ChatGPT-style wrapper extended to contractors reaches the whole workforce quickly and cheaply.", "source": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br", "source_name": "Microsoft News Center Latam", "id": 30, "ID": "CTX-REF-030", "Reference_ID": "REF-030", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R030", "Audit_review": {"id": "CTX-REF-030", "company": "Petrobras", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Microsoft publication is December 5, 2023. It describes ChatPetrobras using Azure OpenAI; accessibility to workers is not a measured active-user count.", "correction": "", "source_url": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Marathon Petroleum", "segment": "Refining", "hq": "US", "title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "date": "Undated; refers to 2024 AFPM Summit", "theme": "Production & operations", "summary": "Imubit reports that Marathon Petroleum presented a 25% reduction in suboptimal coke-drum cycles at the 2024 AFPM Summit. The vendor describes AI-based closed-loop process optimization. The linked page is undated and does not establish a separate Flyscan investment.", "metrics": "Vendor-reported: 25% fewer suboptimal drum cycles; facility, measurement period and realized savings not specified here.", "vendors": "Imubit", "takeaway": "Analyst interpretation: Reinforcement-learning optimizers layered on existing APC can deliver measurable unit gains without replacing controls.", "source": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes", "source_name": "Imubit", "id": 66, "ID": "CTX-REF-066", "Reference_ID": "REF-066", "Category": "Date eligibility needs checking", "Evidence": "Context source excerpt reviewed", "Source_ID": "R066", "Audit_review": {"id": "CTX-REF-066", "company": "Marathon Petroleum", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The undated Imubit page attributes the 25% drum-cycle result to Marathon at the 2024 AFPM Summit. It does not support the separate Flyscan investment claim; removed that claim.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes"}], "residual_gaps": []}}, {"company": "Palantir", "segment": "Big tech", "hq": "United States", "title": "APA extends multi-year Palantir deal, adds AIP", "date": "2024-09-25", "theme": "Enterprise GenAI rollout", "summary": "APA signed a multi-year, multimillion-dollar extension of its 2021 Palantir enterprise agreement, adding the AIP platform. Palantir already supports APA's operational planning, supply chain, maintenance, production optimization and contract management, and AIP is used to flag anomalies in contracts and invoices against source systems.", "metrics": "Multi-year, multimillion-dollar", "vendors": "Palantir AIP, Foundry", "takeaway": "Analyst interpretation: Palantir expands inside E&Ps from an operational data layer into back-office genAI such as invoice and contract checks, so land-and-expand from one ontology is the model to beat.", "source": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html", "source_name": "StockTitan (Business Wire release)", "id": 88, "ID": "CTX-REF-088", "Reference_ID": "REF-088", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R088", "Audit_review": {"id": "CTX-REF-088", "company": "Palantir", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Issuer release reproduced by StockTitan is September 25, 2024, one day before the window. Retained as historical context, not a current-period new initiative.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html"}], "residual_gaps": []}}, {"company": "AWS", "segment": "Big tech", "hq": "United States", "title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "date": "2024-08-07", "theme": "Production & operations", "summary": "bpx energy, which runs bp's US onshore assets, built Xpert with the AWS Generative AI Innovation Center. Xpert is a Bedrock-based chat tool that turns employee questions into SQL or Python, runs the code against bpx production data and answers in text, and is intended to extend to mobile and voice for field staff.", "metrics": "No quantified Xpert-specific operating result established by this review.", "vendors": "Amazon Bedrock, AWS Generative AI Innovation Center, bpx energy", "takeaway": "Analyst interpretation: conversational access to operating data is a useful workflow; company production growth cannot be attributed to the assistant.", "source": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/", "source_name": "AWS for Industries blog", "id": 89, "ID": "CTX-REF-089", "Reference_ID": "REF-089", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R089", "Audit_review": {"id": "CTX-REF-089", "company": "AWS", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "AWS article is August 7, 2024. The 30% bpx production ambition is not an Xpert result; removed it from the metric field.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/"}], "residual_gaps": []}}, {"company": "Databricks", "segment": "Big tech", "hq": "United States", "title": "Data Intelligence Platform for Energy with genAI", "date": "2024-04-09", "theme": "Enterprise GenAI rollout", "summary": "Databricks launched an energy-specific Data Intelligence Platform bringing generative AI to the sector, with partners including AVEVA, Seeq, Deloitte and Capgemini. Shell's digital innovation VP said Databricks had accelerated its analytics and AI.", "metrics": "No energy-specific customer count or operating benefit established by this launch release.", "vendors": "Shell, TotalEnergies, Chevron Phillips Chemical, Cosmo Energy, Wood Mackenzie, AVEVA, Seeq", "takeaway": "Analyst interpretation: Majors' AI and ML work often runs on Databricks, so building natively on its lakehouse (Unity Catalog, model serving) lowers the barrier to buying.", "source": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html", "source_name": "PR Newswire", "id": 91, "ID": "CTX-REF-091", "Reference_ID": "REF-091", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R091", "Audit_review": {"id": "CTX-REF-091", "company": "Databricks", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Databricks issuer release is April 9, 2024. Its organization count is company-wide, not energy adoption; excluded from initiative performance metrics.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Enverus", "segment": "Software incumbent", "hq": "United States", "title": "Instant Analyst genAI with cited answers", "date": "2024-02-20", "theme": "Trading & commercial", "summary": "Enverus launched Instant Analyst, generative AI over its Intelligence Vault research that returns natural-language answers with document citations to limit hallucinations. It runs on Amazon Bedrock and later expanded to other products, including Courthouse records.", "metrics": "Searches tens of thousands of proprietary research docs", "vendors": "Amazon Bedrock, Enverus Intelligence Vault, PRISM, Tenaris, Ironwood Midstream", "takeaway": "Analyst interpretation: Answers with citations over proprietary data are what buyers now expect as a minimum for O&G genAI.", "source": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/", "source_name": "Enverus", "id": 97, "ID": "CTX-REF-097", "Reference_ID": "REF-097", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R097", "Audit_review": {"id": "CTX-REF-097", "company": "Enverus", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Enverus dates the launch February 20, 2024. Later product extensions require their own disclosure evidence.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "McKinsey & Company", "segment": "Survey & research", "hq": "United States", "title": "AI could unlock $230B a year in upstream", "date": "2026-08", "theme": "Adoption data", "summary": "McKinsey sized AI value across the upstream value chain: $65B a year near-term with today's technology, $125B as proven tools scale, and $230B at full potential with autonomous operations. The figures are net of more than $30B a year in implementation costs. Value is concentrated: the top 10 use cases account for nearly half of it, and production optimization and drilling are the largest pools.", "metrics": "$65B / $125B / $230B annual value; >$30B/yr implementation cost; top 10 use cases ~50% of value, top 60 ~95%; <20% of deployments track AI value KPIs; +$35B/yr reserves value from exploration; OFS revenue displacement $17B near-term to $60B", "vendors": "McKinsey", "takeaway": "Analyst interpretation: Fewer than 20% of deployments track AI value, so building value measurement into the enablement program is part of the product.", "source": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value", "source_name": "Oil & Gas Journal", "id": 108, "ID": "CTX-REF-108", "Reference_ID": "REF-108", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R108", "Audit_review": {"id": "CTX-REF-108", "company": "McKinsey & Company", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Oil & Gas Journal reports McKinsey scenarios on August 26, 2026. They are modeled industry opportunities, not realized AI returns; underlying McKinsey model was not independently audited.", "correction": "", "source_url": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Rystad Energy", "segment": "Survey & research", "hq": "Norway", "title": "Digital and AI: a $500B upstream opportunity", "date": "Undated source; 2026–2030 forecast period", "theme": "Adoption data", "summary": "Rystad estimates nearly $500 billion of cumulative value from digitalization and AI in E&P during 2026–2030, with cost reductions and production increases contributing roughly equally. It groups workflows into asset development; operations and maintenance; exploration and reservoir development; and drilling, wells and production. This is a modeled opportunity, not realized industry returns.", "metrics": "Rystad forecast: about $500 billion cumulative value in 2026–2030; $80 billion additional annual value in 2030 versus 2025.", "vendors": "Rystad Energy, ADNOC, Equinor", "takeaway": "Analyst interpretation: Frame use-case portfolios around the few value pools the analysts agree on (drilling, maintenance, reservoir) instead of spreading effort thin.", "source": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas", "source_name": "Rystad Energy", "id": 109, "ID": "CTX-REF-109", "Reference_ID": "REF-109", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R109", "Audit_review": {"id": "CTX-REF-109", "company": "Rystad Energy", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Rystad describes four workflow categories for the $500 billion estimate. The earlier five-area wording came from a separate digitalization article and was removed; publication month is not visible in the opened page.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas"}], "residual_gaps": []}}, {"company": "Rystad Energy", "segment": "Survey & research", "hq": "Norway", "title": "Digitalization could save $320B by 2030", "date": "2025-12", "theme": "Adoption data", "summary": "Rystad's supply-chain team estimated $320B of oil and gas savings from digitalization through 2030 and called the figure modest. It named high upfront hardware and software costs, cybersecurity requirements and aging infrastructure as the main barriers, especially for smaller operators and service firms.", "metrics": "$320B savings through 2030; SLB digital targeting ~35% margins in 2025; Viridien digital/data revenue $787M (+17%)", "vendors": "Rystad Energy, SLB, Halliburton, NOV, Baker Hughes, Viridien", "takeaway": "Analyst interpretation: Smaller operators face cost and cyber barriers, so modular, low-infrastructure enablement offerings are an underserved gap.", "source": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/", "source_name": "World Oil", "id": 110, "ID": "CTX-REF-110", "Reference_ID": "REF-110", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R110", "Audit_review": {"id": "CTX-REF-110", "company": "Rystad Energy", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "World Oil attributes the December 5, 2025 estimate to Rystad. It is a digitalization opportunity including AI, not an AI-only savings result.", "correction": "", "source_url": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Boston Consulting Group", "segment": "Survey & research", "hq": "United States", "title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "date": "2025-08", "theme": "Strategy & org", "summary": "BCG presents a potential 30–70% improvement in oil-and-gas EBIT over five years and organizes AI strategy into deployment of tools, redesign of workflows and new business models. This is consulting analysis rather than measured sector-wide profitability.", "metrics": "BCG modeled potential: 30–70% EBIT improvement over five years; not an achieved return.", "vendors": "BCG, Cognite, SparkCognition, Microsoft, OpenAI (ChatGPT Enterprise), SLB (Delfi), Aker BP, bp", "takeaway": "Analyst interpretation: Seat licences (Deploy) are table stakes. Most of the value sits in redesigning a few core workflows, so enablement has to go past prompt training.", "source": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf", "source_name": "BCG", "id": 111, "ID": "CTX-REF-111", "Reference_ID": "REF-111", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R111", "Audit_review": {"id": "CTX-REF-111", "company": "Boston Consulting Group", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "BCG PDF executive summary presents 30–70% potential EBIT improvement. It is a scenario, not achieved industry profit; other chart statistics require a separate detailed methodology review.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Bain & Company", "segment": "Survey & research", "hq": "United States", "title": "Energy execs: fewer than 20% scaling AI with results", "date": "2026-03-19", "theme": "Adoption data", "summary": "Bain surveyed 859 executives across oil and gas, utilities, chemicals, mining and agribusiness in December 2025–January 2026. Its headline says fewer than 20% are scaling AI with results; the body separately says about a quarter are scaling applications and reimagining functions with measurable impact. These formulations are retained as distinct source statements, not interchangeable adoption rates.", "metrics": "Mixed-sector survey, n=859. Headline: fewer than 20% scaling with results. Body: about one-quarter scaling applications/reimagining functions; fewer than 10% companywide transformation.", "vendors": "Bain & Company", "takeaway": "Analyst interpretation: The top barrier is unclear business value, not technology, so tie each use case to an owner and a P&L metric before building it.", "source": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/", "source_name": "Bain & Company", "id": 112, "ID": "CTX-REF-112", "Reference_ID": "REF-112", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R112", "Audit_review": {"id": "CTX-REF-112", "company": "Bain & Company", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Bain publication is March 19, 2026, not February. Its headline says fewer than 20% scaling with results while body says a quarter scaling applications/reimagining functions; preserve the source distinction and mixed-sector survey population.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/"}], "residual_gaps": []}}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "Oil, gas and chemicals executives: 85% value reskilling; 29% retrain", "date": "2025-03", "theme": "Workforce enablement", "summary": "EY’s survey of 1,020 energy-industry respondents in November–December 2024 found a gap between reskilling intentions and activity. Within the combined oil-and-gas-and-chemicals executive cohort, 85% said reskilling would determine success, while 29% reported current retraining and 58% said culture impeded digital adoption.", "metrics": "Combined oil/gas/chemicals cohort: 85% reskilling importance; 29% retraining; 58% cultural impediment. Total energy survey n=1,020.", "vendors": "EY", "takeaway": "Analyst interpretation: The market gap is in doing the reskilling, not in convincing people it matters.", "source": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey", "source_name": "EY", "id": 113, "ID": "CTX-REF-113", "Reference_ID": "REF-113", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R113", "Audit_review": {"id": "CTX-REF-113", "company": "EY", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "EY groups oil and gas AND chemicals executives together for the 85%/29% findings. Earlier text incorrectly treated these as oil-and-gas-only statistics.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey"}], "residual_gaps": []}}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "Energy leaders struggle to link AI to productivity", "date": "2026-04-13", "theme": "Adoption data", "summary": "Wave 4 of EY's US AI Pulse Survey found that energy leaders see AI as a catalyst for transformation but have trouble proving its impact. 56% of energy senior leaders struggle to link productivity gains directly to AI, compared with 35% across all industries. 72% say more training is needed to communicate AI productivity gains accurately. EY also notes that employee job-security concerns are holding back experimentation.", "metrics": "December 2025 survey wave: energy leaders 56% versus all-industry 35% struggle with AI-productivity attribution; 72% versus 50% see a training need. Not oil-and-gas-only.", "vendors": "EY", "takeaway": "Analyst interpretation: Energy lags other sectors on measuring AI gains, so offering an attribution method alongside training is a differentiator.", "source": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption", "source_name": "EY", "id": 114, "ID": "CTX-REF-114", "Reference_ID": "REF-114", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R114", "Audit_review": {"id": "CTX-REF-114", "company": "EY", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "EY article is April 13, 2026; December 2025 is the survey wave. The statistics describe energy respondents and should not be presented as an oil-and-gas-only sample.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption"}], "residual_gaps": []}}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "96% of O&G execs cite data and vendor challenges", "date": "2025-01", "theme": "Adoption data", "summary": "EY reports that 96% of oil and gas executives see managing alliance and vendor relationships, along with data accessibility, as a challenge for AI and emerging technology. 96% also cite connecting data across organizational silos. EY says the hard part starts when a proof of concept is handed to the team responsible for scaling it.", "metrics": "96% cite vendor/alliance management and data accessibility; 96% cite cross-silo data connection", "vendors": "EY", "takeaway": "Analyst interpretation: Plan the pilot-to-production handoff (data plumbing, vendor orchestration, ownership) as part of any use case, not as an afterthought.", "source": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company", "source_name": "EY", "id": 115, "ID": "CTX-REF-115", "Reference_ID": "REF-115", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R115", "Audit_review": {"id": "CTX-REF-115", "company": "EY", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "EY January 6, 2025 article supports the 96% data/access/silo challenge figures. Survey methodology was not fully reviewed; this is context, not a deployment.", "correction": "", "source_url": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Deloitte", "segment": "Survey & research", "hq": "United States", "title": "AI heading past 50% of O&G IT spend by 2029", "date": "2025-10", "theme": "Adoption data", "summary": "Deloitte's 2026 Oil and Gas Industry Outlook projects that AI and GenAI will grow from less than 20% of US O&G IT spending today to more than 50% by 2029. About half of AI investment goes to process optimization. Early adopters of AI-driven maintenance report up to 40% fewer equipment failures and $10M a year in savings. The outlook expects GenAI, agentic AI and real-time analytics to move from pilots to enterprise deployment in 2026.", "metrics": "AI <20% of IT spend now to >50% by 2029; up to 40% fewer failures; $10M/yr savings; 140+ hours downtime prevented, 1.6% uptime protected (one case); 66% of workforce in mechanically intensive roles", "vendors": "Deloitte", "takeaway": "Analyst interpretation: As budgets swing toward AI, the scarce input becomes people who can use it, which is a direct opening for enablement.", "source": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html", "source_name": "Deloitte Insights", "id": 116, "ID": "CTX-REF-116", "Reference_ID": "REF-116", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R116", "Audit_review": {"id": "CTX-REF-116", "company": "Deloitte", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Deloitte presents IT-spending forecasts and selected maintenance examples. Spending share is a projection; isolated vendor-derived maintenance results are not typical industry outcomes.", "correction": "", "source_url": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Deloitte", "segment": "Survey & research", "hq": "United States", "title": "1.2M US energy and chemicals workers need upskilling", "date": "2025-05", "theme": "Workforce enablement", "summary": "Deloitte forecasts that 1.2 million US energy-and-chemicals workers need upskilling and 534,000 need reskilling. Of analyzed task hours, 64% remain human-driven; 36% involve humans in the loop, split between mechanical/electronic automation (26%) and generative/agentic AI (10%). These are workforce projections, not observed job displacement.", "metrics": "Projected: 1.2 million upskilling; 534,000 reskilling; 26% mechanical/electronic automation task hours and 10% generative/agentic AI task hours.", "vendors": "Deloitte", "takeaway": "Analyst interpretation: Upskilling demand is workforce-wide, not only for engineers, so programs need role-based tracks that include field and operations staff.", "source": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html", "source_name": "Deloitte via PR Newswire", "id": 117, "ID": "CTX-REF-117", "Reference_ID": "REF-117", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R117", "Audit_review": {"id": "CTX-REF-117", "company": "Deloitte", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Deloitte describes 36% of task hours as human-in-the-loop: 26% mechanical/electronic automation plus 10% generative/agentic AI. It is incorrect to imply one-third of all tasks are agentic-AI-led.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html"}], "residual_gaps": []}}, {"company": "KPMG", "segment": "Survey & research", "hq": "Netherlands", "title": "Energy CEOs: only 40% actively reskilling for AI", "date": "2025-11", "theme": "Workforce enablement", "summary": "KPMG's 2025 Global Energy, Natural Resources and Chemicals CEO Outlook found that CEOs are confident in AI's potential but lag on workforce action. 80% agree their leadership understands AI's disruptive potential, but only 40% are actively reskilling or upskilling the roles AI affects. 82% believe AI can support emissions reduction and energy efficiency.", "metrics": "40% actively reskilling AI-affected roles; 80% say leadership understands AI disruption; 82% see AI supporting emissions reduction", "vendors": "KPMG", "takeaway": "Analyst interpretation: CEO awareness has outrun action on skills, so position enablement as how the AI strategy gets executed.", "source": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html", "source_name": "KPMG", "id": 118, "ID": "CTX-REF-118", "Reference_ID": "REF-118", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R118", "Audit_review": {"id": "CTX-REF-118", "company": "KPMG", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "KPMG landing-page figures are 40%, 80% and 82% for the energy, natural resources and chemicals CEO cohort, not exclusively oil and gas.", "correction": "", "source_url": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "DNV", "segment": "Survey & research", "hq": "Norway", "title": "Only 12% of O&G firms use AI in live operations", "date": "2024-10", "theme": "Adoption data", "summary": "DNV describes low operational AI maturity among its oil-and-gas respondents: 12% reported live day-to-day use and 3% advanced integration, while 47% described planning or pilots. Its separate strategic-intention table reports net agreement, defined as agreement minus disagreement; those values are not simple adoption shares.", "metrics": "Reported maturity: 12% live use; 3% advanced use; 47% planning/piloting. Separate 62% digital-strategy and 47% forward-intention figures are net agreement.", "vendors": "DNV, Mordor Intelligence", "takeaway": "Analyst interpretation: Assurance (showing that AI is safe, secure and validated) is itself an adoption lever in a safety-critical industry.", "source": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "source_name": "DNV", "id": 119, "ID": "CTX-REF-119", "Reference_ID": "REF-119", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R119", "Audit_review": {"id": "CTX-REF-119", "company": "DNV", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "DNV table percentages are NET agreement (agree minus disagree). Its prose also distinguishes 12% live use and 3% advanced use. Removed simple adoption-percentage interpretation of the net-agreement table.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/"}], "residual_gaps": []}}, {"company": "DNV", "segment": "Survey & research", "hq": "Norway", "title": "60% plan AI deployment; human-AI interaction a key risk", "date": "2025-09", "theme": "Safety & emissions", "summary": "At SPE Offshore Europe 2025, DNV reported that 60% of 1,160 surveyed energy companies planned to deploy AI-driven applications in 2025, up from 47% in 2024. It proposed a systemic safety framework that combines safety engineering with AI assurance across the lifecycle, and it flagged human-AI interaction as a primary risk area.", "metrics": "60% planning AI deployment in 2025 (vs 47% in 2024); 1,160 companies surveyed", "vendors": "DNV, SPE", "takeaway": "Analyst interpretation: Teach how to supervise AI (when to override, how to verify), not just how to use it.", "source": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/", "source_name": "Oilfield Technology", "id": 120, "ID": "CTX-REF-120", "Reference_ID": "REF-120", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R120", "Audit_review": {"id": "CTX-REF-120", "company": "DNV", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Oilfield Technology reports DNV intentions for the year ahead, not observed completed deployment. The surveyed population spans energy companies, not solely oil and gas.", "correction": "", "source_url": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "DNV Cyber", "segment": "Survey & research", "hq": "Norway", "title": "65% of energy firms see cyber as top risk", "date": "2025-01", "theme": "Safety & emissions", "summary": "DNV's Energy Cyber Priority 2025 surveyed 375 energy professionals. 65% say their leadership views cybersecurity as the greatest risk to the business, and 71% say their organization is more vulnerable to OT cyber events than ever. 57% admit their OT defences lag their IT defences, and the report names embracing AI and embedding new workforce skills as core challenges.", "metrics": "65% cyber = greatest risk; 71% more OT-vulnerable (vs 64% 2023); 57% OT defences lag IT; 67% expect more OT security investment; n=375", "vendors": "DNV Cyber", "takeaway": "Analyst interpretation: Any agentic or OT-adjacent AI program has to be designed with OT security from the start; cyber is a gating objection.", "source": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/", "source_name": "DNV", "id": 121, "ID": "CTX-REF-121", "Reference_ID": "REF-121", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R121", "Audit_review": {"id": "CTX-REF-121", "company": "DNV Cyber", "outcome": "supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "DNV January 22, 2025 release supports the retained 375-respondent energy-sector cyber figures; these are respondent views rather than measured incident rates.", "correction": "", "source_url": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/"}], "residual_gaps": []}}, {"company": "International Energy Agency", "segment": "Survey & research", "hq": "France", "title": "IEA: oil and gas an early AI adopter", "date": "2025-04", "theme": "Adoption data", "summary": "IEA’s April 10, 2025 Energy and AI report examines both electricity needs for AI and AI’s potential to change energy operations. This review confirmed the report and scope but did not recheck the detailed oil-and-gas supercomputer statistics in its underlying chapter.", "metrics": "Detailed supercomputer counts are withheld pending chapter-level verification.", "vendors": "IEA", "takeaway": "Analyst interpretation: Even the IEA ranks skills alongside data and security as a core barrier, which supports treating enablement as infrastructure.", "source": "https://www.iea.org/reports/energy-and-ai", "source_name": "IEA", "id": 122, "ID": "CTX-REF-122", "Reference_ID": "REF-122", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R122", "Audit_review": {"id": "CTX-REF-122", "company": "International Energy Agency", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.iea.org/reports/energy-and-ai"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "IEA report landing page dated April 10, 2025 was reopened. The detailed supercomputer statistics were not checked in the underlying chapter during this context audit.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.iea.org/reports/energy-and-ai"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "PwC", "segment": "Survey & research", "hq": "United Kingdom", "title": "20% of companies capture 74% of AI value", "date": "2026-04", "theme": "Adoption data", "summary": "PwC's 2026 AI Performance Study of 1,217 senior executives across 25 sectors found that 20% of organizations capture about three-quarters of AI's economic value. The rest are largely stuck in pilot mode. Leaders use AI for growth and business-model reinvention rather than productivity alone. The study is cross-industry, not specific to oil and gas.", "metrics": "74% of value to 20% of firms; n=1,217 across 25 sectors", "vendors": "PwC", "takeaway": "Analyst interpretation: AI value is winner-take-most, so the case for enablement is competitive positioning, not just efficiency.", "source": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html", "source_name": "PwC", "id": 123, "ID": "CTX-REF-123", "Reference_ID": "REF-123", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R123", "Audit_review": {"id": "CTX-REF-123", "company": "PwC", "outcome": "supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "PwC April 13, 2026 release supports 74%/20% and the 1,217-executive, 25-sector population. This is a cross-industry self-report study, not oil-and-gas operating evidence.", "correction": "", "source_url": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html"}], "residual_gaps": []}}, {"company": "Novi Labs", "segment": "Survey & research", "hq": "United States", "title": "Reservoir engineer AI use rises to 56%", "date": "Undated article; compares 2024–2025 survey waves", "theme": "Subsurface & exploration", "summary": "Novi Labs surveyed more than 150 US reservoir engineers and found AI adoption rose from 43% in 2024 to 56% in 2025. Well forecasting use rose from 58% to 75%. 84% reported problems with the quality and accessibility of upstream public data, and teams using proprietary data reported much higher success than those relying only on public sources.", "metrics": "Vendor survey: adoption 43% to 56%; forecasting use 58% to 75%; 84% report public-data challenges. More than 150 US reservoir engineers; not an operator census.", "vendors": "Novi Labs", "takeaway": "Analyst interpretation: Engineers now own data quality (74%), so enablement should teach data stewardship alongside AI tools.", "source": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/", "source_name": "Novi Labs", "id": 124, "ID": "CTX-REF-124", "Reference_ID": "REF-124", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R124", "Audit_review": {"id": "CTX-REF-124", "company": "Novi Labs", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Novi vendor survey supports headline adoption and well-forecasting figures among more than 150 US reservoir engineers. Self-selected/vendor survey is not a census of operators; publication month not confirmed.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Shell", "segment": "Supermajor", "hq": "United Kingdom", "title": "Shell's DIY citizen developer program passes 4,000 builders", "date": "2023-11-07", "theme": "Workforce enablement", "summary": "Microsoft’s November 2023 Shell case describes a Power Platform citizen-development program with more than 4,000 active DIY developers. It concerns low-code applications and automation; the developer count is not an AI-specific user count.", "metrics": "More than 4,000 active DIY developers, partner-reported in 2023; not an AI-adoption metric.", "vendors": "Microsoft (Power Platform)", "takeaway": "Analyst interpretation: Shell's playbook (CoE, embedded coaches, persona-based paths, hackathons) transfers directly from low-code to GenAI and agent building.", "source": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk", "source_name": "Microsoft Customer Stories", "id": 125, "ID": "CTX-REF-125", "Reference_ID": "REF-125", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R125", "Audit_review": {"id": "CTX-REF-125", "company": "Shell", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Microsoft case is November 7, 2023, not June. DIY is low-code/citizen development; the disclosed developer count does not establish AI adoption.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk"}], "residual_gaps": []}}, {"company": "Society of Petroleum Engineers", "segment": "Industry body", "hq": "United States", "title": "SPE Forum: AI in Upstream E&P 2030", "date": "2026-04", "theme": "Workforce enablement", "summary": "SPE held a closed forum in Santa Fe (April 7-10, 2026), following its 2024 AI Forum, on how AI is reshaping upstream engineering. Sessions covered which skills preserve the human edge, whether AI can acquire tacit knowledge, how to build trust in models, and how to design human-AI teaming and ethical AI use by 2030.", "metrics": "", "vendors": "SPE", "takeaway": "Analyst interpretation: The profession's own agenda is about trust, tacit knowledge and human-AI teaming, so enablement content should address these as well as tools.", "source": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond", "source_name": "SPE", "id": 127, "ID": "CTX-REF-127", "Reference_ID": "REF-127", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R127", "Audit_review": {"id": "CTX-REF-127", "company": "Society of Petroleum Engineers", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Event page establishes April 7–10, 2026 forum and topics. Agenda is not evidence of completed operator deployments or measured benefits.", "correction": "", "source_url": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "The Open Group OSDU Forum", "segment": "Industry body", "hq": "United States", "title": "OSDU pushes an AI-ready open data platform", "date": "2025-12", "theme": "Subsurface & exploration", "summary": "The Open Group’s December 4, 2025 summit recap describes an OSDU community focus on analytics, AI-ready data access and seismic-interpretation workflows. Testimonials identify community implementation and certification as future priorities; the post alone does not establish those deliverables as complete.", "metrics": "", "vendors": "The Open Group, OSDU Forum", "takeaway": "Analyst interpretation: Subsurface AI enablement depends on standardized data foundations such as OSDU; data literacy belongs in the curriculum.", "source": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/", "source_name": "The Open Group", "id": 128, "ID": "CTX-REF-128", "Reference_ID": "REF-128", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R128", "Audit_review": {"id": "CTX-REF-128", "company": "The Open Group OSDU Forum", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The Open Group post describes work and forward priorities. Community implementation and certification should not be represented as completed deliverables based on forward-looking testimonials.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/"}], "residual_gaps": []}}, {"company": "Rice University / Chevron", "segment": "Industry body", "hq": "United States", "title": "Chevron Digital Scholars earn Rice data-science master's", "date": "Historical context; exact article date not re-established", "theme": "Workforce enablement", "summary": "Chevron sponsored employees as Digital Scholars to complete a full-time Master of Computer Science at Rice University, concentrating in data science and machine learning. The first cohort had 21 scholars. Scholars completed D2K Lab capstones on real Chevron problems, such as modeling reservoir productivity decline, and returned as digital change agents.", "metrics": "Cohort counts not reverified in this pass; retained as workforce-development context.", "vendors": "Chevron, Rice University (D2K Lab)", "takeaway": "Analyst interpretation: Deep, degree-level programs for a small cadre of domain experts complement broad AI literacy and create in-house translators.", "source": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation", "source_name": "Rice University D2K Lab", "id": 129, "ID": "CTX-REF-129", "Reference_ID": "REF-129", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R129", "Audit_review": {"id": "CTX-REF-129", "company": "Rice University / Chevron", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Rice interview supports Chevron Digital Scholars and a reservoir-decline capstone. Exact publication date and complete cohort metrics were not re-established from the reviewed excerpt.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "US Department of Energy", "segment": "Government", "hq": "United States", "title": "Genesis Mission targets AI for subsurface energy", "date": "2026-02-12", "theme": "Subsurface & exploration", "summary": "DOE announced 26 Genesis Mission science and technology challenges on February 12, 2026. One challenge concerns AI modeling of underground environments for responsible, cost-effective development of subsurface energy resources. This is a research agenda, not evidence of a deployed operator system or realized production benefit.", "metrics": "26 research challenges; no project-specific funding or operating result established by this source.", "vendors": "DOE, NETL, national laboratories", "takeaway": "Analyst interpretation: Federal subsurface AI programs are a source of open models, datasets and training partners for US operators.", "source": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and", "source_name": "US Department of Energy", "id": 130, "ID": "CTX-REF-130", "Reference_ID": "REF-130", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R130", "Audit_review": {"id": "CTX-REF-130", "company": "US Department of Energy", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "DOE 26-challenge announcement is February 12, 2026. Removed unsupported $5 billion funding and 17-lab numeric claims from this single-source record.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and"}], "residual_gaps": []}}, {"company": "North Sea Transition Authority", "segment": "Government", "hq": "United Kingdom", "title": "NSTA funds AI chatbot for North Sea data", "date": "2025-10", "theme": "Subsurface & exploration", "summary": "The UK regulator received £107,000 from the Regulators' Pioneer Fund for an AI-supported data discovery project. It includes an enhanced geospatial viewer and an AI chatbot to help users navigate NSTA Open Data and the National Data Repository. The regulator holds more than a petabyte of data, and the workstreams were due for assessment in April 2026.", "metrics": "£107K grant; >1 PB of data held; assessment Apr 2026", "vendors": "NSTA, Regulators' Pioneer Fund", "takeaway": "Analyst interpretation: Regulators are building AI access layers over public subsurface data; enablement programs should teach staff to use them.", "source": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/", "source_name": "NSTA", "id": 131, "ID": "CTX-REF-131", "Reference_ID": "REF-131", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R131", "Audit_review": {"id": "CTX-REF-131", "company": "North Sea Transition Authority", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "NSTA source supports a £107,000 grant and planned chatbot/data-viewer work. A grant and planned assessment are not proof of production deployment.", "correction": "", "source_url": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}}, {"company": "Havtil (Norwegian Ocean Industry Authority)", "segment": "Government", "hq": "Norway", "title": "Norway regulator flags AI black-box and data risks", "date": "2024-03", "theme": "Safety & emissions", "summary": "Havtil’s 2024 article discusses AI in Norwegian petroleum operations, including data quality, opaque models and human supervision. It warns that an operator may become overreliant on automated recommendations. The linked article does not substantiate the separate 2025 incident-data pilot previously combined with this record.", "metrics": "", "vendors": "Havtil, DNV, AIICE JIP", "takeaway": "Analyst interpretation: Regulators expect companies to own AI risk, so enablement needs explainability, verification habits and escalation rules.", "source": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/", "source_name": "Havtil", "id": 132, "ID": "CTX-REF-132", "Reference_ID": "REF-132", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "Source_ID": "R132", "Audit_review": {"id": "CTX-REF-132", "company": "Havtil (Norwegian Ocean Industry Authority)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Havtil article supports risk considerations and supervised use. Removed the separate alleged 2025 dropped-object pilot, which the linked 2024 article does not establish.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/"}], "residual_gaps": []}}, {"company": "Journal of Petroleum Technology (SPE)", "segment": "Cautionary", "hq": "United States", "title": "Why O&G AI pilots fail to scale", "date": "2026-08", "theme": "Strategy & org", "summary": "SPE’s The Way Ahead commentary discusses barriers to scaling AI, including data readiness, organizational alignment, workforce skills and cyber risks. Its cited 95% finding concerns a separate cross-industry generative-AI study and measurable P&L impact; it is not evidence that 95% of oil-and-gas AI projects fail.", "metrics": "No oil-and-gas-wide failure percentage established by this commentary.", "vendors": "SPE, ADNOC Offshore", "takeaway": "Analyst interpretation: Scaling is a governance and change-management problem, so budget for data plumbing and trust-building, not just models.", "source": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas", "source_name": "JPT", "id": 134, "ID": "CTX-REF-134", "Reference_ID": "REF-134", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R134", "Audit_review": {"id": "CTX-REF-134", "company": "Journal of Petroleum Technology (SPE)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "SPE commentary cites a cross-industry generative-AI study for 95% without measurable P&L impact. It must not be generalized to an observed 95% failure rate for all oil-and-gas AI.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas"}], "residual_gaps": []}}, {"company": "Enabled Emissions Campaign / Nature study", "segment": "Cautionary", "hq": "United States", "title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "date": "2026-08", "theme": "Safety & emissions", "summary": "Unverified inherited reference: A peer-reviewed Nature study reported by Inside Climate News estimates that AI tools that make extraction faster and cheaper add 0.47-1.8 Gt of CO2 a year. That is 1.2-4.8% of 2024 global energy-related emissions and up to 13 times data-center emissions. The authors call these enabled emissions and want them recognized, measured and governed.", "metrics": "Unverified inherited figures; source access unresolved.", "vendors": "Enabled Emissions Campaign, Microsoft, Google, Chevron, IEA", "takeaway": "Analyst interpretation: Expect ESG scrutiny of O&G AI programs; framing that covers emissions, safety and methane use cases reduces reputational risk.", "source": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/", "source_name": "Inside Climate News", "id": 135, "ID": "CTX-REF-135", "Reference_ID": "REF-135", "Category": "Context / survey / enabling ecosystem", "Evidence": "Source access unresolved", "Source_ID": "R135", "Audit_review": {"id": "CTX-REF-135", "company": "Enabled Emissions Campaign / Nature study", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference.", "correction": "", "source_url": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/"}], "residual_gaps": ["Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference."]}}, {"company": "Microsoft (shareholders and ex-employees)", "segment": "Cautionary", "hq": "United States", "title": "Investors and ex-staff challenge Microsoft's O&G AI work", "date": "2025-12", "theme": "Partnership & investment", "summary": "As You Sow’s December 4, 2025 statement promoted a shareholder proposal asking Microsoft to assess risks from AI products used in oil-and-gas development. This is advocacy and governance context; it does not independently establish operating outcomes or Microsoft’s market share.", "metrics": "No operational benefit, market-share figure or measured emissions result established by this source.", "vendors": "Microsoft, ExxonMobil (XTO), Chevron, Shell, As You Sow, Google", "takeaway": "Analyst interpretation: Vendor reputational pressure can shift tool availability and terms, so avoid single-vendor lock-in in enablement stacks.", "source": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products", "source_name": "As You Sow", "id": 136, "ID": "CTX-REF-136", "Reference_ID": "REF-136", "Category": "Context / survey / enabling ecosystem", "Evidence": "Context source excerpt reviewed", "Source_ID": "R136", "Audit_review": {"id": "CTX-REF-136", "company": "Microsoft (shareholders and ex-employees)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "As You Sow December 4, 2025 statement establishes an investor proposal and advocacy claims. Removed unsupported cloud-market-share and production-uplift figures from this reference.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products"}], "residual_gaps": []}}, {"company": "ConocoPhillips", "segment": "Cautionary", "hq": "United States", "title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "date": "2025-09", "theme": "Strategy & org", "summary": "Unverified inherited reference: ConocoPhillips said it will cut 20-25% of its roughly 13,000-person workforce under a restructuring called Competitive Edge, advised by BCG, citing rising per-barrel costs. The cuts were not attributed to AI, but they came alongside similar reductions at Chevron and bp. EY's surveys find that employee job-security concerns inhibit AI adoption and experimentation.", "metrics": "Unverified inherited figures; source access unresolved.", "vendors": "ConocoPhillips, BCG", "takeaway": "Analyst interpretation: When layoffs coincide with AI rollouts, enablement programs need explicit messaging that AI is for augmentation, or adoption will stall.", "source": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/", "source_name": "Offshore Technology", "id": 137, "ID": "CTX-REF-137", "Reference_ID": "REF-137", "Category": "Context / survey / enabling ecosystem", "Evidence": "Source access unresolved", "Source_ID": "R137", "Audit_review": {"id": "CTX-REF-137", "company": "ConocoPhillips", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained.", "correction": "", "source_url": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/"}], "residual_gaps": ["Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained."]}}], "disclosures": [{"ID": "ORIG-AI-001", "Initiative_ID": "AI-001", "Title": "APOLO well planning", "Date": "2025-11-12", "Finding": "Machine learning guides prospectivity; development extends to spacing, completions and timing.", "Metric": "Training data: >50,000 wells", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Chevron described training on more than 50,000 wells and extending the approach to spacing, completion design and development timing.", "Context": "Well planning must account for interacting geology and development choices. APOLO uses machine learning to make prospectivity assessment more granular than conventional type-curve groupings.", "Assessment": "A larger training set establishes scale of inputs. It does not establish forecast accuracy on new acreage or incremental recovery. Development decisions still require economic and geological judgment."}, {"ID": "ORIG-AI-002", "Initiative_ID": "AI-002", "Title": "Exploration workflows", "Date": "2025-11-12", "Finding": "AI supports seismic interpretation, inventory assessment and prospect decisions.", "Metric": "Not separately disclosed", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The investor-day discussion links AI to interpretation and prospect decisions. High-performance computing and proprietary imaging also support the broader workflow.", "Context": "Chevron places AI within exploration work, where interpreting seismic and evaluating inventory inform which prospects receive further capital.", "Assessment": "AI is one element of a technical workflow. Seismic computing capacity alone cannot demonstrate a better discovery rate, and the source does not isolate AI’s financial contribution."}, {"ID": "ORIG-AI-003", "Initiative_ID": "AI-003", "Title": "Workforce and workflows", "Date": "2025-11-12", "Finding": "Employee tools and workflow redesign support productivity and decisions, including abandonment preparation.", "Metric": "Over half of employees have tool access", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "A well-abandonment example connects an India-based team with Australian operations to prepare information, permit applications and orders overnight.", "Context": "Chevron is combining employee tools with workflow redesign as part of a standardized operating model. More than half of employees had access at the disclosure date.", "Assessment": "The case illustrates cross-team preparation work. Access and a faster example do not establish sustained usage, a measured reduction in headcount, or enterprise-wide savings."}, {"ID": "ORIG-AI-004", "Initiative_ID": "AI-004", "Title": "AI value ambition", "Date": "2025-11-12", "Finding": "Management sees annual AI value beyond existing forecasts by decade-end.", "Metric": "Potential billions annually", "Source_IDs": "S01", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The ambition is potential annual value in the billions by the end of the decade. The disclosure does not provide a project-by-project bridge to that amount.", "Context": "Management described AI as an additional source of future value beyond the company’s existing forecasts.", "Assessment": "Treat this as strategic upside. Adding it to a valuation would require timing, implementation costs, attribution and a definition of value. It should not be presented as realized savings."}, {"ID": "ORIG-AI-005", "Initiative_ID": "AI-005", "Title": "West Texas data-center power", "Date": "2025-11-12 / 2026-06-22", "Finding": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Metric": "~2.67 GW; 20-year PPA; first power targeted 2028", "Source_IDs": "S01; S44", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "June 2026 update: approximately 2.67 GW, developed with Engine No. 1, using GE Vernova and Solar turbines. Chevron expected FID by year-end 2026 and first power in 2028.", "Context": "Project Kilby would supply a co-located Microsoft data center in West Texas with Permian natural gas generation. Energy Forge One, a Chevron subsidiary, signed a 20-year power agreement.", "Assessment": "The signed PPA advances commercial commitment, while investment approval and construction remain separate milestones. The 2028 date supersedes the 2027 first-power target in the November 2025 investor discussion."}, {"ID": "ORIG-AI-006", "Initiative_ID": "AI-006", "Title": "Artificial lift optimization", "Date": "2026-02-18", "Finding": "Expanding AI-enabled lift optimization and advanced analytics beyond pilots.", "Metric": "No AI-only savings disclosed", "Source_IDs": "S02", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The February 2026 call connects second-half 2025 trials, including gas lift, with broader rollout in the first and second quarters of 2026. Named vendors and fleet size are absent.", "Context": "Devon is expanding AI-enabled artificial lift optimization and analytics from trials into wider operations. The business objective is sustaining production and reducing operating intervention.", "Assessment": "Lift optimization concerns how existing wells produce. Devon’s broader $1 billion business-optimization program also includes other actions, so its savings cannot be assigned wholly to this AI initiative."}, {"ID": "ORIG-AI-007", "Initiative_ID": "AI-007", "Title": "D&C monitoring and AI agents", "Date": "2026-02", "Finding": "Models identify unproductive time and practices to improve drilling and completion cycles.", "Metric": "Company reported 85% of broader $1bn program achieved", "Source_IDs": "S03", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The Q4 2025 earnings presentation places this work inside the broader optimization program, which had achieved 85% of its $1 billion target.", "Context": "Devon’s drilling and completion work uses models and agents to identify unproductive time and repeat better operating practices.", "Assessment": "Faster drilling cycles can reduce required capital, but the presentation does not isolate an AI contribution. The relevant comparison would hold geology, well design and service conditions constant."}, {"ID": "ORIG-AI-008", "Initiative_ID": "AI-008", "Title": "Tela agentic assistant", "Date": "2026-01-23", "Finding": "Upstream assistant supports geoscience workflows and autonomous production applications.", "Metric": "More than a dozen customers engaging", "Source_IDs": "S04; S45", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Powered by Lumi, Tela combines language models with domain foundation models. Examples include well-log interpretation, drilling-issue prediction and equipment optimization. Customers can build or integrate agents.", "Context": "Tela is SLB’s upstream agentic assistant, launched in November 2025. It is intended to help technical users perform work within existing energy applications.", "Assessment": "The investor disclosure reports engagement with more than a dozen customers. Engagement is an adoption signal, but the source does not state how many are paying or running production-critical workflows."}, {"ID": "ORIG-AI-009", "Initiative_ID": "AI-009", "Title": "Lumi data and AI platform", "Date": "2026-01-23", "Finding": "Platform adoption connects geoscience, drilling and production workflows.", "Metric": "More than 50 customers", "Source_IDs": "S04", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB reported more than 50 customers. Lumi also underpins Tela and the announced Shell collaboration, so those records represent related layers of the product architecture.", "Context": "Lumi provides a common data and AI platform across geoscience, drilling and production applications. Its role is to connect data with the tools that technical teams already use.", "Assessment": "Customer count supports commercial adoption. It gives no common denominator for seats, data volume or recurring AI spend. SLB’s digital recurring revenue includes products beyond this platform."}, {"ID": "ORIG-AI-010", "Initiative_ID": "AI-010", "Title": "Upstream agentic collaboration", "Date": "2026-01-23", "Finding": "Plans an open data and AI environment across subsurface, well construction and production.", "Metric": "Not disclosed", "Source_IDs": "S05; S46", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The planned open environment uses Lumi to connect subsurface, well construction and production data. Agentic applications would support technical experts and decision makers across these workflows.", "Context": "Shell and SLB signed a December 2025 agreement to develop digital and AI solutions for upstream operations and potentially the wider industry.", "Assessment": "The agreement defines a development direction. It does not establish a deployed Shell-wide system, a production schedule, contract economics or measured efficiency gains."}, {"ID": "ORIG-AI-011", "Initiative_ID": "AI-011", "Title": "AiPSO production optimization", "Date": "2026-01-23", "Finding": "Combines live data, proprietary ML and Cognite Data Fusion to optimize wells and facilities.", "Metric": "Eight fields", "Source_IDs": "S05; S06", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The solution combines Lumi, Cognite Data Fusion and ADNOC machine-learning models. It processes live operating information to help engineers diagnose and optimize production.", "Context": "AiPSO addresses fragmented production information across wells and facilities. ADNOC and SLB initially deployed it across eight fields.", "Assessment": "The company describes analysis moving from days to minutes. That is a workflow claim, with no disclosed field-level production uplift or independently verified cash return in these sources."}, {"ID": "ORIG-AI-012", "Initiative_ID": "AI-012", "Title": "Azerbaijan remote drilling", "Date": "2026-01-23", "Finding": "AI automation and remote operations support offshore drilling decisions.", "Metric": "66% smaller placement crew; >1,500 onboard days saved", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB names Performance Live, DrillOps advisory and Neuro autonomous directional drilling. It reports a 66% smaller placement crew and more than 1,500 personnel-on-board days avoided.", "Context": "bp’s Azerbaijan operations use a combination of remote operations and drilling automation to reduce the need for well-placement staff offshore.", "Assessment": "These results belong to an integrated digital operating model. They cannot be assigned solely to an AI algorithm. Personnel-on-board days measure offshore exposure, rather than direct cash savings."}, {"ID": "ORIG-AI-013", "Initiative_ID": "AI-013", "Title": "Oman autonomous drilling", "Date": "2026-01-23", "Finding": "AI optimization improves rotary steerable drilling.", "Metric": "Up to 50% faster penetration on three wells", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB reports up to 50% faster rate of penetration across three wells using Neuro and PowerDrive. A separate PowerDrive One application reported a 40% improvement.", "Context": "OQ Exploration & Production applied autonomous directional drilling in Oman to improve drilling performance.", "Assessment": "The three-well result is a limited field comparison. Rate of penetration measures drilling speed, and should not be treated as a same-sized reduction in total well cost or full-cycle duration."}, {"ID": "ORIG-AI-014", "Initiative_ID": "AI-014", "Title": "AI seismic processing", "Date": "2026-01-23", "Finding": "Five-year contract incorporates AI workflows into seismic interpretation.", "Metric": "Five-year term", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The scope covers multiple seismic types and onshore, offshore and transition environments. AI supports interpretation within the larger processing program.", "Context": "Kuwait Oil Company awarded a five-year seismic processing and reprocessing contract that incorporates AI workflows.", "Assessment": "A contract award establishes commercial scope. The source does not give a separated AI fee, realized interpretation gain or improved discovery outcome. See AI-066 for a distinct KOC program."}, {"ID": "ORIG-AI-015", "Initiative_ID": "AI-015", "Title": "Data-center equipment capacity", "Date": "2026-01-23", "Finding": "Expanded Shreveport manufacturing for AI-related infrastructure demand.", "Metric": "Facility footprint nearly doubled", "Source_IDs": "S05", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The facility footprint nearly doubled. The source describes equipment capacity expansion but does not disclose a plant-specific investment amount or output capacity.", "Context": "SLB expanded manufacturing in Shreveport, Louisiana, to meet data-center infrastructure demand.", "Assessment": "This is a manufacturing exposure to data-center growth. It should be assessed through demand, utilization and margins, separately from the operating benefits of AI inside oil and gas."}, {"ID": "ORIG-AI-016", "Initiative_ID": "AI-016", "Title": "LOGIX autonomous drilling", "Date": "2025-Q1", "Finding": "Combines physics models and machine learning in autonomous drilling.", "Metric": "No AI-only financial metric", "Source_IDs": "S07", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Halliburton’s Q1 2025 investor material positions autonomous drilling inside a wider digital well-construction system.", "Context": "LOGIX combines physics-based models and machine learning in a drilling automation offering. Its purpose is consistent execution of the well plan.", "Assessment": "The source establishes a commercial capability, without a named trial in this extract. Well conditions, operator procedures and the automation boundary would matter when assessing performance."}, {"ID": "ORIG-AI-017", "Initiative_ID": "AI-017", "Title": "DS365.ai", "Date": "2025-Q1", "Finding": "AI and ML augment subsurface, drilling and production decisions in cloud E&P services.", "Metric": "Not disclosed", "Source_IDs": "S07", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The broader suite runs on iEnergy and uses an open architecture. The investor material presents subscription services that can connect with other applications.", "Context": "DS365.ai adds AI and machine learning to subsurface, drilling and production workflows within Halliburton’s DecisionSpace 365 cloud services.", "Assessment": "The AI module sits within a larger software product family. The disclosure does not separate its users, revenue, deployment cost or economic return from the wider suite."}, {"ID": "ORIG-AI-018", "Initiative_ID": "AI-018", "Title": "Asset performance collaboration", "Date": "2026-07-21", "Finding": "Unifies subsurface and surface information for predictive asset decisions.", "Metric": "Not disclosed", "Source_IDs": "S08", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2026 earnings release describes a unified asset view connecting subsurface and surface information, combining domain science with applied AI for predictive decisions over the production lifecycle.", "Context": "Halliburton and Shape Digital announced a strategic collaboration to improve digital asset performance management.", "Assessment": "The announcement explains intended workflow integration. Named customer deployments, implementation milestones and measured reliability or production outcomes are absent from the reviewed disclosure."}, {"ID": "ORIG-AI-019", "Initiative_ID": "AI-019", "Title": "Data-center orders", "Date": "2026-01-26", "Finding": "Raised order target as AI-related data-center demand accelerated.", "Metric": "Three-year order target $3bn, previously $1.5bn", "Source_IDs": "S09", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Its January 2026 presentation reports $1 billion of data-center power orders in 2025 and a $3 billion target for 2026–2028, raised from $1.5 billion. A roughly 1 GW NovaLT slot reservation is a separate item.", "Context": "Baker Hughes is selling power equipment into data-center demand, including demand associated with AI.", "Assessment": "Orders, reservations and future targets have different commitment levels. These amounts concern equipment and power infrastructure, and should not be labeled AI software revenue."}, {"ID": "ORIG-AI-020", "Initiative_ID": "AI-020", "Title": "Cordant asset strategy", "Date": "2025-04-15", "Finding": "Azure AI supports asset strategy, unstructured data processing and engineering document analysis.", "Metric": "Not disclosed", "Source_IDs": "S10", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The partner webinar discusses generative AI, unstructured text processing and Azure services, including Azure Data Factory in SAP S/4HANA environments.", "Context": "Baker Hughes and EPAM describe AI features for Cordant Asset Strategy, addressing slow data preparation and the specialist knowledge required for reliability engineering.", "Assessment": "This is a partner-described implementation case. The public webinar description does not establish deployment scale, measured time savings or the accuracy of generated asset strategies."}, {"ID": "ORIG-AI-021", "Initiative_ID": "AI-021", "Title": "Enterprise AI partnership renewal", "Date": "2025-06", "Finding": "Renewed multi-year relationship covering co-selling, joint investment and deployment scaling.", "Metric": "Not disclosed", "Source_IDs": "S11; S12", "Evidence": "Primary filing indexed passage reviewed", "Origin": "Original register", "Implementation": "The filings and earnings material describe co-selling, joint investment and scaling. The commercial relationship should be read against the actual renewal terms.", "Context": "Baker Hughes and C3 AI renewed a multi-year enterprise AI relationship intended to support joint sales and deployment.", "Assessment": "A renewal establishes continued commercial cooperation. It does not, by itself, confirm that earlier exclusive oil-and-gas resale rights remain unchanged or that every deployment generated a return."}, {"ID": "ORIG-AI-022", "Initiative_ID": "AI-022", "Title": "Permian and operating-site AI", "Date": "2025-10-31", "Finding": "CEO describes AI deployed downhole and at global operating sites to learn from data and optimize production.", "Metric": "No standalone metric", "Source_IDs": "S13", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q3 2025 call connects learning from operating data with production optimization. The reproduced transcript gives no named model, vendor, field count or controlled result.", "Context": "ExxonMobil’s CEO describes AI operating downhole in the Permian and at company sites around the world.", "Assessment": "This provides management confirmation of deployment but little implementation detail. It cannot support a claim that a particular percentage of Permian improvement came from AI."}, {"ID": "ORIG-AI-023", "Initiative_ID": "AI-023", "Title": "Low-carbon data-center opportunity", "Date": "2025-10-31", "Finding": "Pursuing data-center decarbonization opportunities using capabilities associated with Denbury.", "Metric": "Not disclosed", "Source_IDs": "S13", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q3 2025 call frames these as development opportunities. The evidence reviewed here does not identify a completed project-level investment decision.", "Context": "ExxonMobil discussed lower-carbon power opportunities serving data centers, drawing on capabilities associated with the Denbury acquisition.", "Assessment": "The economic exposure would come through power and carbon-management services. The discussion does not establish AI model deployment inside ExxonMobil or contracted returns from a named campus."}, {"ID": "ORIG-AI-024", "Initiative_ID": "AI-024", "Title": "ERP and logistics AI", "Date": "2026-01-30", "Finding": "Standardizing enterprise systems facilitates AI applications such as marine logistics planning.", "Metric": "No AI-only benefit disclosed", "Source_IDs": "S14", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The Q4 2025 call gives marine logistics planning as an example of AI-enabled work supported by the ERP transformation.", "Context": "ExxonMobil’s enterprise-system standardization creates a more consistent data and process environment for AI applications.", "Assessment": "The ERP program is an enabling investment with broader benefits. System simplification counts and enterprise savings should not be reclassified as a measured AI return."}, {"ID": "ORIG-AI-025", "Initiative_ID": "AI-025", "Title": "Technology and AI cost reduction", "Date": "2025-03", "Finding": "AI is one component of structural simplification and cost reduction.", "Metric": "Total structural reduction target $5–7bn, 2022–2028", "Source_IDs": "S15", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Capital Markets Day 2025 presents a company-wide structural reduction target of $5–7 billion over 2022–2028. Portfolio actions and other non-AI measures also contribute.", "Context": "Shell includes technology and AI within its structural simplification and cost-reduction plans.", "Assessment": "This is evidence that AI participates in a cost program. The disclosure provides no separate AI savings bridge, making the full corporate target unsuitable as an AI benefit estimate."}, {"ID": "ORIG-AI-026", "Initiative_ID": "AI-026", "Title": "Immersion cooling fluids", "Date": "2025-03", "Finding": "GTL-derived fluids support AI data-center cooling.", "Metric": "Up to 33% claimed improvement; footnote describes cost savings", "Source_IDs": "S15", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Its investor presentation cites an improvement of up to 33%, while the associated footnote describes cost savings. The wording does not support treating that figure as a universal efficiency gain.", "Context": "Shell markets gas-to-liquids-derived immersion cooling fluids for data-center applications, including AI computing.", "Assessment": "The commercial route is specialty fluids. Assess the exact test conditions and cost denominator before using the percentage in a customer business case."}, {"ID": "ORIG-AI-027", "Initiative_ID": "AI-027", "Title": "ENI AI portfolio", "Date": "2026", "Finding": "Disclosed expanding use-case portfolio and AI agents.", "Metric": "Around 300 use cases; +40% vs 2024; about 20 agents", "Source_IDs": "S16", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The presentation places AI alongside computing and industrial digitalization initiatives. It does not provide a complete application-level deployment list.", "Context": "Eni’s 2026 Capital Markets Update describes an expanding portfolio of approximately 300 AI use cases, up about 40% from 2024, plus around 20 agents.", "Assessment": "Use-case counts measure portfolio activity. Nearby digital-plant and drilling-automation figures cannot automatically be attributed to all AI use cases, and agent count does not measure usage or value."}, {"ID": "ORIG-AI-028", "Initiative_ID": "AI-028", "Title": "Ferrera Erbognone campus", "Date": "2025-07-11", "Finding": "Heads of terms for a joint venture developing an AI data-center campus.", "Metric": "500 MW IT capacity", "Source_IDs": "S17", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2025 announcement combines Khazna’s data-center expertise with Eni’s energy capabilities. A proposed gas-fired Blue Power plant, designed for CO₂ capture, would support the campus.", "Context": "Eni and Khazna agreed heads of terms for a joint venture to develop a 500 MW IT-capacity campus at Ferrera Erbognone, near Milan.", "Assessment": "IT capacity measures computing load, which differs from generation capacity. Heads of terms set development principles. Budget, equity shares, commissioning date and FID are not established by this source."}, {"ID": "ORIG-AI-029", "Initiative_ID": "AI-029", "Title": "Research assistant and innovation lab", "Date": "2025-06-12", "Finding": "Joint lab to build researcher assistance and industrial decision-support tools.", "Metric": "Assistant intended for 1,000 researchers", "Source_IDs": "S18", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Planned work includes an assistant for 1,000 researchers, industrial decision support, emissions and energy-efficiency applications, and customer energy advice.", "Context": "TotalEnergies and Mistral AI announced a joint innovation laboratory in June 2025 to build AI tools with company-specific expertise.", "Assessment": "The researcher count describes intended reach. The source gives no achieved adoption rate or assistant-level economics. The lab’s planned tools should be distinguished from TotalEnergies’ existing AI estate."}, {"ID": "ORIG-AI-030", "Initiative_ID": "AI-030", "Title": "Digital Factory AI portfolio", "Date": "2025-06-12", "Finding": "Company reports AI in earth sciences, equipment diagnostics and predictive maintenance.", "Metric": "60 of >100 digital solutions use ML through generative AI", "Source_IDs": "S18", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The applications span machine learning through generative AI, including earth sciences, equipment diagnostics and predictive maintenance.", "Context": "TotalEnergies describes an existing Digital Factory with 300 experts and more than 100 digital solutions, of which 60 use AI.", "Assessment": "The portfolio predates the Mistral lab. Counting all 60 as Mistral deployments would overstate the partnership’s scope, and the disclosure does not isolate the portfolio’s financial return."}, {"ID": "ORIG-AI-031", "Initiative_ID": "AI-031", "Title": "Multi-agent platform", "Date": "2025-10-15", "Finding": "Piloting a platform that enables teams to construct coordinated AI agents.", "Metric": "22 agents across three pilot teams; >50 people supervising", "Source_IDs": "S19", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "In Digitalization and Services, three teams used 22 agents for more than four months. More than 50 people subsequently worked together supervising agent teams.", "Context": "Repsol and Accenture developed a pilot platform for coordinated AI agents, supported by dedicated knowledge bases and a professional collaboration interface.", "Assessment": "The pilot demonstrates a defined organizational setting. Its 22 agents are distinct from the 22 scaled GenAI use cases in AI-032. No platform-specific financial return is disclosed."}, {"ID": "ORIG-AI-032", "Initiative_ID": "AI-032", "Title": "GenAI competence center", "Date": "2025-10-15", "Finding": "Scaling use cases and employee capability across the company.", "Metric": ">60 production use cases; 22 scaled; >5,000 employees trained", "Source_IDs": "S19", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The October 2025 release reports more than 60 use cases in production, 22 scaled and more than 5,000 employees trained in GenAI or advanced prompting.", "Context": "Repsol’s Generative AI Competence Center supports production use cases and workforce capability across the company.", "Assessment": "These are different measures of maturity. Training does not establish active use, while the release does not enumerate all production cases or provide comparable application-level economics."}, {"ID": "ORIG-AI-033", "Initiative_ID": "AI-033", "Title": "Industrial AI portfolio value", "Date": "2026-01-07", "Finding": "Reports value creation and savings for Equinor and partners.", "Metric": "$130m in 2025; >$330m since 2020", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Management says traditional machine learning on operating data provides much of the industrial value. AI-034 to AI-036 describe contributing maintenance, planning and seismic examples.", "Context": "Equinor reports $130 million of AI-related value and savings for itself and partners in 2025, and more than $330 million since 2020.", "Assessment": "These are company-defined benefits, including partners. The component examples overlap the portfolio total, so adding them again would double-count value."}, {"ID": "ORIG-AI-034", "Initiative_ID": "AI-034", "Title": "Predictive maintenance", "Date": "2026-01-07", "Finding": "Sensor models identify equipment failure and maintenance needs.", "Metric": ">700 machines; $120m value since 2020", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Models predict failures and maintenance needs. Equinor attributes $120 million of cumulative value to the application since 2020.", "Context": "Predictive maintenance monitors more than 700 rotating machines using 24,000 sensors across Equinor facilities.", "Assessment": "Earlier detection may avoid shutdowns, but this figure is cumulative and belongs within the broader AI value total. It is not annual recurring savings or solely Equinor net income."}, {"ID": "ORIG-AI-035", "Initiative_ID": "AI-035", "Title": "Well and field planning", "Date": "2026-01-07", "Finding": "AI generated an alternative Johan Sverdrup phase 3 development solution.", "Metric": "$12m partnership savings", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "For Johan Sverdrup phase 3, management says an AI-generated alternative saved the partnership $12 million.", "Context": "Equinor uses AI to generate alternative well and field-development plans for experts to assess.", "Assessment": "The example shows value from a changed development choice. It does not establish that all planning decisions are automated, or that the saving repeats annually across the portfolio."}, {"ID": "ORIG-AI-036", "Initiative_ID": "AI-036", "Title": "Seismic interpretation", "Date": "2026-01-07", "Finding": "AI expands seismic interpretation capacity.", "Metric": "10x capacity; 2m km² interpreted in 2025", "Source_IDs": "S20", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The company reports tenfold interpretation capacity and two million square kilometres interpreted in 2025.", "Context": "Equinor’s AI interpretation tools increase the volume of seismic information that geoscientists can evaluate.", "Assessment": "Interpretation throughput is an intermediate operating measure. Discovery success and commercial reserves depend on additional geological and economic factors that the cited metric does not capture."}, {"ID": "ORIG-AI-037", "Initiative_ID": "AI-037", "Title": "AI-enabled technology value", "Date": "2026-03-10", "Finding": "CEO attributes more than half of reported technology value to AI.", "Metric": ">$2.65bn implied AI-enabled share of $5.3bn total", "Source_IDs": "S21", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Multiplying the disclosed total by 50% implies an AI-enabled share above $2.65 billion. That threshold is calculated here, rather than quoted as a separate company figure.", "Context": "Aramco’s CEO said more than half of $5.3 billion in 2025 Technology Realized Value came from AI.", "Assessment": "TRV is a company-defined technology measure. It should not be treated as audited incremental net income, and comparability with another operator’s AI savings requires matching definitions."}, {"ID": "ORIG-AI-038", "Initiative_ID": "AI-038", "Title": "AI deployment portfolio", "Date": "2025-11-24", "Finding": "Models address downtime, maintenance and production optimization.", "Metric": "442 identified use cases; >200 solutions deployed; >100 in development", "Source_IDs": "S22", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The company identified 442 use cases, with more than 200 solutions deployed and over 100 in development. It also described GPU computing and AI workforce training.", "Context": "Aramco’s late-2025 portfolio spans facility downtime, maintenance and production optimization, using large volumes of operating data.", "Assessment": "Identified, developing and deployed counts refer to different stages. They are a dated portfolio snapshot, rather than an independently verified count of economically successful applications."}, {"ID": "ORIG-AI-039", "Initiative_ID": "AI-039", "Title": "HUMAIN minority investment plan", "Date": "2025-11-24", "Finding": "Proposed stake combines AI capabilities, data centers, models and cloud infrastructure.", "Metric": "Significant minority stake; no value given", "Source_IDs": "S22", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "HUMAIN’s intended scope includes data centers, cloud infrastructure, advanced models such as ALLAM, and AI applications.", "Context": "Aramco proposed a significant minority investment in PIF-backed HUMAIN to combine AI assets, capabilities and talent.", "Assessment": "The reviewed source establishes a proposed sovereign-capital transaction. It does not confirm closing, valuation or ownership percentage, and the activity differs from conventional buyout-fund investing."}, {"ID": "ORIG-AI-040", "Initiative_ID": "AI-040", "Title": "ENERGYai", "Date": "2025-01-16", "Finding": "Agentic AI combines company data and an LLM for seismic interpretation, reservoir monitoring and production insights.", "Metric": "70% accuracy improvement in selected seismic aspects; 70bn-parameter LLM", "Source_IDs": "S23", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "A 90-day proof of concept used data from over 15% of ADNOC wells. The January 2025 report cites a 70% accuracy improvement in selected seismic aspects and planned five-agent test deployment in H1 2025.", "Context": "ENERGYai combines a 70-billion-parameter language model with ADNOC knowledge and proprietary data to support upstream technical work.", "Assessment": "The accuracy claim concerns selected trial tasks. The next-stage date was a target at that disclosure, and the source does not prove completion or a general 70% productivity gain."}, {"ID": "ORIG-AI-041", "Initiative_ID": "AI-041", "Title": "Lumina, Copilot and ChatGPT Enterprise", "Date": "2026-05-27", "Finding": "Enterprise agents and employee tools support business workflows and prototype testing.", "Metric": ">50 Lumina agents", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Copilot serves personal productivity in Office 365. ChatGPT Enterprise provides an analytics and prototype environment before custom agents enter Lumina.", "Context": "Woodside’s Lumina platform supports more than 50 agents across business functions, alongside Microsoft Copilot and ChatGPT Enterprise.", "Assessment": "The architecture distinguishes employee assistance from integrated workflows. Agent count alone does not reveal frequency of use, human review effort or realized financial benefit."}, {"ID": "ORIG-AI-042", "Initiative_ID": "AI-042", "Title": "Maint Intel", "Date": "2026-05-27", "Finding": "Maintenance scheduling model compares historical performance with reliability targets.", "Metric": "Potential 15% annual maintenance-hour reduction at Angel over five years", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Woodside launched the tool at the North West Shelf Project. At Angel, it identified potential to reduce annual maintenance hours by 15% over the next five years.", "Context": "Maint Intel recommends maintenance schedules by comparing historical equipment performance with reliability targets.", "Assessment": "This is identified potential. The source does not report those hours as already avoided. Reliability outcomes and implementation costs would determine the value of changing maintenance intervals."}, {"ID": "ORIG-AI-043", "Initiative_ID": "AI-043", "Title": "Startup Advisor", "Date": "2026-05-27", "Finding": "AI and digital procedures support LNG startup decisions using historical and live performance.", "Metric": "Not disclosed", "Source_IDs": "S24", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "It combines AI, automation and digital procedures, comparing real-time plant performance with historical startup experience. Woodside says operators have used it in live operations.", "Context": "Woodside’s Startup Advisor supports operators during complex LNG plant startups.", "Assessment": "The disclosure supports operational use. It does not quantify startup duration, failure reduction or cash savings, and it describes support for operators rather than unattended plant control."}, {"ID": "ORIG-AI-044", "Initiative_ID": "AI-044", "Title": "AI-enhanced subsurface development", "Date": "2026-05-06", "Finding": "Plans combine technical data with AI to improve development and performance.", "Metric": "No AI-only metric", "Source_IDs": "S25", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The Q1 2026 earnings call presents an expansion of AI-enabled technical work, without isolating a named deployment’s economics.", "Context": "Occidental describes plans to combine technical subsurface data with AI to improve development decisions and operating performance.", "Assessment": "Historical well results and reserves replacement provide business context, but the call does not attribute them solely to AI. The disclosed ambition exceeds the implementation detail available here."}, {"ID": "ORIG-AI-045", "Initiative_ID": "AI-045", "Title": "Lift optimization and downtime", "Date": "2026-08-10", "Finding": "Early AI and automation deployment targets daily lift optimization and reduced downtime.", "Metric": "No realized financial metric", "Source_IDs": "S26", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The August 2026 call explores the importance of improving base production and decline performance. It does not report a measured AI-specific decline-rate improvement.", "Context": "Diamondback discusses early AI and automation applications for daily artificial lift optimization and reduced downtime.", "Assessment": "A 1% decline-rate discussion is illustrative. Using it as an achieved AI result would convert a scenario into evidence. Named field scope and realized savings remain undisclosed."}, {"ID": "ORIG-AI-046", "Initiative_ID": "AI-046", "Title": "Agora edge AI", "Date": "2026-07-24", "Finding": "MoU combines low-power compute and SLB edge solutions for well, facility and production decisions.", "Metric": "Not disclosed", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The proposal pairs Qualcomm’s low-power processing with Agora edge AI and industrial connectivity.", "Context": "SLB and Qualcomm signed an MoU focused on AI at wells and production facilities.", "Assessment": "Edge processing places computation near equipment. The agreement does not establish completed joint deployments, latency improvements or autonomous control authority. Its operating value depends on a subsequent implemented use case."}, {"ID": "ORIG-AI-047", "Initiative_ID": "AI-047", "Title": "Digital Marketplace for Tela", "Date": "2026-07-24", "Finding": "Governed marketplace distributes AI capabilities from SLB and ecosystem developers.", "Metric": "Not disclosed", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Offerings connect with Tela and require security, compatibility and interoperability certification before listing.", "Context": "SLB’s Digital Marketplace distributes agents, domain models and supporting tools through a governed channel.", "Assessment": "A common distribution channel may reduce integration effort. Availability alone gives no evidence of paid adoption, deployed agents or incremental customer value. This is a product layer related to AI-008 and AI-009."}, {"ID": "ORIG-AI-048", "Initiative_ID": "AI-048", "Title": "Intelligent Production Studio", "Date": "2026-07-24", "Finding": "Demonstration environment combines AI, edge intelligence, twins and autonomous well-pad workflows.", "Metric": "Autonomous pad with real three-phase flow", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "SLB describes an autonomous well pad operating with real three-phase flow in a demonstration environment.", "Context": "Intelligent Production Studio demonstrates production workflows using AI, edge intelligence and digital twins.", "Assessment": "A physical demonstration is stronger than a software concept but differs from a customer field rollout. Commercial reliability, intervention rates and economics require evidence beyond the demonstration setting."}, {"ID": "ORIG-AI-049", "Initiative_ID": "AI-049", "Title": "Atlas AI partnership expansion", "Date": "2025-09-16", "Finding": "Expanded agent adoption aims to improve efficiency and workforce productivity.", "Metric": "Partner reports >70% less engineering time for root-cause analysis", "Source_IDs": "S28", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The September 2025 announcement names document processing, equipment registration, barrier management and production chemistry. Root-cause agents retrieve information, build cause maps and suggest analysis steps.", "Context": "Aker BP expanded Cognite Atlas AI adoption, using pre-built agents and a low-code workbench connected to Cognite Data Fusion.", "Assessment": "Cognite reports over 70% less engineering time for root-cause analysis. This is a partner-reported task result. The release does not disclose sample size, total deployment cost or a company-wide cash benefit."}, {"ID": "ORIG-AI-050", "Initiative_ID": "AI-050", "Title": "TIRAI subsurface assistant", "Date": "2025-11-03", "Finding": "Chatbot supports subsurface exploration and retrieval of reports, maps, seismic and well data.", "Metric": "Released October 18, 2025", "Source_IDs": "S29", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Released on 18 October 2025, the GenAI chatbot appears in SmartViewer, SmartData Catalog and MBR/MBR+ pages for MBR2025 blocks. ElasticDocs links users to reports, maps, seismic and well markers.", "Context": "PETRONAS launched TIRAI to help myPROdata users explore subsurface information during block evaluation.", "Assessment": "The disclosed function is data discovery and retrieval. The source does not establish autonomous interpretation, faster drilling decisions or measured user productivity."}, {"ID": "ORIG-AI-051", "Initiative_ID": "AI-051", "Title": "Enverus acquisition and AI scaling", "Date": "2025-08-06 / 2026-04-23", "Finding": "Buyout announcement links the platform’s growth to GenAI, asset decisions and energy-market analytics.", "Metric": "Financial terms not disclosed", "Source_IDs": "S30; S47", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Enverus reported 8,000 customers across 50 countries. Blackstone’s April 2026 results subsequently include Enverus in first-quarter capital deployed, updating the original agreement-only status.", "Context": "Blackstone agreed in August 2025 to acquire Enverus from Hellman & Friedman and Genstar. The investment thesis connects energy intelligence with AI and rising power demand.", "Assessment": "The aggregate private-equity deployment figure is not the Enverus purchase price. Acquisition terms and an AI-only valuation remain undisclosed in these sources. See AI-065 for the later product launch."}, {"ID": "ORIG-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum Energy Cloud", "Date": "Undated; accessed 2026-09-26", "Finding": "Microsoft/Databricks-based foundation unifies portfolio financial, operating and market data for AI.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Quantum names Microsoft and Databricks as platform partners and describes a shared foundation for AI.", "Context": "Quantum Energy Cloud unifies financial, operational and market data across the sponsor’s portfolio.", "Assessment": "A shared data foundation can make portfolio analytics repeatable. The undated firm page does not establish when deployment began, how many companies participate, or whether a new initiative launched within this review window."}, {"ID": "ORIG-AI-053", "Initiative_ID": "AI-053", "Title": "Quantum IQ", "Date": "Undated; accessed 2026-09-26", "Finding": "Proprietary GenAI platform answers questions, automates analysis and triggers workflows.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Quantum describes teams using QIQ to answer questions, fine-tune models and standardize practices.", "Context": "Quantum IQ is the sponsor’s proprietary generative AI platform for shared knowledge, analysis and workflow assistance.", "Assessment": "This is a firm-described operating capability. Public disclosure does not supply usage, cost or independently tested output quality. The source is undated, so it is included as current context with a timing exception."}, {"ID": "ORIG-AI-054", "Initiative_ID": "AI-054", "Title": "Asset and capital planning", "Date": "Undated; accessed 2026-09-26", "Finding": "Models support single-well economics, scenarios, acreage planning and well/pad combinations.", "Metric": "Millions of combinations evaluated", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Single-well economics, geospatial analysis and constraint-based optimization evaluate millions of well and pad combinations.", "Context": "Quantum describes scenario models for asset acquisition and capital deployment.", "Assessment": "Optimization explores feasible development choices under assumptions. It should not be interpreted as delegated investment authority, proven investment alpha or a substitute for reserves and land diligence. The public description is undated."}, {"ID": "ORIG-AI-055", "Initiative_ID": "AI-055", "Title": "Production optimization and surveillance", "Date": "Undated; accessed 2026-09-26", "Finding": "Models recommend completion designs, optimize lift and predict production anomalies and downtime.", "Metric": "Not disclosed", "Source_IDs": "S31", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The firm describes completion recommendations, lift optimization, anomaly detection and downtime prediction.", "Context": "Quantum applies models to production optimization and surveillance across the investment and operating process.", "Assessment": "These functions could inform portfolio-company operations, but the page names no specific company, field or quantified outcome. The source is undated and should not be treated as a newly announced deployment in this window."}, {"ID": "ORIG-AI-056", "Initiative_ID": "AI-056", "Title": "AI logistics acquisition", "Date": "2026-06-25", "Finding": "Acquired VOR intelligence layer to connect planning, logistics and execution; agentic response is part of roadmap.", "Metric": "Terms not disclosed", "Source_IDs": "S32; S33", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Streamba’s Visibility, Orchestration and Response layer combines with DaWinci to cover personnel and cargo. The roadmap includes agents that respond to disruptions, reassign resources and revise plans under human oversight.", "Context": "Francisco Partners-backed Quorum acquired Streamba in June 2026 to connect logistics planning with field execution.", "Assessment": "The acquisition is disclosed as completed. Fully autonomous response remains a product-development ambition. Deal terms, integration timing and customer-level savings are not disclosed."}, {"ID": "ORIG-AI-057", "Initiative_ID": "AI-057", "Title": "Lease provision extraction", "Date": "2025-09-25", "Finding": "AI embedded in land management accelerates lease research and provision processing.", "Metric": "Vendor claim: hours to seconds", "Source_IDs": "S34; S36", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The function processes lease information inside the existing system. PakEnergy claims work previously taking hours can take seconds and positions this as the first release in a wider AI roadmap.", "Context": "Charlesbank-backed PakEnergy embedded AI lease-provision extraction in its Land Management product in September 2025.", "Assessment": "The time claim comes from the vendor, without a disclosed accuracy benchmark or customer sample. Faster extraction does not remove the need to verify obligations in the underlying lease."}, {"ID": "ORIG-AI-058", "Initiative_ID": "AI-058", "Title": "PakCAPTURE ticket and receipt processing", "Date": "2026-05-20", "Finding": "Photo capture and AI extraction automate transportation paperwork and back-office entry.", "Metric": "No quantified return disclosed", "Source_IDs": "S35; S36", "Evidence": "Company-issued release text reviewed", "Origin": "Original register", "Implementation": "Drivers or staff capture document images so information can enter transportation and back-office workflows with less manual re-entry. The May 2026 release adds another product to PakEnergy’s AI rollout.", "Context": "PakCAPTURE brings AI document extraction into PakEnergy Transportation for tickets and receipts.", "Assessment": "The launch establishes availability. The reviewed announcement does not provide quantified payback, extraction accuracy or a measured reduction in billing disputes."}, {"ID": "ORIG-AI-059", "Initiative_ID": "AI-059", "Title": "Bosque data-center campus", "Date": "2025-07-30", "Finding": "Campus adjacent to gas generation uses a dedicated long-term power contract.", "Metric": "190 MW campus; initial 144 MW IT; investment approaching $4bn", "Source_IDs": "S37", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "CyrusOne and ECP are developing a 190 MW campus with initial IT capacity of 144 MW, next to Calpine gas generation. July 2025 disclosure put investment near $4 billion and targeted Q4 2026 operation.", "Context": "ECP and KKR announced the Bosque campus as the first project under their broader $50 billion infrastructure partnership.", "Assessment": "Construction was underway at announcement. The $50 billion figure refers to the broader partnership, while $4 billion refers to this project. Campus and IT MW describe different capacity boundaries."}, {"ID": "ORIG-AI-060", "Initiative_ID": "AI-060", "Title": "Refining digital / AI program", "Date": "2024-10-29", "Finding": "Management reported progress applying Palantir and Infosys to refining workflows.", "Metric": "No AI-only benefit disclosed", "Source_IDs": "S38; S39", "Evidence": "Third-party-hosted transcript reviewed", "Origin": "Original register", "Implementation": "The call discusses digital improvement. A separate Palantir announcement on 9 September 2024 establishes the AI relationship but falls just outside the review window.", "Context": "bp’s Q3 2024 call described progress applying Palantir and Infosys to refining workflows.", "Assessment": "The AI linkage uses corroborating partner evidence rather than explicit AI detail in every call remark. This distinction prevents a general digital program from becoming an unsupported claim of specific AI savings."}, {"ID": "ORIG-AI-061", "Initiative_ID": "AI-061", "Title": "Productivity and decision-support pilots", "Date": "2026-02-18", "Finding": "Describes governed AI introduction beginning with productivity and decision-support use cases.", "Metric": "Not disclosed", "Source_IDs": "S40", "Evidence": "Original filing reverified 17 October 2026; program-level pilots only", "Origin": "Original register", "Implementation": "The 2025 Management’s Discussion and Analysis supplies the program-level disclosure, without naming individual applications or external vendors.", "Context": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "Assessment": "This supports the existence of a pilot program. More detailed implementation, operating scope and value claims would require additional disclosure. A generic program should not be expanded into invented use cases."}, {"ID": "ORIG-AI-062", "Initiative_ID": "AI-062", "Title": "San Marcos gas supply", "Date": "2025-02-10", "Finding": "Long-term gas agreement for an AI-focused campus via Oasis Pipeline.", "Metric": "Up to 450,000 MMBtu/day", "Source_IDs": "S41", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The February 2025 announcement specifies up to 450,000 MMBtu/day, sufficient for roughly 1.2 GW of direct power, for at least ten years. Supply depends on CloudBurst reaching customer FID.", "Context": "Energy Transfer signed a long-term agreement to supply CloudBurst’s proposed AI-focused campus near San Marcos through Oasis Pipeline.", "Assessment": "Q3 2026 operation was an original conditional target, not a verified operating date in this review. Contract capacity is not actual gas consumption, and the source gives no realized revenue."}, {"ID": "ORIG-AI-063", "Initiative_ID": "AI-063", "Title": "Homer City gas supply", "Date": "2025-07-15", "Finding": "EQT selected in principle as exclusive gas sourcing and supply partner for AI/HPC campus power.", "Metric": "4.4 GW gas facility; power targeted 2027", "Source_IDs": "S42", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The July 2025 release names Knighthead as financing sponsor, GE Vernova as turbine supplier and Kiewit as construction lead. It targeted first power in 2027.", "Context": "EQT reached an agreement in principle to source and supply gas for the 4.4 GW Homer City power facility serving an AI/HPC campus.", "Assessment": "An agreement in principle differs from completed definitive supply terms. The 4.4 GW figure concerns generation capacity. Gas volumes, pricing and realized supply economics remain undisclosed here."}, {"ID": "ORIG-AI-064", "Initiative_ID": "AI-064", "Title": "Socrates turnkey power", "Date": "2026-08-04", "Finding": "Gas supply, pipeline and generation solution serves data centers.", "Metric": "556 MW ISO; Plato South in service, North targeted Q4 2026", "Source_IDs": "S43", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The August 2026 presentation describes 556 MW at ISO conditions, with Plato South in service and Plato North targeted for the fourth quarter of 2026.", "Context": "Williams’ Socrates program combines gas supply, pipelines and power generation for data-center customers.", "Assessment": "Partial service is a more advanced milestone than a development announcement. ISO rating is a standardized capacity measure, and the presentation does not identify the share of customer load devoted to AI."}, {"ID": "ORIG-AI-065", "Initiative_ID": "AI-065", "Title": "Enverus ONE workflow platform", "Date": "2026-04-07", "Finding": "Enverus ONE launched in April 2026 as an AI platform for energy workflows. It pairs frontier models with Enverus’ proprietary energy model, Astra.", "Metric": "Four named launch workflows", "Source_IDs": "S48", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Four launch workflows cover AFE evaluation, current-production valuation, project siting and secure document Q&A. The vendor describes private tenancy, role-based access and audit trails.", "Context": "Enverus ONE launched in April 2026 as an AI platform for energy workflows. It pairs frontier models with Enverus’ proprietary energy model, Astra.", "Assessment": "This is a named product launch within a PE-backed energy platform. Availability and vendor time-saving claims do not establish adoption, accuracy or independent customer returns."}, {"ID": "ORIG-AI-066", "Initiative_ID": "AI-066", "Title": "Jurassic gas AI collaborations", "Date": "2026-07-24", "Finding": "KOC awarded four Innovation Factori collaborations for Jurassic gas.", "Metric": "Four collaborations, including a three-month academy", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The scope covers rig scheduling, GenAI production analytics, log quality and reconstruction, plus a three-month AI academy.", "Context": "KOC awarded four Innovation Factori collaborations for Jurassic gas.", "Assessment": "This is a distinct program from KOC’s seismic-processing contract in AI-014. The award establishes intended work, while field scale, completion milestones and economics remain unreported in the cited results release."}, {"ID": "ORIG-AI-067", "Initiative_ID": "AI-067", "Title": "Offshore production AI pilot", "Date": "2026-07-24", "Finding": "PTTEP contracted SLB to pilot digital production tools at a domestic offshore asset.", "Metric": "One domestic offshore asset", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The pilot combines Cognite Data Fusion, OptiFlow and Tela advisory agents.", "Context": "PTTEP contracted SLB to pilot digital production tools at a domestic offshore asset.", "Assessment": "The defined pilot connects data preparation with production assurance and well optimization. The source gives no named asset, pilot result or measured uplift. A pilot award should not be presented as an enterprise rollout."}, {"ID": "ORIG-AI-068", "Initiative_ID": "AI-068", "Title": "Offshore-terminal process optimization", "Date": "2026-07-24", "Finding": "SLB worked with CNOOC subsidiary China France Bohai Geoservices on offshore-terminal process optimization.", "Metric": "No quantified financial outcome", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "Operational data and Symmetry digital twins connect with process-control systems for real-time economic optimization.", "Context": "SLB worked with CNOOC subsidiary China France Bohai Geoservices on offshore-terminal process optimization.", "Assessment": "The disclosure describes a closed-loop implementation. It does not quantify profit, energy savings or the AI contribution relative to the broader control system. The operating outcome remains a vendor-reported qualitative claim."}, {"ID": "ORIG-AI-069", "Initiative_ID": "AI-069", "Title": "Frontier seismic AI interpretation", "Date": "2026-07-24", "Finding": "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic-processing project.", "Metric": "Nearly 6,000 km²; two-year project", "Source_IDs": "S27", "Evidence": "Primary source text reviewed", "Origin": "Original register", "Implementation": "The nearly 6,000 km² program combines advanced imaging with AI-enabled horizon interpretation to create a first 3D image.", "Context": "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic-processing project.", "Assessment": "The AI component sits inside a larger imaging workflow. A processing award does not establish a discovery or commercial reserves, and the source does not isolate the AI contract value."}, {"ID": "REF-000", "Initiative_ID": "AI-037", "Title": "AI drives over half of $5.3B technology value", "Date": "2026-03", "Finding": "Aramco reported $5.3B of Technology Realized Value (TRV) in 2025 from AI, digital and other technology, with CEO Amin Nasser saying over half was enabled by AI. Cumulative TRV since 2023 is $11.3B, and the company guided to an additional $3-5B in 2026.", "Metric": "$5.3B TRV 2025 (>50% AI-enabled); $11.3B cumulative since 2023; $3-5B additional TRV expected 2026", "Source_IDs": "S21", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Publishing a single audited-style 'realized value' KPI for AI/tech turns AI from a cost center into a board-level P&L story; builders should instrument value capture from day one.", "Partners": "", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-001", "Initiative_ID": "CTX-REF-001", "Title": "Aramco launches Metabrain 250B-parameter industrial LLM", "Date": "2024-03", "Finding": "Aramco unveiled METABRAIN, a 250-billion-parameter generative AI model trained on about 7 trillion tokens of public and internal data spanning ~90 years of company history. It targets drilling-plan analysis, well recommendations and refined-product price/market forecasting, with a 1-trillion-parameter version planned.", "Metric": "250B parameters; 7T training tokens; ~90 years of data; 80+ AI professionals", "Source_IDs": "R001", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Proprietary operational history is the moat; NOCs with deep archives are building domain LLMs rather than relying only on general-purpose models.", "Partners": "Aramco Digital", "Original_date": "2024-03", "Related_IDs": []}, {"ID": "REF-002", "Initiative_ID": "AI-039", "Title": "Aramco takes significant minority stake in HUMAIN", "Date": "2025-10", "Finding": "PIF and Aramco signed a non-binding term sheet for Aramco to acquire a significant minority stake in HUMAIN, combining PIF and Aramco AI programs and assets under the PIF company. Nasser said it would strengthen Aramco's leadership in industrial AI applications.", "Metric": "Non-binding term sheet; PIF retains majority", "Source_IDs": "R002", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Majors are consolidating AI compute, models and talent into shared vehicles rather than building everything in-house; expect sovereign AI stacks to become preferred suppliers in the Gulf.", "Partners": "HUMAIN, PIF, ALLaM", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-003", "Initiative_ID": "AI-070", "Title": "Aramco-Microsoft MoU on industrial AI and digital talent", "Date": "2026-02", "Finding": "Aramco and Microsoft signed a non-binding MoU to accelerate industrial AI adoption on Azure with sovereign controls for Saudi data residency. Focus areas include operational efficiency, an industry alliance framework, industrial AI IP co-innovation, and workforce programs in AI engineering, cybersecurity, data governance and product management.", "Metric": "No quantified metric supplied", "Source_IDs": "R003", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Sovereignty/data residency is a gating requirement for Gulf NOCs; enablement offerings must pair tooling with local-cloud deployment and talent programs.", "Partners": "Microsoft, Azure", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-004", "Initiative_ID": "AI-038", "Title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "Date": "2025-06", "Finding": "Aramco says it processes ~10 billion data points daily and is using AI to reprocess old seismic data, optimize autonomous drilling and detect pipeline corrosion. It is training 6,000+ AI developers with Imperial College, Caltech and KAUST, and partnered with Groq on a regional AI inferencing center.", "Metric": "10B data points/day; 1,500 PB storage; 6,000+ AI developers being trained; 5 WEF Lighthouse awards", "Source_IDs": "R004", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Pair large-scale technical upskilling with high-value 'old data, new answers' use cases like seismic reprocessing to show quick ROI.", "Partners": "Groq, Qualcomm, Imperial College, Caltech, KAUST", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-005", "Initiative_ID": "AI-071", "Title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "Date": "2025-11", "Finding": "ADNOC deployed generative AI company-wide in November 2023; by ADIPEC 2025 more than 40,000 employees had completed AI training, with utilization above 90% and over 70,000 productivity hours gained per month. At the same time ADNOC signed an agreement with Microsoft to co-develop and deploy AI agents across its operations.", "Metric": "40,000+ trained; >90% utilization; 70,000+ hours/month saved", "Source_IDs": "R005", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Enterprise-wide training plus usage and hours-saved telemetry is the adoption playbook to copy; report utilization, not just licenses.", "Partners": "Microsoft, AIQ, Masdar, XRG, Gecko Robotics", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-006", "Initiative_ID": "AI-040", "Title": "ENERGYai deployment contract", "Date": "2025-03-12", "Finding": "AIQ announced a three-year, $340 million contract covering ENERGYai and associated AI solutions for ADNOC. The plan covers more than 28 producing fields, with an initial five-agent version expected in mid-2025 before wider scaling.", "Metric": "$340m / three years / planned coverage of 28+ producing fields", "Source_IDs": "R006", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Buying agentic AI as a multi-year outcome contract with a captive AI vendor is an alternative to piecemeal pilots.", "Partners": "AIQ, G42, Microsoft, OpenAI, OSDU", "Original_date": "2025-03", "Related_IDs": []}, {"ID": "REF-007", "Initiative_ID": "AI-040", "Title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "Date": "2025-01", "Finding": "ADNOC and AIQ completed a 90-day trial of ENERGYai, a 70B-parameter LLM tuned on 50+ years of ADNOC data and drawing on over 15% of its onshore and offshore wells. The trial showed a 70% improvement in accuracy on major aspects of seismic interpretation plus better anomaly detection in reservoir monitoring.", "Metric": "70B-parameter LLM; 90-day PoC; >15% of wells; 70% seismic interpretation accuracy gain", "Source_IDs": "R007", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-boxed PoCs with a hard accuracy metric on a core technical workflow make the case for scaling agentic tools.", "Partners": "AIQ", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "REF-008", "Initiative_ID": "AI-040", "Title": "ADNOC moves AI from pilots to enterprise-wide operations", "Date": "2026-08", "Finding": "ADNOC is embedding AI across exploration, drilling, production, maintenance and decision-making, aiming to become 'the world's most AI-enabled energy company.' Tools include ENERGYai agents, the Neuron 5 asset-performance platform for predicting equipment failure, and an internal AI Lab that cuts proof-of-concept-to-deployment time 3x.", "Metric": "3x faster PoC-to-deployment; months-to-days on technical tasks", "Source_IDs": "R008", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A dedicated internal AI lab with a mandate to shorten PoC-to-production is what turns pilots into operating change.", "Partners": "AIQ, Microsoft, G42", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-009", "Initiative_ID": "AI-040", "Title": "AIQ and SLB co-develop ENERGYai for subsurface", "Date": "2025-05", "Finding": "AIQ and SLB agreed to advance development and deployment of the ENERGYai agentic solution across ADNOC's subsurface operations, spanning geology, seismic and reservoir workflows. SLB brings its digital and agentic AI portfolio, including its Tela assistant.", "Metric": "No quantified metric supplied", "Source_IDs": "R009", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "OFS majors are becoming the channel for agentic subsurface AI; enablement builders should plan to integrate with SLB and Halliburton platforms.", "Partners": "AIQ, SLB (Tela)", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "REF-010", "Initiative_ID": "AI-072", "Title": "First fully autonomous AI-drilled well offshore Guyana", "Date": "2026-03", "Finding": "ExxonMobil completed what it called the industry's first fully closed-loop automated well placement with full rig automation in the Stabroek Block. VP of Exploration John Ardill said no driller touched the controls and AI handled drilling and geosteering through the reservoir section.", "Metric": "No quantified metric supplied", "Source_IDs": "R010", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Closed-loop autonomy is arriving first in drilling, where outcomes are measurable and operations are repeatable; human oversight shifts to remote centers.", "Partners": "Halliburton (LOGIX, EarthStar), Sekal (DrillTronics), Noble", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-011", "Initiative_ID": "AI-073", "Title": "AI identifies four new Guyana exploration opportunities", "Date": "2026-08", "Finding": "ExxonMobil said it used AI, including deep learning, machine learning, HPC and advanced seismic imaging, applied to historical discoveries and drilling results to identify four new exploration opportunities in Guyana's Stabroek Block. It is expanding use of these tools to evaluate prospects that were previously hard to assess.", "Metric": "4 new opportunities identified", "Source_IDs": "R011", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Reinterpreting legacy subsurface data with AI can surface new prospects in mature acreage; data readiness is the prerequisite.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-012", "Initiative_ID": "AI-024", "Title": "AI plus process mining behind $15B cost-cut push", "Date": "2024-11", "Finding": "ExxonMobil tied its $15B operating-cost reduction target for 2027 to AI combined with process intelligence run through its Global Business Solutions group. It uses Celonis process mining across procurement, supply chain, logistics, maintenance and AP/AR, and the CFO acknowledged that historically siloed, inconsistent processes had held back automation.", "Metric": "$15B cost-savings target by 2027", "Source_IDs": "R012", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardize and mine processes before layering on AI agents; back-office AI ROI depends on clean end-to-end process data.", "Partners": "Celonis, AWS, Snowflake, SAP, Salesforce", "Original_date": "2024-11", "Related_IDs": []}, {"ID": "REF-013", "Initiative_ID": "AI-074", "Title": "ApEX multi-agent GenAI searches 1M+ exploration files", "Date": "2025-12", "Finding": "Chevron's in-house generative AI tool ApEX, launched in August 2024 and first used in Gulf of America deepwater exploration, lets explorationists query through a chat interface. Multiple specialized search agents comb more than a million files, cutting pre-drill studies from weeks of manual review to minutes.", "Metric": "1M+ files; weeks to minutes", "Source_IDs": "R013", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A multi-agent RAG over technical archives is the most repeatable subsurface GenAI win; specialize agents by data type (geospatial, assessments, rock properties).", "Partners": "In-house", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-014", "Initiative_ID": "CTX-REF-014", "Title": "Up to 20% workforce cut, technology and hubs cited", "Date": "2025-02", "Finding": "Chevron said it would cut 15-20% of its global workforce (up to ~9,000 roles) by end-2026 to save $2-3B. Levers included leveraging technology to raise productivity and changing where work is done via global capability centers such as Manila, Buenos Aires and its Bengaluru ENGINE center.", "Metric": "15-20% of workforce; ~9,000 roles; $2-3B savings by 2026", "Source_IDs": "R014", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Majors are pairing technology-enabled productivity with centralized service hubs; AI enablement pitches should target work consolidated into these centers.", "Partners": "", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-015", "Initiative_ID": "AI-075", "Title": "Agentic AI for robotic inspections", "Date": "2026-06-02", "Finding": "DataRobot and Chevron described edge agents for robotic mission planning and continuous Safe Start assessment using gas sensors, vision and local AI.", "Metric": "No quantified financial benefit disclosed", "Source_IDs": "R015", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Agentic AI adds most value where it removes human approval bottlenecks on already automated physical workflows, with governance built in.", "Partners": "DataRobot, NVIDIA NIM", "Original_date": "2026-06", "Related_IDs": []}, {"ID": "REF-016", "Initiative_ID": "AI-076", "Title": "Shell extends C3 AI with agent-based root-cause analysis", "Date": "2026-06", "Finding": "Shell signed a new multi-year agreement extending its C3 AI Reliability program, which has monitored equipment across its assets since 2018. The expansion adds AI agent-based root-cause analysis and remediation on top of anomaly detection, running on Microsoft Azure.", "Metric": "13,000+ pieces of equipment monitored; program since 2018", "Source_IDs": "R016", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Mature predictive-maintenance estates are the natural base for agents: move from 'alert' to 'diagnose and recommend fix.'", "Partners": "C3 AI, Microsoft Azure", "Original_date": "2026-06", "Related_IDs": []}, {"ID": "REF-017", "Initiative_ID": "AI-010", "Title": "SLB and Shell co-develop agentic AI for upstream", "Date": "2025-12", "Finding": "SLB and Shell signed a strategic collaboration to develop agentic AI solutions for technical experts across subsurface, well construction and production. The work builds on SLB's Lumi data and AI platform and Shell's earlier deployment of Petrel across its global assets.", "Metric": "No quantified metric supplied", "Source_IDs": "R017", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Supermajors are co-developing agents with OFS vendors on shared data platforms; independent builders should interoperate with Lumi/OSDU instead of competing with them.", "Partners": "SLB (Lumi, Petrel)", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-018", "Initiative_ID": "CTX-REF-018", "Title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "Date": "2024-09", "Finding": "Shell and NVIDIA used Fourier Neural Operators in NVIDIA PhysicsNeMo to build surrogate models of CO2 storage that run about 100,000x faster than numerical simulators while keeping R-squared above 0.97. This enables screening tens of thousands of storage configurations for CCS site selection.", "Metric": "~100,000x speedup; R-squared >0.97; 1,000-year simulation horizon", "Source_IDs": "R018", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Physics-ML surrogates turn slow simulation into fast scenario screening, a strong pattern for subsurface and CCS workflows.", "Partners": "NVIDIA PhysicsNeMo", "Original_date": "2024-09", "Related_IDs": []}, {"ID": "REF-019", "Initiative_ID": "CTX-REF-019", "Title": "bp renews Palantir for 5 years with AIP LLMs", "Date": "2024-09", "Finding": "bp renewed its decade-long Palantir relationship with a five-year strategic agreement centered on Palantir's Artificial Intelligence Platform (AIP). AIP brings LLMs into bp's operations on top of digital twins fed by more than 2 million sensors to speed human decision-making.", "Metric": "5-year deal; 2M+ sensors", "Source_IDs": "R019", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Operators with an existing ontology and digital-twin layer can add LLM copilots quickly; the data model is the durable asset.", "Partners": "Palantir (Foundry, AIP)", "Original_date": "2024-09", "Related_IDs": []}, {"ID": "REF-020", "Initiative_ID": "CTX-REF-020", "Title": "bp expands job cuts as AI drives efficiency", "Date": "2025-08", "Finding": "bp raised expected job losses to about 6,200 (roughly 15% of office-based staff) plus contractor cuts. CEO Murray Auchincloss said technology is helping improve capital productivity and drive cost reductions, with AI part of the overhaul toward a $5B cost-cut target by end-2027.", "Metric": "~6,200 roles; $900M savings H1 2025; $5B cost-cut target by 2027", "Source_IDs": "R020", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "AI is now explicitly linked to headcount plans at majors; change management and redeployment belong in the business case alongside productivity.", "Partners": "", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "REF-021", "Initiative_ID": "AI-077", "Title": "TotalEnergies and Mistral launch €100M frontier subsurface AI program", "Date": "2026-09", "Finding": "TotalEnergies and Mistral AI announced a three-year joint program worth more than €100M to build frontier, agentic AI models for geoscientists in exploration, reservoir characterization and development. A joint lab pairs TotalEnergies subsurface experts with Mistral ML researchers on about 10 petaflops of data.", "Metric": ">€100M over 3 years", "Source_IDs": "R021", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "European majors are backing sovereign model vendors to co-train domain models on proprietary subsurface data.", "Partners": "Mistral AI", "Original_date": "2026-09", "Related_IDs": []}, {"ID": "REF-022", "Initiative_ID": "AI-029", "Title": "Mistral joint lab: assistant for 1,000 researchers", "Date": "2025-06", "Finding": "TotalEnergies and Mistral AI set up a joint innovation lab. First use cases include an AI assistant for TotalEnergies' 1,000 researchers, decision support for industrial asset performance and CO2 reduction, customer energy-saving tools, and specialized assistants for support functions.", "Metric": "1,000 researchers; 300-person Digital Factory; 100+ digital solutions, 60 using ML/GenAI", "Source_IDs": "R022", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Start with a bounded expert population (R&D) for a domain assistant, then extend to support functions.", "Partners": "Mistral AI", "Original_date": "2025-06", "Related_IDs": ["AI-030"]}, {"ID": "REF-023", "Initiative_ID": "CTX-REF-023", "Title": "Early adopter of Microsoft Copilot across the workforce", "Date": "2024-02", "Finding": "TotalEnergies said it was one of the first companies to adopt Microsoft's generative AI assistant to boost employee productivity. Its 300-person Digital Factory runs AI for production forecasting, predictive maintenance and renewables siting, with experts keeping the final say.", "Metric": "300 AI/digital experts in Digital Factory", "Source_IDs": "R023", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Horizontal copilots plus a central digital factory for vertical use cases is the standard two-track model among majors.", "Partners": "Microsoft Copilot", "Original_date": "2024-02", "Related_IDs": []}, {"ID": "REF-024", "Initiative_ID": "AI-078", "Title": "Industrial AI across operated upstream assets", "Date": "2025-09-26", "Finding": "TotalEnergies and Cognite announced a three-year deployment of industrial data and AI capabilities across all TotalEnergies-operated upstream assets worldwide, spanning drilling through production.", "Metric": "Three-year program / all operated upstream assets", "Source_IDs": "R024", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "A contextualized industrial data layer comes before agents; budget for it explicitly.", "Partners": "Cognite (Data Fusion)", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "REF-025", "Initiative_ID": "AI-033", "Title": "AI saved Equinor $130M in 2025", "Date": "2026-01", "Finding": "Equinor said AI created $130M in value and savings in 2025 and more than $330M since 2020. Predictive maintenance on 700+ rotating machines with 24,000 sensors accounts for $120M since 2020; AI interpreted 2 million km2 of seismic in 2025 at 10x capacity; and a well-planning tool saved $12M on Johan Sverdrup Phase 3.", "Metric": "$130M in 2025; $330M+ since 2020; 700+ machines, 24,000 sensors; 10x seismic interpretation; $12M Sverdrup saving", "Source_IDs": "S20", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Publish value by use case; a few industrial ML workhorses (rotating equipment, seismic) still deliver most of the dollars while GenAI tools spread to employees.", "Partners": "", "Original_date": "2026-01", "Related_IDs": ["AI-034", "AI-035", "AI-036"]}, {"ID": "REF-026", "Initiative_ID": "AI-027", "Title": "HPC6 supercomputer and company-wide responsible-AI training", "Date": "2025-06", "Finding": "Eni runs HPC6, launched November 2024 and ranked in the global TOP10, for AI-driven geoscience, safety (the 'Safety Pre-sense' tool) and other uses, and has opened HPC6 access to startups through Call4Innovators. From 2025 it rolled out a company-wide training program on responsible AI, regulation, ethics and data literacy, with advanced modules.", "Metric": "HPC6 in global TOP10; 10 Call4Innovators projects selected (Dec 2025)", "Source_IDs": "R026", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Owning compute is a differentiator for subsurface AI; pairing it with mandatory AI literacy helps meet EU AI Act obligations.", "Partners": "HPE, AMD, CINECA, Plug and Play", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-027", "Initiative_ID": "AI-032", "Title": "Copilot to 25,000 staff; pilot saved 121 min/week", "Date": "2026-02", "Finding": "Repsol extended Copilot to about 25,000 employees after a four-month M365 Copilot pilot with 550 users showed 121 minutes saved per person per week and 16% higher work quality. It runs 60+ GenAI use cases and 34 agents, and has trained 5,000+ employees plus all senior management in prompting.", "Metric": "~25,000 employees; 121 min/week saved; +16% quality; 60+ use cases; 34 agents; 5,000+ trained; 200k+ lines via GitHub Copilot", "Source_IDs": "R027", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Run a measured pilot (time saved, quality, would-not-give-up rate) before mass licensing; it gives the board-ready business case.", "Partners": "Microsoft 365 Copilot, GitHub Copilot", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-028", "Initiative_ID": "AI-031", "Title": "Multi-agent platform: 22 agents supervised by 50 staff", "Date": "2025-10", "Finding": "Repsol launched a multi-agent platform built with Accenture in which AI agents self-organize and coordinate tasks within workflows. The pilot ran three teams with 22 agents in the 700-person Digitalization and Services department, supervised by 50+ employees for over four months.", "Metric": "22 agents; 3 agent teams; 50+ supervisors; 4+ months", "Source_IDs": "S19", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Pilot agents inside your own IT/digital function first, with explicit human supervisors, before pushing to operations.", "Partners": "Accenture", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-029", "Initiative_ID": "AI-031", "Title": "Accenture and NVIDIA to scale AI agents", "Date": "2025-02", "Finding": "Repsol expanded its Digital Program by deploying AI agent systems with Accenture's AI Refinery and the NVIDIA AI platform, targeting process optimization, planning, forecasting, maintenance and customer personalization. It also plans to explore NVIDIA Omniverse digital twins.", "Metric": "650+ digital initiatives; 1,200+ people on digital projects; 50 tech partners", "Source_IDs": "R029", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "SI-led agent factories are a common way integrateds get to scale fast; independent tools need to plug into SI delivery models.", "Partners": "Accenture AI Refinery, NVIDIA AI Enterprise, NVIDIA Omniverse", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-030", "Initiative_ID": "CTX-REF-030", "Title": "ChatPetrobras GenAI rolled out to 110,000 workers", "Date": "2023-12", "Finding": "Petrobras built ChatPetrobras with Microsoft on Azure OpenAI Service (GPT models) and made it available through an internal portal to 110,000 employees and contractors. It supports reports, summaries, presentations and Q&A to cut time and cost in administrative and operational processes.", "Metric": "110,000 workers (employees + contractors)", "Source_IDs": "R030", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A secure internal ChatGPT-style wrapper extended to contractors reaches the whole workforce quickly and cheaply.", "Partners": "Microsoft, Azure OpenAI", "Original_date": "2023-12", "Related_IDs": []}, {"ID": "REF-031", "Initiative_ID": "AI-079", "Title": "Microsoft collaboration on agentic AI and Copilot", "Date": "2025-07-31", "Finding": "PETRONAS and Microsoft signed an MoU to explore agentic AI, Copilot, analytics, cloud and cybersecurity across the value chain, supported by the Malaysia West cloud region.", "Metric": "No deployment or financial metric disclosed", "Source_IDs": "R031", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "In-country cloud regions are unlocking Copilot/agent rollouts at Asian NOCs.", "Partners": "Microsoft Copilot, Azure (Malaysia West)", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "REF-032", "Initiative_ID": "AI-080", "Title": "Agentic AI added to myPROdata upstream platform", "Date": "2026-09", "Finding": "Petronas, through Malaysia Petroleum Management, partnered with Iraya Energies to add agentic AI to its myPROdata upstream data platform so it can autonomously analyze and connect G&G and engineering data for investors. The aim is to speed E&P evaluation and support RM50-60B a year of upstream investment.", "Metric": "RM50-60B/yr investment target; discovery-to-first-oil goal cut from 100 to 50 months", "Source_IDs": "R032", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Agentic AI is being pointed outward at investors and partners; data-room agents are a monetizable NOC use case.", "Partners": "Iraya Energies", "Original_date": "2026-09", "Related_IDs": []}, {"ID": "REF-033", "Initiative_ID": "AI-081", "Title": "CNPC releases 300B-parameter Kunlun energy model", "Date": "2025-05", "Finding": "CNPC released an upgraded 300B-parameter Kunlun large model, which it described as the first large model approved by regulators in China's energy and chemical sector (approved August 2024). The industry LLM scaled from 70B to 300B parameters, alongside larger vision (4.4B) and multimodal (80B) models.", "Metric": "LLM 70B to 300B; vision 300M to 4.4B; multimodal 16B to 80B", "Source_IDs": "R033", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Chinese NOCs are building sovereign domain foundation models; Western vendors are largely shut out of that stack.", "Partners": "China Mobile, Huawei, iFlytek", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "REF-034", "Initiative_ID": "AI-082", "Title": "Offshore predictive maintenance cuts failures ~20%", "Date": "2026-04", "Finding": "CNOOC deployed AI predictive maintenance across offshore platforms including Bohai Oilfield, using real-time sensor data from subsea pumps and turbines. China Daily reported roughly a 20% cut in critical equipment failure rates; CNPC's Changqing field reported ~30% shorter exploration cycles and 10%+ higher per-well output.", "Metric": "~20% fewer critical equipment failures (CNOOC); ~30% shorter exploration cycle, >10% per-well output (CNPC Changqing)", "Source_IDs": "R034", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Rotating-equipment predictive maintenance remains the most consistently quantified AI win across all regions.", "Partners": "", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-035", "Initiative_ID": "AI-041", "Title": "Lumina enterprise AI platform runs 50+ agents", "Date": "2025-12", "Finding": "Woodside's enterprise AI platform Lumina supports more than 50 AI agents handling day-to-day tasks across business functions, overseen by an AI Centre of Excellence and a senior-leader AI Council. Microsoft Copilot covers personal productivity and ChatGPT Enterprise is used as a sandbox for testing custom agents before they move into Lumina (date approximate; undated company page).", "Metric": "50+ AI agents in production", "Source_IDs": "S24", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A layered stack (Copilot for everyone, ChatGPT Enterprise as sandbox, an in-house agent platform for production) plus a governance council is a template worth copying.", "Partners": "Microsoft Copilot, ChatGPT Enterprise", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-036", "Initiative_ID": "AI-083", "Title": "Xecta AI digital twin expanded to East Australia and PNG", "Date": "2025-07", "Finding": "After a successful rollout in the Cooper Basin, Santos signed a five-year agreement with Xecta to build an AI-driven digital twin for production optimization across its Eastern Australia and Papua New Guinea operations. This builds on its 2023 Kelvin.ai agreement for autonomous control of wellsite and facility operations.", "Metric": "5-year agreement", "Source_IDs": "R036", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Mid-size integrateds scale AI by proving value in one basin and then signing enterprise-wide multi-year deals.", "Partners": "Xecta, Kelvin.ai, SIG ML", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "REF-037", "Initiative_ID": "AI-084", "Title": "AI Hub and enterprise adoption", "Date": "2026-04-02", "Finding": "OMV’s 2025 report describes an Azure-based AI Hub launched in 2025, with over 2,500 monthly GPT users and over 10,000 staff trained.", "Metric": "50+ projects in use / 40+ in development / 210+ ideas", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "A single AI front door with a visible pipeline (ideas to development to live) makes adoption measurable.", "Partners": "Microsoft Azure, SAP, Salesforce, SLB, TerraAI", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-038", "Initiative_ID": "AI-085", "Title": "AI Innovation Center delivers G Agent drilling win", "Date": "2025-09", "Finding": "KOC opened its first AI Innovation Center in August 2025 with Microsoft, Halliburton and Ghaia.ai. Its first result is an AI-powered rig drilling project on Ghaia.ai's G Agent platform, a human-agent collaboration mesh that improves productivity and planning of drilling rig operations.", "Metric": "No quantified metric supplied", "Source_IDs": "R038", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Co-located innovation centers with a hyperscaler, an OFS firm and an agent startup can get a first production use case within weeks.", "Partners": "Microsoft, Halliburton, Ghaia.ai (G Agent)", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "REF-039", "Initiative_ID": "AI-086", "Title": "ChatDVN internal AI platform plus AI-101 training", "Date": "2025-07", "Finding": "Devon built ChatDVN, a proprietary platform combining AI tools, machine learning and LLMs, and paired it with a company-wide AI-101 course. CTO Trey Lowe cited productivity gains among engineers, software developers and geoscientists, and use cases such as gas-lift optimization and simul-frac planning done in hours. Devon is standardizing wellsite sensors and cameras so AI can scale, including planned video-to-data analytics in 2026.", "Metric": "About 1,000 employees completed AI-101 by end of June 2025; 15-30% productivity gain for engineers, developers and geoscientists; 5% oil uplift at one Delaware gas-lift site", "Source_IDs": "R039", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A branded internal assistant plus a mandatory foundational course gets adoption started faster than rolling out tools on their own.", "Partners": "ChatDVN (in-house)", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "REF-040", "Initiative_ID": "AI-006", "Title": "AI-enabled lift and maintenance drive $1B optimization plan", "Date": "2026-02", "Finding": "On the Q4 2025 call, Devon said it is taking AI-enabled artificial lift optimization and advanced analytics beyond pilots, and adding condition-based maintenance and faster drilling cycle times. Management said investments in AI and platforms built over the prior year were 'coming to fruition' in production. The work sits inside a $1B pre-tax business optimization program.", "Metric": "85% of $1B optimization target captured by YE2025; more than 100 active workstreams; downtime cut from about 7% to under 5%; base production about 5,000 BOE/d ahead of plan", "Source_IDs": "R040", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Track AI as named workstreams inside a cost program with a dollar target, so the impact shows up in guidance and not only in anecdotes.", "Partners": "", "Original_date": "2026-02", "Related_IDs": ["AI-007"]}, {"ID": "REF-041", "Initiative_ID": "AI-006", "Title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "Date": "2026-05", "Finding": "Devon's Smart Gas Lift program pairs AI models with a physics-based calculation to set gas-lift injection rates in a closed loop directly at the wells. A 2025 pilot delivered a 2-3% uplift. Devon then moved to full implementation in the Delaware and is piloting similar AI models on ESPs and rod pumps.", "Metric": "Pilot uplift 2-3%; more than 850 wells live, heading to about 1,500; full-scale uplift said to exceed the pilot", "Source_IDs": "R041", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Closed-loop setpoint control on artificial lift, a physics model plus ML, is a proven first 'agentic' use case in production ops that scales well by well.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-042", "Initiative_ID": "AI-006", "Title": "AI speeds Coterra integration; in-house plus vendor model", "Date": "2026-08", "Finding": "CTO Trey Lowe said Devon 'leaned in really hard' on enterprise AI after ChatGPT launched and is merging its operations-focused AI stack with Coterra's ML well-prediction models. Closed-loop gas lift went from a few wells to about 1,000, and was extended to legacy Coterra wells. Service-company partners supply pieces of the solution, but Lowe called 'empowering our employees' the 'secret sauce'. AI tools also benchmark every rig and crew against a best-performance curve to find D&C gaps.", "Metric": "Closed-loop gas-lift wells rose from 'a few' to a couple of hundred (March 2026) to about 1,000; more than 2,000 more Permian candidates; new data tables built in 2 days instead of a couple of months", "Source_IDs": "R042", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "In M&A, a mature AI and data stack becomes a synergy engine: onboarding the acquired staff onto the tools is itself a value lever.", "Partners": "ChatDVN, Smart Gas Lift (in-house); unnamed service-provider partners", "Original_date": "2026-08", "Related_IDs": ["AI-007"]}, {"ID": "REF-043", "Initiative_ID": "AI-007", "Title": "AI agents for well design and completions sequencing", "Date": "2026-05", "Finding": "At IADC Drilling Onshore, Devon's VP of D&C described ChatDVN plus the proprietary InnDrill and InnFrac data applications, which were built over about 2.5 years. Uses include well-design optimization agents, AI-driven completions sequencing, basin-wide pre-drill performance prediction, and supply-chain and contract optimization. These tools support the post-Coterra synergy plan.", "Metric": "$350M of the $1B merger synergy target comes from D&C capital optimization", "Source_IDs": "R043", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Owning the D&C data layer (InnDrill/InnFrac) is what makes design agents and pre-drill prediction possible; build the data apps before the agents.", "Partners": "ChatDVN, InnDrill, InnFrac (in-house)", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-044", "Initiative_ID": "AI-087", "Title": "ML well-prediction models fed capital planning", "Date": "2026-08", "Finding": "According to Devon's CTO after the merger closed, Coterra had invested heavily in AI and machine-learning models for well prediction that informed its capital plans. Those models are now being combined with Devon's operations-focused AI systems.", "Metric": "No quantified metric supplied", "Source_IDs": "R042", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Well-performance prediction models used in capital allocation are a high-value, low-integration-risk AI use case that survives M&A.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-045", "Initiative_ID": "AI-088", "Title": "Generative AI across geology, drilling and safety", "Date": "2025-11", "Finding": "EOG said it uses 'quite a bit of generative AI' to organize geologic data and find hidden trends, alongside real-time drilling optimization, predictive maintenance and some autonomous field operations. The work builds on almost 10 years of smart centralized gas-lift technology and newer deep-learning tools that capture human observation and experiential learning. EOG said it is 'not quite to true agentic intelligence' and continues to favor building apps in-house next to field operations.", "Metric": "About 10 years of ML-driven gas-lift optimization", "Source_IDs": "R045", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Tie in-house software to proprietary field sensors and downhole tools, and capture expert tacit knowledge as training data.", "Partners": "In-house apps", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-046", "Initiative_ID": "AI-089", "Title": "ML 'production optimizers' and 24-hour control room", "Date": "2026-02", "Finding": "EOG described its proprietary production optimizers program, which uses machine learning to optimize base production for better run time and lower cost. EOG is also rolling automation across acquired Utica operations so wells can be managed remotely from its 24-hour control room.", "Metric": "No quantified metric supplied", "Source_IDs": "R046", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardize automation and remote monitoring on acquired assets right away so the ML optimizers have data to work on.", "Partners": "Production optimizers (in-house)", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-047", "Initiative_ID": "AI-045", "Title": "ML and AI in field ops cut downtime post-Endeavor", "Date": "2026-05", "Finding": "Diamondback management said that, after the Endeavor merger, layering machine learning onto its data streams and putting AI into field operations was reducing downtime. They cited this as a big part of the Q1 2026 production beat, together with completion-design and workover optimization.", "Metric": "No quantified metric supplied", "Source_IDs": "R047", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Combining merged teams' playbooks with ML on unified data streams can produce a production beat within a few quarters.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-048", "Initiative_ID": "AI-045", "Title": "Moving to 'pump by exception' with AI and automation", "Date": "2026-08", "Finding": "CFO Jere Thompson said AI 'is helping Diamondback in the office today' and that AI and automation will be very big drivers of a shallower or cheaper-to-maintain production base as the company moves to a pump-by-exception model. Operations leadership said AI and automation are applied first to day-to-day artificial-lift optimization and downtime management. CEO Kaes Van't Hof called it the 'first inning' and said even a 1% improvement in decline rate matters.", "Metric": "No quantified metric supplied", "Source_IDs": "S26", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Framing AI value as lower base decline and sustaining capex makes the investor case concrete, even when adoption is early.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-049", "Initiative_ID": "AI-090", "Title": "CEO: embedding AI into core processes at scale", "Date": "2026-05", "Finding": "Answering a shareholder question at the 2026 annual meeting, CEO Ryan Lance called AI 'an increasingly important enabler'. He said it builds on long-standing analytics and that the company is focused on embedding AI into core business processes to cut manual effort, speed decision cycles and improve consistency across geographically diverse operations. He said all applications run under an enterprise framework for governance, security and responsible use.", "Metric": "No quantified metric supplied", "Source_IDs": "R049", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Enterprise framing (governance first, scale proven use cases, redeploy people to higher-value work) is the template large E&Ps are signaling to shareholders.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-050", "Initiative_ID": "AI-044", "Title": "AI center of excellence and Gulf of America AI", "Date": "2025-02", "Finding": "Oxy said it created an AI center of excellence to align all intercompany AI initiatives. Its Gulf of America operations use AI for supply chain management, asset integrity and reservoir characterization. Oxy also cited large new seismic datasets, including Algeria's largest-ever seismic acquisition, as fuel for AI applications.", "Metric": "No quantified metric supplied", "Source_IDs": "R050", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A central AI CoE that aligns business-unit initiatives, applied first to a complex asset, is a common starting operating model.", "Partners": "", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-051", "Initiative_ID": "AI-044", "Title": "AI subsurface project targets complex Gulf reservoirs", "Date": "2025-08", "Finding": "CEO Vicki Hollub said Oxy is building internal AI capabilities, with a project aimed at the highly complex Gulf of America subsurface, and hoped to begin acting on results within 1-2 years. She said AI is already helping with Permian and other onshore subsurface work. Oxy also credited teams 'using AI' in its operations center with record artificial-lift reliability and workover performance.", "Metric": "No quantified metric supplied", "Source_IDs": "R051", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Point AI at the hardest subsurface problems where incremental recovery is worth the most, and report ops-center reliability wins alongside.", "Partners": "", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "REF-052", "Initiative_ID": "AI-091", "Title": "AI now part of workforce efficiency agenda", "Date": "2026-05", "Finding": "On the Q1 2026 call, Oxy leadership said 'AI is taking on a larger role across all disciplines' and named workforce efficiency as a priority. Richard Jackson said Oxy has been doing 'quite a bit of work on workforce', including AI and re-examining processes and priorities. In Q2 2026, Oxy listed AI with CO2, power and water as core technology and partnership areas behind its $4B incremental cash-flow target.", "Metric": "$4B incremental sustainable cash flow target by 2030 (AI one of several levers)", "Source_IDs": "R052", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Expect AI to be tied explicitly to workforce redesign and process re-engineering, not only to technical use cases.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-053", "Initiative_ID": "AI-092", "Title": "In-house AI across production, drilling and completions", "Date": "2025-11", "Finding": "CEO Brendan McCracken said Ovintiv is 'just at the very front end' of applying AI, and is active in three areas: production operations (uptime, artificial-lift optimization), drilling times and costs, and completions. For the NuVista acquisition, Ovintiv planned to connect wells to its Grand Prairie operations control center and use in-house AI tools and algorithms to optimize lift setpoints, with remote shut-in and restart in minutes. It listed AI as both a near-term and a long-term synergy bucket.", "Metric": "No quantified metric supplied", "Source_IDs": "R053", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Put AI and automation in the acquisition synergy case and connect acquired wells to the control center on day one.", "Partners": "Operations control center; in-house AI algorithms", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-054", "Initiative_ID": "AI-093", "Title": "DRIVE Center AI optimizes rigs from day one", "Date": "2026-02", "Finding": "Weeks after closing an acquisition, Ovintiv had linked all producing wells to its operations control center and the drilling rigs to its DRIVE Center, an optimization tool that uses AI to improve drilling performance. Management said this lets synergies start immediately.", "Metric": "About $1M per well synergy savings built into guidance", "Source_IDs": "R054", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A centralized AI drilling-optimization center makes acquired rigs productive in weeks, which strengthens bid economics.", "Partners": "DRIVE Center (in-house)", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-055", "Initiative_ID": "AI-094", "Title": "AI tunes 70-odd frac design inputs in real time", "Date": "2026-05", "Finding": "McCracken called AI paired with Ovintiv's private dataset 'the big technology frontier'. The company develops in-house algorithms for production operations centers (uptime, artificial lift) and for frac design, tuning 'the 70 odd input design factors' of each frac job in real time.", "Metric": "About 70 frac design input factors optimized in real time", "Source_IDs": "R055", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A proprietary operating dataset plus in-house models is presented as a moat that competitors cannot copy.", "Partners": "In-house algorithms", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-056", "Initiative_ID": "AI-095", "Title": "AI tools to target service rebids and procurement", "Date": "2026-02", "Finding": "EQT said it will aggressively rebid oilfield services and that AI tools will help pinpoint those efforts. It expects low-single-digit cost reductions on the procurement side, and said it is focused on 'the procurement and the backside' as well as the field.", "Metric": "Low-single-digit % procurement savings targeted", "Source_IDs": "R056", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Procurement analytics (spend categorization, rebid targeting) is a quick back-office AI win with measurable savings.", "Partners": "", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-057", "Initiative_ID": "AI-096", "Title": "Vendor-led AI: 'fast follower' approach", "Date": "2026-02", "Finding": "CEO Joe Foran said Matador's most effective path has been working with its vendors, who 'most often' already have the AI programs and are ahead. He described 'baby steps' taken deliberately with partners as a win-win.", "Metric": "No quantified metric supplied", "Source_IDs": "R057", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Mid-caps can reach AI value by partnering with service vendors instead of building platforms, but need their own governance.", "Partners": "Unnamed oilfield vendors", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-058", "Initiative_ID": "AI-096", "Title": "AI committee; AI used for lateral targeting", "Date": "2026-05", "Finding": "Matador formed a cross-functional committee, led by Glenn Stetson with Jordan Ellington, to adopt AI in a 'controlled, organized fashion'. Uses include AI-driven analytics that surface actionable alerts to cut downtime, real-time frac monitoring (pressures, water and sand volumes, logistics) as simul- and trimul-frac expand, and AI-assisted geosteering to stay in the best part of the target zone.", "Metric": "More than 36 drilling records set in Q1 2026 (MaxCom-supported)", "Source_IDs": "R058", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A small named AI committee with a clear 'fast follower' mandate is a lightweight governance model for mid-cap operators.", "Partners": "MaxCom (in-house real-time center)", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-059", "Initiative_ID": "AI-097", "Title": "AI rod-pump optimization arrests base decline", "Date": "2026-05", "Finding": "Chord adjusted its artificial-intelligence rod-pump controls to maximize productivity from older Bakken wells, alongside lowering pumps deeper and adding workover rigs. Management showed a positive impact on base production and a halt in decline for the targeted well group.", "Metric": "No quantified metric supplied", "Source_IDs": "R059", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Retuning AI lift controllers for legacy wells is a cheap, short-cycle barrel source when prices call for it.", "Partners": "", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-060", "Initiative_ID": "AI-097", "Title": "AI rod-pump program deployed field-wide", "Date": "2026-08", "Finding": "Chord said AI optimization of its rod-pump program is now 'pretty broadly implemented' across its Williston field, where nearly every well ends up on rod pump. The system sets pump loading to reduce wear and improve production, and Chord is looking for other places to apply similar computer-driven optimization.", "Metric": "No quantified metric supplied", "Source_IDs": "R060", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Standardized lift type (rod pump) across a field makes one AI optimizer broadly deployable.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-061", "Initiative_ID": "AI-098", "Title": "Formal AI policy: no AI in production-critical processes", "Date": "2026-03", "Finding": "CNRL's 2025 Annual Information Form says it uses AI 'responsibly' to widen employee access to knowledge, deliver data-driven insights and improve efficiency through 'incremental, cost-justified deployments'. It will not rely on AI for time-sensitive or production-critical processes until controls meet necessary standards. Employees must stay involved in and accountable for AI outputs, and AI risk is reported to board committees.", "Metric": "No quantified metric supplied", "Source_IDs": "R061", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A conservative, human-in-the-loop policy disclosed in securities filings shows how cost-focused operators govern GenAI.", "Partners": "", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-062", "Initiative_ID": "AI-099", "Title": "Autonomous haul fleet nears 140 trucks", "Date": "2025-07", "Finding": "Suncor runs Canada's largest autonomous haul truck fleet at its Base Plant oil sands mines, with nearly 120 Komatsu 980E trucks in 2025 and a plan for 140 by year-end. CEO Rich Kruger said the technology 'makes us a better operator'. The program was announced in 2018 with an expected loss of about 400 heavy-equipment operator jobs over six years.", "Metric": "About 120 autonomous trucks (target 140 by YE2025); about 400 operator positions affected (2018 plan)", "Source_IDs": "R062", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Autonomy at fleet scale changes the workforce, so a transition plan needs to be part of the business case.", "Partners": "Komatsu (AHS)", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "REF-063", "Initiative_ID": "AI-100", "Title": "Machine learning shortens refinery turnarounds", "Date": "2026-02", "Finding": "CEO Mark Lashier said Phillips 66 is taking work out of turnarounds and using tools 'like machine learning' to plan and execute them. This shortens duration and financial impact, and the company cut its turnaround spend forecast substantially mid-2025 and still beat it.", "Metric": "Turnaround spend forecast cut substantially mid-2025, then beaten", "Source_IDs": "R063", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "ML-driven turnaround scoping (moving work outside the outage window) is a direct availability and cost lever for refiners.", "Partners": "", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-064", "Initiative_ID": "AI-101", "Title": "Company AI program 'normalizing' AI for front line", "Date": "2026-08", "Finding": "Lashier said Phillips 66 has implemented an AI program that is 'normalizing the use of AI inside the company', with dedicated staff finding use cases and extending them across the organization. Front-line engineers use AI to extract data and deploy fixes in 'days and weeks' instead of months or years. Refining leadership said AI helps it find opportunities hidden in 'the fog of too much data' across about 200 cost-out projects.", "Metric": "About 200 cost-reduction projects in flight; refining cost cut by more than $1/bbl, heading to $1.50/bbl (AI one of several contributors)", "Source_IDs": "R064", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A central use-case team that replicates front-line wins across sites, described as 'human-centered', is a scalable refinery AI operating model.", "Partners": "", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-065", "Initiative_ID": "AI-102", "Title": "'Cautiously optimistic' on AI; data quality first", "Date": "2025-10", "Finding": "Valero's refining SVP Greg Bram said Valero is evaluating several AI and ML applications and has deployed some to improve availability. He stressed that success depends on high-quality operating data, which Valero began standardizing 10-15 years ago. Valero also uses robotics for tank cleaning, drones for inspections, and control-room consolidation.", "Metric": "No quantified metric supplied", "Source_IDs": "R065", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A decade of consistent operating-data practices is the prerequisite for refinery AI; start with availability and inspection use cases.", "Partners": "", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-066", "Initiative_ID": "CTX-REF-066", "Title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "Date": "2024", "Finding": "Marathon Petroleum reported that Imubit's closed-loop AI optimization, layered on its existing advanced process controls, cut suboptimal coker drum cycles and the associated giveaway. The result was presented at the 2024 AFPM Summit. MPC is also a strategic investor in Flyscan, an AI aerial pipeline leak-detection company.", "Metric": "25% reduction in suboptimal coke-drum cycles", "Source_IDs": "R066", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Reinforcement-learning optimizers layered on existing APC can deliver measurable unit gains without replacing controls.", "Partners": "Imubit, Flyscan", "Original_date": "2024", "Related_IDs": []}, {"ID": "REF-067", "Initiative_ID": "AI-103", "Title": "New CIO; AI model out-trades 10-year gas trader", "Date": "2025-08", "Finding": "CEO Chad Zamarin said Williams hired a new chief information officer with 'a lot of passion around AI' to drive cost efficiency and business transformation, including AI in project execution and analytics in market intelligence. In a month-long internal test marketing gas in Dallas-Fort Worth, young data-analytics staff building AI models beat a 10-year Sequent trader. Williams is also applying data to maintenance strategy.", "Metric": "AI modeler scored 96% of perfect-foresight execution vs. 94% for the veteran trader", "Source_IDs": "R067", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Head-to-head pilots against top human performers are a persuasive way to build the internal case for AI in trading.", "Partners": "", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "REF-068", "Initiative_ID": "AI-104", "Title": "Strategic investment in Flyscan AI leak detection", "Date": "2025-01", "Finding": "Kinder Morgan joined Flyscan Systems' strategic investor group alongside Marathon Petroleum, Enbridge and Hatch. Flyscan uses remote sensing and AI to detect liquid hydrocarbon leaks and automate right-of-way inspection from patrol aircraft, and is expanding into natural-gas detection. KM COO James Holland called the technology 'state-of-the-art'.", "Metric": "No quantified metric supplied", "Source_IDs": "R068", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Midstream operators are using corporate venture stakes to shape AI inspection tools for their own networks.", "Partners": "Flyscan Systems", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "REF-069", "Initiative_ID": "AI-009", "Title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "Date": "2025-06", "Finding": "SLB said its Lumi data and AI platform would be available on Mistral Compute, Mistral AI's new European AI infrastructure. Mistral's LLMs have been the primary generative models inside Lumi since the partnership began in early 2024, and this deal lets the platform be hosted entirely within European borders for data residency.", "Metric": "Lumi launched Sept 2024 (SLB Digital Forum, Monaco)", "Source_IDs": "R069", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Where a buyer needs data sovereignty, offering models and hosting that stay in-region wins deals, so a sovereign or on-prem deployment option should be designed in from the start.", "Partners": "Mistral AI, Lumi, Delfi", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-070", "Initiative_ID": "AI-008", "Title": "Tela agentic AI assistant launched for upstream", "Date": "2025-11", "Finding": "SLB launched Tela, an agentic AI assistant built on the Lumi platform that combines LLMs with domain foundation models. SLB says it can interpret well logs, predict drilling issues and optimize equipment, working alongside humans or on its own. It is available across SLB's digital platforms, in the cloud or on premises.", "Metric": "Five-step agent loop: observe, plan, generate, act, learn", "Source_IDs": "R070", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The largest service company is now selling agents as a layer across its whole software estate, so independent agent vendors need either deep workflow niches or to interoperate with Tela.", "Partners": "Lumi, Delfi", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-071", "Initiative_ID": "AI-105", "Title": "SLB and NVIDIA build 'AI Factory for Energy'", "Date": "2026-03", "Finding": "SLB expanded its NVIDIA collaboration into three areas: modular DSX AI-factory data centers, an 'AI Factory for Energy' reference environment with domain generative models and industrial-scale agentic AI on SLB platforms, and accelerated computing across SLB software. SLB will make modular data-center components at its Louisiana technology center.", "Metric": "3.1 million sq ft Louisiana manufacturing center; partnership dates to 2008", "Source_IDs": "R071", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Oilfield services companies are bundling compute infrastructure with domain AI, so pure-software AI vendors should partner with them for delivery rather than compete on infrastructure.", "Partners": "NVIDIA, Lumi, Delfi", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-072", "Initiative_ID": "AI-047", "Title": "SLB Digital Marketplace lists ~200 AI and digital products", "Date": "2026-06", "Finding": "SLB launched a curated marketplace where customers can find and deploy AI agents, domain models, skills, data connectors and apps, including Tela AI skills and agents. Every listing is certified against SLB standards for security and interoperability, and ISVs can apply to publish through the SLB partner program.", "Metric": "~200 products at launch; 30+ partners", "Source_IDs": "R072", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "This is effectively an app store for O&G agents, so startups can use it as a distribution channel, but SLB controls certification and the customer relationship.", "Partners": "Tela, Lumi, Delfi, Ocean store, ISVs", "Original_date": "2026-06", "Related_IDs": []}, {"ID": "REF-073", "Initiative_ID": "AI-106", "Title": "SLB acquires Tachyus physics-ML reservoir startup", "Date": "2026-05", "Finding": "SLB agreed to acquire Houston-based Tachyus, whose Aqueon platform combines machine learning with reservoir physics for waterflood management, pressure forecasting and production optimization. SLB plans to fold it into Delfi and Lumi. Terms were not disclosed.", "Metric": "Aqueon deployed on 7,500+ wells; customers include ConocoPhillips, Shell, Equinor, PDO", "Source_IDs": "R073", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Startups with proven physics-plus-ML products and named major customers are being bought by oilfield services companies, which makes an exit to SLB, Halliburton or Baker Hughes a realistic plan.", "Partners": "Tachyus Aqueon, Delfi, Lumi, ConocoPhillips, Shell, Equinor, Petroleum Development Oman", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-074", "Initiative_ID": "AI-107", "Title": "SLB buys Kingdom, Petra upstream software for agentic AI", "Date": "2026-04", "Finding": "SLB agreed to acquire S&P Global Energy's geoscience and petroleum-engineering software (Kingdom, Petra, Harmony Enterprise), which is widely used by US onshore operators. The two companies will co-develop AI models on S&P's upstream data using Lumi and Tela. Closing is expected in H2 2026 or early 2027.", "Metric": "~110,000 users across ~4,000 organizations (reported); data spans 113 countries", "Source_IDs": "R074", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "SLB is buying the desktop tools independents already use in order to push agentic AI to US onshore operators, the same mid-market segment many AI startups target.", "Partners": "Kingdom, Petra, Harmony Enterprise, Lumi, Tela, S&P Global Energy", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-075", "Initiative_ID": "AI-108", "Title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "Date": "2025-11", "Finding": "ADNOC, Masdar, XRG and Microsoft signed an agreement to co-develop and deploy AI agents across ADNOC's value chain and to supply energy for Microsoft's data centers. The deal builds on ADNOC's enterprise-wide Microsoft Copilot rollout, which began in November 2023.", "Metric": "40,000+ employees AI-trained; Copilot utilization >90%; 70,000+ hours/month productivity gained", "Source_IDs": "R075", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Hyperscalers trade power-supply deals for AI footprint inside NOCs, and ADNOC's Copilot adoption numbers are a useful benchmark for workforce-enablement pitches.", "Partners": "Microsoft Copilot, Azure, ADNOC, Masdar, XRG", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-076", "Initiative_ID": "AI-109", "Title": "Octiv Auto Frac: first fully automated frac with Coterra", "Date": "2025-01", "Finding": "Halliburton and Coterra launched Octiv Auto Frac on the ZEUS electric frac platform. It automates stage execution, making thousands of pumping decisions from the operator's design without a human at the controls, and Coterra extended it to its remaining Permian completion programs.", "Metric": "17% increase in stage efficiency on initial rollout", "Source_IDs": "R076", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Autonomy in completions reaches the field through service companies that own the equipment and control loop, which leaves software-only players to compete on design, analytics and planning.", "Partners": "Coterra Energy, ZEUS, Sensori", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "REF-077", "Initiative_ID": "AI-110", "Title": "ZEUS IQ closed-loop intelligent frac with Chevron", "Date": "2025-06", "Finding": "Halliburton debuted ZEUS IQ in May 2025, describing it as the industry's first closed-loop fracturing platform, combining Octiv Auto Frac with Sensori subsurface monitoring. With Chevron it ran feedback-driven completions in Colorado's DJ Basin, where the job plan was adjusted in real time without human intervention.", "Metric": "No quantified metric supplied", "Source_IDs": "R077", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The frontier has moved from alerts to closed-loop control, so AI builders need live subsurface feedback and actuation partners to compete.", "Partners": "Chevron, Octiv, Sensori", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-078", "Initiative_ID": "AI-111", "Title": "Halliburton acquires Sekal drilling-automation software", "Date": "2026-04", "Finding": "Halliburton bought Sekal from Sumitomo Corporation, combining Sekal's DrillTronics with its LOGIX automation and remote operations. In 2025 the two had delivered the first automated on-bottom drilling system, for Equinor on the Norwegian Continental Shelf.", "Metric": "DrillTronics used on 1,300+ wells; up to 25% reduction in well delivery time (reported)", "Source_IDs": "R078", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Oilfield services companies are buying automation IP they first proved through partnerships, so a joint pilot with a major service company is a common route to acquisition.", "Partners": "Sekal DrillTronics, LOGIX, Equinor, Sumitomo", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-079", "Initiative_ID": "AI-021", "Title": "BakerHughesC3.ai JV renewed through 2028", "Date": "2025-05", "Finding": "C3 AI and Baker Hughes renewed and expanded their joint venture through June 2028 to sell enterprise AI to oil, gas and chemicals, integrated with Cordant. Baker Hughes will also expand internal use of C3 sourcing, inventory and sustainability apps.", "Metric": "JV has generated >$500M revenue from O&G and chemicals since 2019; 5th amendment", "Source_IDs": "R079", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Distribution through an oilfield services company moved more than $500M of AI software, so a channel JV with a service company can be worth more than direct sales.", "Partners": "C3 Agentic AI Platform, Cordant, Shell, Eni, QatarEnergy LNG, Petronas, ExxonMobil, LyondellBasell, Flint Hills", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "REF-080", "Initiative_ID": "AI-020", "Title": "Cordant built on Azure AI Foundry with Microsoft", "Date": "2025-02", "Finding": "Baker Hughes signed an MOU with Microsoft to speed development and adoption of its Cordant industrial software suite, which is built with Azure AI Foundry. It targets predictive maintenance, operations optimization, safety and energy efficiency, and later releases (25.2.1, 26.1) added AI-driven risk intelligence.", "Metric": "No quantified metric supplied", "Source_IDs": "R080", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Incumbents are standardizing on hyperscaler agent frameworks, so building on the same foundry or model stack makes integration and co-selling easier.", "Partners": "Microsoft Azure AI Foundry, Cordant", "Original_date": "2025-02", "Related_IDs": []}, {"ID": "REF-081", "Initiative_ID": "AI-112", "Title": "Leucipa adds genAI assistant, AI lift optimizer for Repsol", "Date": "2025-06", "Finding": "Baker Hughes and Repsol agreed to deploy new Leucipa capabilities, including a generative AI virtual assistant. Repsol is piloting the AI-powered Leucipa Lift Optimizer on plunger-lift wells to reduce the time production engineers spend on individual wells, following an October 2024 MOU on AI field automation.", "Metric": "No quantified metric supplied", "Source_IDs": "R081", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Production-engineer time on lift optimization is a proven, fundable use case that majors now co-develop with oilfield services companies.", "Partners": "Repsol, Leucipa", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-082", "Initiative_ID": "AI-113", "Title": "Weatherford Industrial Intelligence portfolio launched", "Date": "2025-10", "Finding": "At FWRD 2025, Weatherford combined ForeSite, Centro, PetroVisor and new tools into 'Industrial Intelligence'. New pieces include the StrataCore production foundation model, the MetaFlow AI virtual flow meter, Canvas no-code data science, and the WFRD Software Launchpad SaaS hub.", "Metric": "Operates in ~75 countries", "Source_IDs": "R082", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Even mid-sized oilfield services companies now pitch production foundation models and no-code tools, so 'foundation model for production' is no longer a differentiator.", "Partners": "ForeSite, Centro, PetroVisor, StrataCore, Canvas", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-083", "Initiative_ID": "AI-114", "Title": "NOV and AIQ pilot AI drilling on Max platform", "Date": "2026-01", "Finding": "NOV and AIQ agreed to combine NOV's Max industrial IoT and edge-computing platform with AIQ's AI to design and test AI-driven drilling optimization. Work will run through proof-of-concept and pilot deployments in live operations.", "Metric": "No quantified metric supplied", "Source_IDs": "R083", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Equipment makers with edge platforms are looking for AI partners, so an AI builder can reach the rig without owning hardware by plugging into their edge layer.", "Partners": "AIQ, NOV Max", "Original_date": "2026-01", "Related_IDs": []}, {"ID": "REF-084", "Initiative_ID": "AI-115", "Title": "RigCLOUD Powered by Corva drilling intelligence", "Date": "2025-04", "Finding": "Nabors expanded its alliance with Corva by integrating Corva's AI analytics into its RigCLOUD edge and cloud platform, sold as 'RigCLOUD Powered by Corva'. Corva also offers Corva Copilot, an AI assistant for drilling and completions.", "Metric": "No quantified metric supplied", "Source_IDs": "R084", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Drilling contractors white-label startup analytics rather than building their own, which gives startups distribution across a whole rig fleet.", "Partners": "Corva, Corva Copilot, RigCLOUD", "Original_date": "2025-04", "Related_IDs": []}, {"ID": "REF-085", "Initiative_ID": "AI-116", "Title": "FlexRobotics rig-floor robotics deployed for supermajor", "Date": "2026-02", "Finding": "H&P said it deployed its FlexRobotics rig-floor automation system on a rig for a supermajor in the Permian Basin. The system combines robotic arms, automation software and safety controls to keep people out of high-risk zones during drilling and tripping.", "Metric": "No quantified metric supplied", "Source_IDs": "R085", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Physical automation on land rigs is starting to scale, and it will generate new data streams and workflows for software AI to use.", "Partners": "FlexRobotics", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-086", "Initiative_ID": "AI-117", "Title": "Cortex ML autodriller and REX monitoring drive rig margins", "Date": "2025-07", "Finding": "Patterson-UTI credits its Cortex automation platform (machine-learning autodriller), REX cloud early-alert monitoring and its Digital Performance Center with differentiated rig performance. On completions it uses Vertex frac automation and FleetStream real-time data feeds.", "Metric": "No quantified metric supplied", "Source_IDs": "R086", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Contractors market their digital and ML stack as a margin driver, so AI vendors should frame their value in the contractor's terms, such as revenue per rig and performance pricing.", "Partners": "Cortex, REX, Vertex, FleetStream, EOS Completions Platform", "Original_date": "2025-07", "Related_IDs": []}, {"ID": "REF-087", "Initiative_ID": "AI-118", "Title": "Liberty launches Forge LLM for frac asset orchestration", "Date": "2025-10", "Finding": "Liberty launched Forge, its own large language model for intelligent asset orchestration, which works with StimCommander, its AI-driven automated rate and pressure control software. The company reported fleet-level automation gains in its Q3 2025 results.", "Metric": "StimCommander: 65% faster time to target injection rate; 5-10% hydraulic efficiency gain; digiPrime maintenance savings >30%", "Source_IDs": "R087", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Frac companies are building proprietary LLMs to orchestrate their own equipment fleets, which leaves little room for outside vendors in pressure pumping itself.", "Partners": "Forge, StimCommander, digiPrime", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-088", "Initiative_ID": "CTX-REF-088", "Title": "APA extends multi-year Palantir deal, adds AIP", "Date": "2024-09", "Finding": "APA signed a multi-year, multimillion-dollar extension of its 2021 Palantir enterprise agreement, adding the AIP platform. Palantir already supports APA's operational planning, supply chain, maintenance, production optimization and contract management, and AIP is used to flag anomalies in contracts and invoices against source systems.", "Metric": "Multi-year, multimillion-dollar", "Source_IDs": "R088", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Palantir expands inside E&Ps from an operational data layer into back-office genAI such as invoice and contract checks, so land-and-expand from one ontology is the model to beat.", "Partners": "Palantir AIP, Foundry", "Original_date": "2024-09", "Related_IDs": []}, {"ID": "REF-089", "Initiative_ID": "CTX-REF-089", "Title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "Date": "2024-08", "Finding": "bpx energy, which runs bp's US onshore assets, built Xpert with the AWS Generative AI Innovation Center. Xpert is a Bedrock-based chat tool that turns employee questions into SQL or Python, runs the code against bpx production data and answers in text, and is intended to extend to mobile and voice for field staff.", "Metric": "bpx targeting 30% volume growth by 2025 vs 2022", "Source_IDs": "R089", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Text-to-SQL over production data is a common first genAI use case for E&Ps, and hyperscaler innovation teams build it at little or no cost, so vendors need more defensible depth.", "Partners": "Amazon Bedrock, AWS Generative AI Innovation Center, bpx energy", "Original_date": "2024-08", "Related_IDs": []}, {"ID": "REF-090", "Initiative_ID": "AI-119", "Title": "Snowflake launches Energy Solutions with 30+ partner apps", "Date": "2026-01", "Finding": "Snowflake launched Energy Solutions for its AI Data Cloud, bundling governance, 30+ partner-built apps, industry datasets, Snowflake Intelligence and Cortex AI. Named energy customers include ExxonMobil and Expand Energy.", "Metric": "30+ partner apps; 13,900+ Snowflake customers overall", "Source_IDs": "R090", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Data platforms are becoming the distribution layer for O&G AI apps, so listing as a partner app on Snowflake or Databricks puts a vendor next to data the operator already governs.", "Partners": "Snowflake Cortex AI, Snowflake Intelligence, ExxonMobil, Expand Energy, CARTO, Itron", "Original_date": "2026-01", "Related_IDs": []}, {"ID": "REF-091", "Initiative_ID": "CTX-REF-091", "Title": "Data Intelligence Platform for Energy with genAI", "Date": "2024-04", "Finding": "Databricks launched an energy-specific Data Intelligence Platform bringing generative AI to the sector, with partners including AVEVA, Seeq, Deloitte and Capgemini. Shell's digital innovation VP said Databricks had accelerated its analytics and AI.", "Metric": "10,000+ Databricks orgs globally", "Source_IDs": "R091", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Majors' AI and ML work often runs on Databricks, so building natively on its lakehouse (Unity Catalog, model serving) lowers the barrier to buying.", "Partners": "Shell, TotalEnergies, Chevron Phillips Chemical, Cosmo Energy, Wood Mackenzie, AVEVA, Seeq", "Original_date": "2024-04", "Related_IDs": []}, {"ID": "REF-092", "Initiative_ID": "AI-049", "Title": "Aker BP expands AI-first strategy with Cognite Atlas AI", "Date": "2025-09", "Finding": "Aker BP deepened its Cognite partnership to deploy agent-based applications built with Cognite Atlas AI on Cognite Data Fusion. Use cases include root-cause analysis, document automation, barrier management, equipment registration, predictive maintenance and production chemistry.", "Metric": "Root-cause-analysis agents cut engineering time >70%", "Source_IDs": "R092", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A contextualized knowledge graph underneath is what makes agents credible in operations, and time saved per engineering task is the proof metric buyers respond to.", "Partners": "Aker BP, Cognite Atlas AI, Cognite Data Fusion", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "REF-093", "Initiative_ID": "AI-120", "Title": "Cognite Atlas AI major release, low-code agent workbench", "Date": "2025-09", "Finding": "Cognite shipped a major Atlas AI release, which it describes as a low-code industrial agent workbench on real-time OT, IT and engineering data. Customers named include Aker BP, ADNOC Offshore, Hess and HMH, with SLB among its partners, and the company says it adds new customers weekly.", "Metric": "New customers onboarded weekly (company claim)", "Source_IDs": "R093", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "A low-code agent builder where operators assemble their own agents is the competing model to custom agent services.", "Partners": "Aker BP, ADNOC Offshore, Hess, HMH, Celanese, Idemitsu, SLB, Radix", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "REF-094", "Initiative_ID": "AI-121", "Title": "Cognite integrates NVIDIA accelerated computing for agents", "Date": "2025-10", "Finding": "Cognite integrated its Atlas AI and Data Fusion platform with NVIDIA accelerated computing and software. The first use case is real-time anomaly detection in time-series data with Celanese, and Cognite later integrated NVIDIA NV-Tesseract for forecasting.", "Metric": "No quantified metric supplied", "Source_IDs": "R094", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-series anomaly detection at scale is the core agent workload in operations, and partnering with a GPU vendor adds credibility and lowers compute cost.", "Partners": "NVIDIA, Celanese, NV-Tesseract", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-095", "Initiative_ID": "AI-122", "Title": "Invictus invests $35m in Novi Labs", "Date": "2025-06-25", "Finding": "Novi Labs announced $35 million from Invictus Growth Partners to expand its energy AI analytics business. ML forecasts support well design and upstream capital allocation.", "Metric": "$35m investment / company reports >$50bn annual capital decisions through platform", "Source_IDs": "R095", "Evidence": "Primary source text reviewed", "Origin": "User-supplied reference", "Supplied_interpretation": "Growth equity is backing vertical AI analytics for upstream investing, where accurate forecasts on proprietary data are the moat.", "Partners": "Invictus Growth Partners, Shell, ExxonMobil, Chevron, Devon (reported users)", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-096", "Initiative_ID": "AI-123", "Title": "Novi acquires Turing Analytics and RBN Energy", "Date": "2026-05", "Finding": "Novi Labs acquired Calgary-based Turing Analytics (EVA, an April 2026 deal) and Houston-based RBN Energy (a May 2026 deal). The acquisitions connect upstream analytics with Canadian coverage and midstream and commodity-market intelligence in one decision platform.", "Metric": "RBN reaches 120,000+ monthly users", "Source_IDs": "R096", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Well-funded vertical AI players are consolidating data and audience assets, so owning proprietary content becomes a competitive lever.", "Partners": "Turing Analytics (EVA), RBN Energy", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-097", "Initiative_ID": "CTX-REF-097", "Title": "Instant Analyst genAI with cited answers", "Date": "2024-02", "Finding": "Enverus launched Instant Analyst, generative AI over its Intelligence Vault research that returns natural-language answers with document citations to limit hallucinations. It runs on Amazon Bedrock and later expanded to other products, including Courthouse records.", "Metric": "Searches tens of thousands of proprietary research docs", "Source_IDs": "R097", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Answers with citations over proprietary data are what buyers now expect as a minimum for O&G genAI.", "Partners": "Amazon Bedrock, Enverus Intelligence Vault, PRISM, Tenaris, Ironwood Midstream", "Original_date": "2024-02", "Related_IDs": []}, {"ID": "REF-098", "Initiative_ID": "AI-051", "Title": "Blackstone to acquire Enverus (~$6.5B reported)", "Date": "2025-08", "Finding": "Blackstone funds agreed to acquire Enverus, the energy data analytics and SaaS platform, from Hellman & Friedman and Genstar Capital. Reuters reported the deal at roughly $6.5B.", "Metric": "~$6.5B (reported)", "Source_IDs": "S30", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Private equity pays a premium for proprietary energy data that genAI workflows depend on, so owning data is what drives valuation.", "Partners": "Blackstone, Hellman & Friedman, Genstar", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "REF-099", "Initiative_ID": "AI-065", "Title": "Enverus AI launched as 'AI operating system for energy'", "Date": "2025-12", "Finding": "Enverus unveiled Enverus AI, a governed system built on its energy datasets. It covers asset lifecycle management, document analysis, title verification, field safety monitoring and M&A deal screening.", "Metric": "8,000+ customer companies; claims '10X workforce impact'", "Source_IDs": "R099", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Title, document and deal-screening workflows are being built into the incumbent data vendor's product, so specialist startups there face bundling pressure.", "Partners": "Enverus", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-100", "Initiative_ID": "AI-065", "Title": "Enverus ONE governed AI platform with workflow 'Flows'", "Date": "2026-04", "Finding": "Enverus launched Enverus ONE, which pairs frontier models with Astra, its own energy model, as an execution layer for energy workflows. The first 'Flows' are AFE Evaluation, Current Production Valuation, Project Siting and QuickStart, running in a SOC 2 Type II private environment.", "Metric": "8,000+ companies in 50 countries; 2.7 PB data; 350M+ courthouse records; $500B+ annual transactions tracked", "Source_IDs": "S48", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Packaging agents as named, outcome-specific workflows (such as AFE evaluation) on proprietary data is the product pattern to borrow.", "Partners": "Astra, frontier LLMs", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-101", "Initiative_ID": "AI-124", "Title": "Collide raises $5M seed for O&G genAI platform", "Date": "2025-04", "Finding": "Houston-based Collide closed a $5M seed round led by Mercury Fund, with operator angels Bryan Sheffield, Billy Quinn and David Albin. It serves public, private and PE-backed E&Ps and grew from an industry community platform.", "Metric": "$5M seed; 6,000+ community users growing 11% MoM (at raise)", "Source_IDs": "R101", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Building a practitioner community first gave Collide a distribution base before selling enterprise AI.", "Partners": "Mercury Fund", "Original_date": "2025-04", "Related_IDs": []}, {"ID": "REF-102", "Initiative_ID": "AI-125", "Title": "RIGGS energy LLM automates RRC filings", "Date": "2026-05", "Finding": "Collide rolled out RIGGS, an energy LLM that returns cited answers from authoritative sources. With Winn Resources it automated Texas Railroad Commission regulatory filings, and Collide says RIGGS beat general frontier models on a petroleum-engineering exam subset.", "Metric": "Filing time cut from hours to 20 min for 50 wells (>95%); 67.5% on 40-question PE exam subset vs Grok 4 62.5%, Claude Sonnet 4.5 52.5%", "Source_IDs": "R102", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Regulatory filings are a narrow, measurable back-office wedge for small operators, and domain benchmarks against frontier models are becoming a marketing tool.", "Partners": "Winn Resources", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-103", "Initiative_ID": "AI-126", "Title": "Battalion Oil makes strategic investment in Collide", "Date": "2026-09", "Finding": "Public E&P Battalion Oil made an undisclosed strategic investment in Collide and is deploying its Riggs AI-native operations system across its upstream operations. Battalion plans to unify more than 100 TB of operational data, with Riggs reaching its production engineers around day 90.", "Metric": "100+ TB data; ~90-day time to production engineers; Collide cites $20B addressable market", "Source_IDs": "R103", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Operators investing in their AI vendors aligns incentives and creates reference customers, and a stated 90-day time-to-value target is a useful benchmark.", "Partners": "Battalion Oil, Riggs", "Original_date": "2026-09", "Related_IDs": []}, {"ID": "REF-104", "Initiative_ID": "AI-127", "Title": "$20M Series A for Orbital O&G foundation model", "Date": "2026-07", "Finding": "London-based Applied Computing raised a $20M Series A led by KBR, with Databricks Ventures participating. Its Orbital foundation model combines time-series, physics and language modeling to predict plant state, find anomalies and simulate fixes.", "Metric": "$20M Series A; 'double-digit millions' ARR within ~18 months; facilities use <8% of sensor data", "Source_IDs": "R104", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Strategic investors (an engineering firm and a data platform) give distribution, and a hybrid physics-time-series-language model is attracting funding.", "Partners": "KBR (INSITE 3.0), Databricks Ventures, Wipro", "Original_date": "2026-07", "Related_IDs": []}, {"ID": "REF-105", "Initiative_ID": "AI-128", "Title": "Seeq Intelligence with Agent Q and custom agents", "Date": "2026-03", "Finding": "Seeq launched Seeq Intelligence, adding Agent Q (a natural-language AI analyst), build-your-own agents for multi-step workflows, secure connections to customer systems and unstructured-document extraction. It is included in the Seeq Enterprise package, and ExxonMobil and Marathon Petroleum have presented at Seeq events.", "Metric": "No quantified metric supplied", "Source_IDs": "R105", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Time-series analytics incumbents are bundling agents into existing enterprise tiers rather than charging separately, which compresses what standalone agent vendors can charge.", "Partners": "ExxonMobil, Marathon Petroleum, PBF", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-106", "Initiative_ID": "AI-129", "Title": "AVEVA embeds agentic AI, MCP across CONNECT", "Date": "2026-05", "Finding": "At AVEVA World 2026, AVEVA announced agentic twin-builder mapping, an AI-enriched industrial knowledge graph, and MCP integrations that let LLMs read operational data. It also announced a CONNECT major release planned for Q1 2027 and deeper Microsoft, Snowflake and ServiceNow integrations. Its Industrial AI Assistant runs on Azure OpenAI.", "Metric": "3,500+ delegates; 5,000 partners", "Source_IDs": "R106", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Industrial historians now expose data through MCP, so AI builders should plan to connect via MCP to PI and CONNECT rather than build custom connectors.", "Partners": "Microsoft Azure OpenAI, Snowflake, ServiceNow, Cognizant, Crosser", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-107", "Initiative_ID": "AI-130", "Title": "Emerson completes AspenTech buyout for industrial AI", "Date": "2025-03", "Finding": "Emerson bought the remaining AspenTech shares at $265 per share in cash, making AspenTech a wholly owned subsidiary. The deal combines automation hardware with optimization, APM and subsurface software aimed at autonomous operations.", "Metric": "Minority stake ~$7.2B; total AspenTech investment ~$17B (reported)", "Source_IDs": "R107", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Automation vendors are consolidating industrial software to own the path to autonomy, which leaves independent AI players to become partners or acquisition targets.", "Partners": "Emerson, AspenTech", "Original_date": "2025-03", "Related_IDs": []}, {"ID": "REF-108", "Initiative_ID": "CTX-REF-108", "Title": "AI could unlock $230B a year in upstream", "Date": "2026-08", "Finding": "McKinsey sized AI value across the upstream value chain: $65B a year near-term with today's technology, $125B as proven tools scale, and $230B at full potential with autonomous operations. The figures are net of more than $30B a year in implementation costs. Value is concentrated: the top 10 use cases account for nearly half of it, and production optimization and drilling are the largest pools.", "Metric": "$65B / $125B / $230B annual value; >$30B/yr implementation cost; top 10 use cases ~50% of value, top 60 ~95%; <20% of deployments track AI value KPIs; +$35B/yr reserves value from exploration; OFS revenue displacement $17B near-term to $60B", "Source_IDs": "R108", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Fewer than 20% of deployments track AI value, so building value measurement into the enablement program is part of the product.", "Partners": "McKinsey", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-109", "Initiative_ID": "CTX-REF-109", "Title": "Digital and AI: a $500B upstream opportunity", "Date": "2026-05", "Finding": "Rystad estimates that digitalization and AI could create close to $500B of cumulative value for E&Ps from 2026 to 2030, split roughly equally between cost reduction and production gains. E&Ps already investing are expected to capture an extra $80B a year by 2030 compared with 2025. Rystad points to ADNOC ($500M of AI value in 2023) and Equinor as proof points.", "Metric": "~$500B cumulative 2026-2030; +$80B/yr by 2030 vs 2025; five priority areas: drilling optimization, autonomous robotics, predictive maintenance, reservoir management, logistics", "Source_IDs": "R109", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Frame use-case portfolios around the few value pools the analysts agree on (drilling, maintenance, reservoir) instead of spreading effort thin.", "Partners": "Rystad Energy, ADNOC, Equinor", "Original_date": "2026-05", "Related_IDs": []}, {"ID": "REF-110", "Initiative_ID": "CTX-REF-110", "Title": "Digitalization could save $320B by 2030", "Date": "2025-12", "Finding": "Rystad's supply-chain team estimated $320B of oil and gas savings from digitalization through 2030 and called the figure modest. It named high upfront hardware and software costs, cybersecurity requirements and aging infrastructure as the main barriers, especially for smaller operators and service firms.", "Metric": "$320B savings through 2030; SLB digital targeting ~35% margins in 2025; Viridien digital/data revenue $787M (+17%)", "Source_IDs": "R110", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Smaller operators face cost and cyber barriers, so modular, low-infrastructure enablement offerings are an underserved gap.", "Partners": "Rystad Energy, SLB, Halliburton, NOV, Baker Hughes, Viridien", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-111", "Initiative_ID": "CTX-REF-111", "Title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "Date": "2025-08", "Finding": "BCG's CEO briefing argues AI could raise O&G EBIT by 30-70% within five years (about +30% upstream, +50% refining, +70% fuel retail). It splits AI plays into Deploy (GenAI productivity tools, 10-15% productivity), Reshape (end-to-end workflow redesign) and Invent (new business models). Its playbook calls for executive sponsorship, data and platform upgrades, and upskilling of key roles with AI literacy and change management.", "Metric": "+30-70% EBIT potential in 5 yrs; Deploy/Reshape/Invent chart shows 72%/47%/22% of O&G companies with initiatives; top GenAI everyday uses: knowledge management 69%, code 67%, meeting summaries 62% (N=188); leaders put 80% of AI investment into Reshape/Invent", "Source_IDs": "R111", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Seat licences (Deploy) are table stakes. Most of the value sits in redesigning a few core workflows, so enablement has to go past prompt training.", "Partners": "BCG, Cognite, SparkCognition, Microsoft, OpenAI (ChatGPT Enterprise), SLB (Delfi), Aker BP, bp", "Original_date": "2025-08", "Related_IDs": []}, {"ID": "REF-112", "Initiative_ID": "CTX-REF-112", "Title": "Energy execs: fewer than 20% scaling AI with results", "Date": "2026-02", "Finding": "In Bain's 2026 survey of energy and natural resources executives, about two-thirds said their company is experimenting with AI or running pilots without the outcomes it wanted. Only about a quarter have scaled applications with measurable impact, and fewer than 10% are running a companywide AI transformation. Nearly half named unclear outcomes or no link to business value as the biggest impediment, followed by shortages of technical expertise and poor data governance.", "Metric": "~2/3 stuck in experiments/pilots; ~1/4 scaling with measurable impact; <10% companywide transformation; <20% scaling with results; ~50% cite unclear value link as top barrier; 800+ executives surveyed", "Source_IDs": "R112", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The top barrier is unclear business value, not technology, so tie each use case to an owner and a P&L metric before building it.", "Partners": "Bain & Company", "Original_date": "2026-02", "Related_IDs": []}, {"ID": "REF-113", "Initiative_ID": "CTX-REF-113", "Title": "O&G execs: 85% say reskilling decides success; 29% do it", "Date": "2025-03", "Finding": "EY's Future of Energy Survey (1,020 respondents, Nov-Dec 2024) found a large gap between intent and action on workforce reskilling. 85% of oil and gas executives say reskilling will determine success over the next five years, but only 29% are retraining existing employees. 58% of oil and gas companies say their culture hinders adoption of digital technology.", "Metric": "85% say reskilling critical vs 29% retraining; 81% plan more reliance on vendors/contractors; 58% say culture impedes digital adoption; 92% of energy cos plan digital investment vs 27% retraining", "Source_IDs": "R113", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The market gap is in doing the reskilling, not in convincing people it matters.", "Partners": "EY", "Original_date": "2025-03", "Related_IDs": []}, {"ID": "REF-114", "Initiative_ID": "CTX-REF-114", "Title": "Energy leaders struggle to link AI to productivity", "Date": "2025-12", "Finding": "Wave 4 of EY's US AI Pulse Survey found that energy leaders see AI as a catalyst for transformation but have trouble proving its impact. 56% of energy senior leaders struggle to link productivity gains directly to AI, compared with 35% across all industries. 72% say more training is needed to communicate AI productivity gains accurately. EY also notes that employee job-security concerns are holding back experimentation.", "Metric": "56% energy vs 35% all-industry struggle to attribute gains; 72% need more training (vs 50%); 78% see gains as catalyst (vs 55%); 72% report rising responsible-AI interest (+25 pts since Jul 2024)", "Source_IDs": "R114", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Energy lags other sectors on measuring AI gains, so offering an attribution method alongside training is a differentiator.", "Partners": "EY", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-115", "Initiative_ID": "CTX-REF-115", "Title": "96% of O&G execs cite data and vendor challenges", "Date": "2025-01", "Finding": "EY reports that 96% of oil and gas executives see managing alliance and vendor relationships, along with data accessibility, as a challenge for AI and emerging technology. 96% also cite connecting data across organizational silos. EY says the hard part starts when a proof of concept is handed to the team responsible for scaling it.", "Metric": "96% cite vendor/alliance management and data accessibility; 96% cite cross-silo data connection", "Source_IDs": "R115", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Plan the pilot-to-production handoff (data plumbing, vendor orchestration, ownership) as part of any use case, not as an afterthought.", "Partners": "EY", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "REF-116", "Initiative_ID": "CTX-REF-116", "Title": "AI heading past 50% of O&G IT spend by 2029", "Date": "2025-10", "Finding": "Deloitte's 2026 Oil and Gas Industry Outlook projects that AI and GenAI will grow from less than 20% of US O&G IT spending today to more than 50% by 2029. About half of AI investment goes to process optimization. Early adopters of AI-driven maintenance report up to 40% fewer equipment failures and $10M a year in savings. The outlook expects GenAI, agentic AI and real-time analytics to move from pilots to enterprise deployment in 2026.", "Metric": "AI <20% of IT spend now to >50% by 2029; up to 40% fewer failures; $10M/yr savings; 140+ hours downtime prevented, 1.6% uptime protected (one case); 66% of workforce in mechanically intensive roles", "Source_IDs": "R116", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "As budgets swing toward AI, the scarce input becomes people who can use it, which is a direct opening for enablement.", "Partners": "Deloitte", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-117", "Initiative_ID": "CTX-REF-117", "Title": "1.2M US energy and chemicals workers need upskilling", "Date": "2025-05", "Finding": "Deloitte's workforce study says more than 60% of the US energy and chemicals workforce (about 1.2 million people) will need upskilling in new technologies, process operations and analytics, and another 534,000 will need reskilling. It expects about one-third of the tasks analyzed to be led by machines and agentic AI with humans in the loop.", "Metric": "1.2M workers (>60%) need upskilling; 534K need reskilling; ~1/3 of tasks human-in-the-loop with AI leading; workforce to grow 4.1% by 2033", "Source_IDs": "R117", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Upskilling demand is workforce-wide, not only for engineers, so programs need role-based tracks that include field and operations staff.", "Partners": "Deloitte", "Original_date": "2025-05", "Related_IDs": []}, {"ID": "REF-118", "Initiative_ID": "CTX-REF-118", "Title": "Energy CEOs: only 40% actively reskilling for AI", "Date": "2025-11", "Finding": "KPMG's 2025 Global Energy, Natural Resources and Chemicals CEO Outlook found that CEOs are confident in AI's potential but lag on workforce action. 80% agree their leadership understands AI's disruptive potential, but only 40% are actively reskilling or upskilling the roles AI affects. 82% believe AI can support emissions reduction and energy efficiency.", "Metric": "40% actively reskilling AI-affected roles; 80% say leadership understands AI disruption; 82% see AI supporting emissions reduction", "Source_IDs": "R118", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "CEO awareness has outrun action on skills, so position enablement as how the AI strategy gets executed.", "Partners": "KPMG", "Original_date": "2025-11", "Related_IDs": []}, {"ID": "REF-119", "Initiative_ID": "CTX-REF-119", "Title": "Only 12% of O&G firms use AI in live operations", "Date": "2024-10", "Finding": "DNV's 2024 Energy Industry Insights research found that oil and gas AI adoption is still low even though 62% say digitalization is central to strategy. 47% were at the planning or pilot stage, 12% used AI in live operations, and only 3% had advanced, highly integrated use. DNV says trust is what the industry needs to move beyond this low base.", "Metric": "12% live ops; 3% advanced/integrated; 47% planning/pilot; 62% digitalization central to strategy; 20% cite regulatory barriers; AI-in-O&G market ~$3B (2024) to $5.2B (2029, Mordor)", "Source_IDs": "R119", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Assurance (showing that AI is safe, secure and validated) is itself an adoption lever in a safety-critical industry.", "Partners": "DNV, Mordor Intelligence", "Original_date": "2024-10", "Related_IDs": []}, {"ID": "REF-120", "Initiative_ID": "CTX-REF-120", "Title": "60% plan AI deployment; human-AI interaction a key risk", "Date": "2025-09", "Finding": "At SPE Offshore Europe 2025, DNV reported that 60% of 1,160 surveyed energy companies planned to deploy AI-driven applications in 2025, up from 47% in 2024. It proposed a systemic safety framework that combines safety engineering with AI assurance across the lifecycle, and it flagged human-AI interaction as a primary risk area.", "Metric": "60% planning AI deployment in 2025 (vs 47% in 2024); 1,160 companies surveyed", "Source_IDs": "R120", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Teach how to supervise AI (when to override, how to verify), not just how to use it.", "Partners": "DNV, SPE", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "REF-121", "Initiative_ID": "CTX-REF-121", "Title": "65% of energy firms see cyber as top risk", "Date": "2025-01", "Finding": "DNV's Energy Cyber Priority 2025 surveyed 375 energy professionals. 65% say their leadership views cybersecurity as the greatest risk to the business, and 71% say their organization is more vulnerable to OT cyber events than ever. 57% admit their OT defences lag their IT defences, and the report names embracing AI and embedding new workforce skills as core challenges.", "Metric": "65% cyber = greatest risk; 71% more OT-vulnerable (vs 64% 2023); 57% OT defences lag IT; 67% expect more OT security investment; n=375", "Source_IDs": "R121", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Any agentic or OT-adjacent AI program has to be designed with OT security from the start; cyber is a gating objection.", "Partners": "DNV Cyber", "Original_date": "2025-01", "Related_IDs": []}, {"ID": "REF-122", "Initiative_ID": "CTX-REF-122", "Title": "IEA: oil and gas an early AI adopter", "Date": "2025-04", "Finding": "The IEA's Energy and AI special report describes oil and gas as an early AI adopter in exploration, production, maintenance, safety and methane detection. Oil and gas companies ran 24 of the world's 500 fastest supercomputers in 2024, up from 11 in 2000, and their computing capacity grew about 70% a year. The IEA names missing data access, digital infrastructure, skills, and digital and physical security concerns as the main barriers.", "Metric": "O&G supercomputers in TOP500: 11 (2000) to 24 (2024); ~70%/yr compute capacity growth", "Source_IDs": "R122", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Even the IEA ranks skills alongside data and security as a core barrier, which supports treating enablement as infrastructure.", "Partners": "IEA", "Original_date": "2025-04", "Related_IDs": []}, {"ID": "REF-123", "Initiative_ID": "CTX-REF-123", "Title": "20% of companies capture 74% of AI value", "Date": "2026-04", "Finding": "PwC's 2026 AI Performance Study of 1,217 senior executives across 25 sectors found that 20% of organizations capture about three-quarters of AI's economic value. The rest are largely stuck in pilot mode. Leaders use AI for growth and business-model reinvention rather than productivity alone. The study is cross-industry, not specific to oil and gas.", "Metric": "74% of value to 20% of firms; n=1,217 across 25 sectors", "Source_IDs": "R123", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "AI value is winner-take-most, so the case for enablement is competitive positioning, not just efficiency.", "Partners": "PwC", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-124", "Initiative_ID": "CTX-REF-124", "Title": "Reservoir engineer AI use rises to 56%", "Date": "2025-06", "Finding": "Novi Labs surveyed more than 150 US reservoir engineers and found AI adoption rose from 43% in 2024 to 56% in 2025. Well forecasting use rose from 58% to 75%. 84% reported problems with the quality and accessibility of upstream public data, and teams using proprietary data reported much higher success than those relying only on public sources.", "Metric": "Adoption 43% to 56%; well-forecasting use 58% to 75%; 84% cite public-data quality issues; 73% vs 41% success (proprietary vs public data); hybrid build/buy success 72%; confidence in type curves 39% to 23%", "Source_IDs": "R124", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Engineers now own data quality (74%), so enablement should teach data stewardship alongside AI tools.", "Partners": "Novi Labs", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-125", "Initiative_ID": "CTX-REF-125", "Title": "Shell's DIY citizen developer program passes 4,000 builders", "Date": "2023-06", "Finding": "Shell's DIY (Do IT Yourself) program uses Microsoft Power Platform to let non-IT staff build low-code apps and automations, with a dedicated center of excellence in digital engineering. The original target was 500 trained developers, and the program now has more than 4,000 active citizen developers. It is supported by embedded coaches, bootcamps, persona-based learning paths, hackathons and gamification.", "Metric": "Target 500 to >4,000 active citizen developers; lifting-approval app cut approval time ~60% (2.5 hrs to 1 hr); furnace app covers 20 units x 64 burners", "Source_IDs": "R125", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Shell's playbook (CoE, embedded coaches, persona-based paths, hackathons) transfers directly from low-code to GenAI and agent building.", "Partners": "Microsoft (Power Platform)", "Original_date": "2023-06", "Related_IDs": []}, {"ID": "REF-126", "Initiative_ID": "AI-131", "Title": "Aker BP: Copilot for all staff plus champion network", "Date": "2025-06", "Finding": "Aker BP started with 300 Copilot licences and committed to giving every employee one in 2025. It created an AI adoption champion network with accountability at executive-committee level. Training includes help-desk support, AI-fundamentals learning paths and hands-on prompt-engineering sessions, and 'Digital and AI competence' became one of five core competencies for all employees. Staff and small teams are building their own agents, such as knowledge agents.", "Metric": "300 initial licences to all employees; 96% active users over four weeks (Viva dashboard); AI competence is 1 of 5 core competencies", "Source_IDs": "R126", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Putting AI competence into the formal competency model and making executives accountable for champions is what moves usage to 96%.", "Partners": "Microsoft", "Original_date": "2025-06", "Related_IDs": []}, {"ID": "REF-127", "Initiative_ID": "CTX-REF-127", "Title": "SPE Forum: AI in Upstream E&P 2030", "Date": "2026-04", "Finding": "SPE held a closed forum in Santa Fe (April 7-10, 2026), following its 2024 AI Forum, on how AI is reshaping upstream engineering. Sessions covered which skills preserve the human edge, whether AI can acquire tacit knowledge, how to build trust in models, and how to design human-AI teaming and ethical AI use by 2030.", "Metric": "No quantified metric supplied", "Source_IDs": "R127", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "The profession's own agenda is about trust, tacit knowledge and human-AI teaming, so enablement content should address these as well as tools.", "Partners": "SPE", "Original_date": "2026-04", "Related_IDs": []}, {"ID": "REF-128", "Initiative_ID": "CTX-REF-128", "Title": "OSDU pushes an AI-ready open data platform", "Date": "2025-12", "Finding": "At The Open Group Houston Summit 2025, the OSDU Forum focused heavily on AI and analytics and on making the open, vendor-neutral data platform AI-ready by simplifying data consumption. Deliverables included a community implementation that encodes the standard, a data model for seismic interpretation workflows, and first steps toward certification.", "Metric": "No quantified metric supplied", "Source_IDs": "R128", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Subsurface AI enablement depends on standardized data foundations such as OSDU; data literacy belongs in the curriculum.", "Partners": "The Open Group, OSDU Forum", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-129", "Initiative_ID": "CTX-REF-129", "Title": "Chevron Digital Scholars earn Rice data-science master's", "Date": "2021-08", "Finding": "Chevron sponsored employees as Digital Scholars to complete a full-time Master of Computer Science at Rice University, concentrating in data science and machine learning. The first cohort had 21 scholars. Scholars completed D2K Lab capstones on real Chevron problems, such as modeling reservoir productivity decline, and returned as digital change agents.", "Metric": "First cohort: 21 scholars; multiple cohorts completed", "Source_IDs": "R129", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Deep, degree-level programs for a small cadre of domain experts complement broad AI literacy and create in-house translators.", "Partners": "Chevron, Rice University (D2K Lab)", "Original_date": "2021-08", "Related_IDs": []}, {"ID": "REF-130", "Initiative_ID": "CTX-REF-130", "Title": "Genesis Mission targets AI for subsurface energy", "Date": "2026-03", "Finding": "DOE's Genesis Mission, launched by executive order in November 2025, aims to double US science and engineering productivity by bringing AI into all 17 national labs. One of its 26 challenges applies AI to model subsurface environments for cost-effective development of unconventional oil and gas, geothermal and coalbed methane. NETL, DOE's hydrocarbon lab, is participating.", "Metric": "26 challenges; 17 national labs; >$5B announced for Genesis (White House, Jul 2026)", "Source_IDs": "R130", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Federal subsurface AI programs are a source of open models, datasets and training partners for US operators.", "Partners": "DOE, NETL, national laboratories", "Original_date": "2026-03", "Related_IDs": []}, {"ID": "REF-131", "Initiative_ID": "CTX-REF-131", "Title": "NSTA funds AI chatbot for North Sea data", "Date": "2025-10", "Finding": "The UK regulator received £107,000 from the Regulators' Pioneer Fund for an AI-supported data discovery project. It includes an enhanced geospatial viewer and an AI chatbot to help users navigate NSTA Open Data and the National Data Repository. The regulator holds more than a petabyte of data, and the workstreams were due for assessment in April 2026.", "Metric": "£107K grant; >1 PB of data held; assessment Apr 2026", "Source_IDs": "R131", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Regulators are building AI access layers over public subsurface data; enablement programs should teach staff to use them.", "Partners": "NSTA, Regulators' Pioneer Fund", "Original_date": "2025-10", "Related_IDs": []}, {"ID": "REF-132", "Initiative_ID": "CTX-REF-132", "Title": "Norway regulator flags AI black-box and data risks", "Date": "2024-03", "Finding": "Havtil said AI use on the Norwegian shelf is early-stage, with testing in condition monitoring, maintenance planning, autonomous cranes, automated drilling and ICT security. It warned about poor training data, models that miss rare events, the black-box problem, and workers relying on screens instead of field observation. Havtil says companies must own the risk, and a DNV follow-up for Havtil (2025) piloted an ontology-grounded AI tool for dropped-object incident data that still needed extensive human review.", "Metric": "No quantified metric supplied", "Source_IDs": "R132", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Regulators expect companies to own AI risk, so enablement needs explainability, verification habits and escalation rules.", "Partners": "Havtil, DNV, AIICE JIP", "Original_date": "2024-03", "Related_IDs": []}, {"ID": "REF-133", "Initiative_ID": "AI-037", "Title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "Date": "2025-01", "Finding": "Aramco's CEO said technology delivered $4B of value in 2024, double the 2023 figure, with AI embedded across operations. The company has trained hundreds of AI specialists and more than 6,000 AI-enabled professionals, and it has developed 300+ AI use cases. It is also building the METABRAIN industrial LLM on 90 years of company data. Separately, in late 2025 Aramco agreed to take a significant minority stake in HUMAIN, the PIF-backed national AI company.", "Metric": "$4B tech value 2024 (vs $2B 2023); 6,000+ AI-enabled professionals; 300+ use cases; METABRAIN LLM", "Source_IDs": "R133", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "National programs link NOC workforce training to sovereign AI infrastructure; expect the same pattern in other Gulf NOCs.", "Partners": "Aramco, Aramco Digital, HUMAIN, PIF", "Original_date": "2025-01", "Related_IDs": ["AI-038"]}, {"ID": "REF-134", "Initiative_ID": "CTX-REF-134", "Title": "Why O&G AI pilots fail to scale", "Date": "2026-08", "Finding": "An ADNOC Offshore reservoir engineer writing in JPT cites the finding that about 95% of GenAI pilots fail to produce measurable business value. He attributes this to curated pilot data that fails in production, organizational misalignment and skill gaps. He argues that a hybrid physics-informed model that engineers trust will scale further than a more elegant model that sits unused.", "Metric": "~95% of GenAI pilots show no measurable value; 54% believe workforce lacks skills; 48% of traditional energy workforce over 45; 57% say OT defences lag IT", "Source_IDs": "R134", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Scaling is a governance and change-management problem, so budget for data plumbing and trust-building, not just models.", "Partners": "SPE, ADNOC Offshore", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-135", "Initiative_ID": "CTX-REF-135", "Title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "Date": "2026-08", "Finding": "A peer-reviewed Nature study reported by Inside Climate News estimates that AI tools that make extraction faster and cheaper add 0.47-1.8 Gt of CO2 a year. That is 1.2-4.8% of 2024 global energy-related emissions and up to 13 times data-center emissions. The authors call these enabled emissions and want them recognized, measured and governed.", "Metric": "0.47-1.8 GtCO2/yr; 1.2-4.8% of energy emissions; up to 13x data-center emissions; +470B bbl potentially accessible", "Source_IDs": "R135", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Expect ESG scrutiny of O&G AI programs; framing that covers emissions, safety and methane use cases reduces reputational risk.", "Partners": "Enabled Emissions Campaign, Microsoft, Google, Chevron, IEA", "Original_date": "2026-08", "Related_IDs": []}, {"ID": "REF-136", "Initiative_ID": "CTX-REF-136", "Title": "Investors and ex-staff challenge Microsoft's O&G AI work", "Date": "2025-12", "Finding": "Former Microsoft employees who quit in 2024 founded Enabled Emissions. At the December 2025 AGM, a shareholder proposal backed by As You Sow challenged Microsoft's sale of AI products to oil and gas producers. Critics note that Microsoft holds about 60% of O&G cloud contracts and that Google stopped building custom upstream AI in 2020.", "Metric": "~60% of O&G cloud contracts (2022 study); XTO target +50K boe/d cited", "Source_IDs": "R136", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "Vendor reputational pressure can shift tool availability and terms, so avoid single-vendor lock-in in enablement stacks.", "Partners": "Microsoft, ExxonMobil (XTO), Chevron, Shell, As You Sow, Google", "Original_date": "2025-12", "Related_IDs": []}, {"ID": "REF-137", "Initiative_ID": "CTX-REF-137", "Title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "Date": "2025-09", "Finding": "ConocoPhillips said it will cut 20-25% of its roughly 13,000-person workforce under a restructuring called Competitive Edge, advised by BCG, citing rising per-barrel costs. The cuts were not attributed to AI, but they came alongside similar reductions at Chevron and bp. EY's surveys find that employee job-security concerns inhibit AI adoption and experimentation.", "Metric": "20-25% cut (~2,600-3,250 roles); ~13,000 headcount; costs $13/bbl vs $11 in 2021", "Source_IDs": "R137", "Evidence": "Reference pending review", "Origin": "User-supplied reference", "Supplied_interpretation": "When layoffs coincide with AI rollouts, enablement programs need explicit messaging that AI is for augmentation, or adoption will stall.", "Partners": "ConocoPhillips, BCG", "Original_date": "2025-09", "Related_IDs": []}, {"ID": "ADD-AI-132", "Initiative_ID": "AI-132", "Title": "Norwegian AI Companion", "Date": "2026-04-02", "Finding": "NAIC answers Norwegian Continental Shelf technical questions using more than 2.5 million pages.", "Metric": "2.5m+ pages", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}, {"ID": "ADD-AI-133", "Initiative_ID": "AI-133", "Title": "Agent-based field development", "Date": "2026-04-02", "Finding": "OMV is testing agents with Stanford and TerraAI for Norwegian low-carbon field development.", "Metric": "No quantified return disclosed", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}, {"ID": "ADD-AI-134", "Initiative_ID": "AI-134", "Title": "iDEA well engineering", "Date": "2026-04-02", "Finding": "iDEA links drilling program design to global analogues and lessons learned.", "Metric": "No quantified return disclosed", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}, {"ID": "ADD-AI-135", "Initiative_ID": "AI-135", "Title": "Pump-failure prediction", "Date": "2026-04-02", "Finding": "A Romanian ML pilot analyzes pump histories to anticipate well failures.", "Metric": "Up to 90 days advance warning", "Source_IDs": "R037", "Evidence": "Primary source text reviewed", "Origin": "Primary-source expansion"}, {"ID": "FU-DVN-01", "Initiative_ID": "AI-006", "Title": "Issuer confirmation of Smart Gas Lift deployment", "Date": "2026-05-06", "Finding": "Physics-informed AI calculates gas-lift injection settings and controls wells in a closed loop. Management reported more than 850 Delaware wells operating, with ESP and rod-pump models still in pilot.", "Metric": "2025 gas-lift pilot: approximately 2–3% production uplift; 850+ Delaware wells live by May 2026 (management-reported).", "Source_IDs": "FU001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "followup-ai-200", "Initiative_ID": "AI-200", "Title": "Neuro autonomous geosteering", "Date": "2024-12-09", "Finding": "AI interprets live downhole measurements and changes the well trajectory.", "Metric": "25 autonomous trajectory changes across a 2,392-ft lateral.", "Source_IDs": "FU002", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "followup-ai-201", "Initiative_ID": "AI-201", "Title": "Kaybob integrated AI operations and lift control", "Date": "2026-04-01", "Finding": "IOCaaS links SCADA and historian data to surveillance and bounded gas-lift actions.", "Metric": "5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.", "Source_IDs": "FU003", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "followup-ai-202", "Initiative_ID": "AI-202", "Title": "Montney AI surveillance and flow assurance", "Date": "2026-04-01", "Finding": "IOCaaS monitors lift, hydrate risk and compressor condition.", "Metric": "3–4% production above forecast after four months of operation; about 5% lower LOE.", "Source_IDs": "FU003", "Evidence": "Provider-authored source reviewed", "Origin": "Opportunity research follow-up"}, {"ID": "M29-D-AI-203", "Initiative_ID": "AI-203", "Title": "AI-assisted pipeline right-of-way surveillance", "Date": "2025-04-09", "Finding": "Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.", "Metric": "No quantified detection, incident reduction, or cost result disclosed.", "Source_IDs": "M29-001", "Evidence": "Full company page opened; article text directly accessible.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-204", "Initiative_ID": "AI-204", "Title": "Group-wide AI rollout with market-monitoring and supply-chain assistants", "Date": "2025-07-08", "Finding": "Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.", "Metric": "More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.", "Source_IDs": "M29-002", "Evidence": "Full English company announcement opened; exact application wording accessible.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-205", "Initiative_ID": "AI-205", "Title": "AI use in emissions reporting, logistics and reservoir-management decisions", "Date": "2024-11", "Finding": "Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.", "Metric": "None disclosed.", "Source_IDs": "M29-003", "Evidence": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-206", "Initiative_ID": "AI-206", "Title": "AI/ML-assisted continuous cyber monitoring and response", "Date": "2025-08-06", "Finding": "Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.", "Metric": "No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.", "Source_IDs": "M29-004", "Evidence": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-207", "Initiative_ID": "AI-207", "Title": "Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit", "Date": "2025-11-11", "Finding": "Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.", "Metric": "Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.", "Source_IDs": "M29-005", "Evidence": "Full primary announcement reviewed.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-208", "Initiative_ID": "AI-208", "Title": "AI extraction of gas-plant design data and equipment relationships", "Date": "2025-11-05", "Finding": "Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.", "Metric": "Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.", "Source_IDs": "M29-006", "Evidence": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-209", "Initiative_ID": "AI-209", "Title": "PROTEAN predictive rotating-equipment analytics", "Date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "Finding": "Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.", "Metric": ">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.", "Source_IDs": "RS017", "Evidence": "Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-210", "Initiative_ID": "AI-210", "Title": "OneDEC producer-statement automation and contract repository for revenue accounting and settlements", "Date": "2026-05-28", "Finding": "Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.", "Metric": "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.", "Source_IDs": "M29-007", "Evidence": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-211", "Initiative_ID": "AI-211", "Title": "AI-enabled maritime voyage optimization across controlled fleet", "Date": "2025-07-09", "Finding": "AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.", "Metric": "No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.", "Source_IDs": "M29-008", "Evidence": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-212", "Initiative_ID": "AI-212", "Title": "Machine-learning gas-lift injection-rate optimization in Midland Basin production operations", "Date": "2025-11-21", "Finding": "A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.", "Metric": "Management reports incremental production without a numerical uplift, baseline or measurement period.", "Source_IDs": "M29-009", "Evidence": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-165", "Initiative_ID": "AI-165", "Title": "AI sourcing optimization across enterprise procurement (implementation-detail enrichment)", "Date": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "Finding": "Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.", "Metric": "Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.", "Source_IDs": "RS084", "Evidence": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-125", "Initiative_ID": "AI-125", "Title": "AI-supported Texas Railroad Commission monthly W-10/G-10 filings (detail enrichment; regulatory workflow with accounting adjacency)", "Date": "2026-03-05", "Finding": "Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.", "Metric": "One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.", "Source_IDs": "M29-010", "Evidence": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "Origin": "29 September 2026 gap review"}, {"ID": "M29-D-AI-143", "Initiative_ID": "AI-143", "Title": "TriCipta AI third JDA: surface equipment optimization", "Date": "2026-07-08", "Finding": "Third joint development agreement extends the TriCipta partnership into AI for upstream surface equipment optimization, targeting production, maintenance decisions and asset value.", "Metric": "No operating or economic metric for surface equipment optimization disclosed. Prior-JDA AI.SEEK and GEB deployment is reported without quantified result.", "Source_IDs": "M29-011", "Evidence": "Full issuer media release reviewed.", "Origin": "29 September 2026 gap review"}, {"ID": "SW2-D-AI-213", "Initiative_ID": "AI-213", "Title": "LLM-assisted physical gas trade capture from chat into ETRM", "Date": "2025-09-18", "Finding": "An LLM converts abbreviated trader chat into structured physical-gas trade fields for booking in Koch’s ETRM system.", "Metric": "Koch reports 25 traders, more than 35,000 trades through the platform, and approximately 75% less trade-input time.", "Source_IDs": "SW2-001; SW2-002", "Evidence": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-214", "Initiative_ID": "AI-214", "Title": "Intelligent Oilfield steam-injection optimization with Quantum", "Date": "2025-05-02", "Finding": "An AI-based thermal-management tool supports lower steam-injection rates in California heavy-oil operations.", "Metric": "For 2023, the case study reports more than 3 million barrels less steam injection and over 120,000 metric tons less Scope 1 GHG.", "Source_IDs": "RS076", "Evidence": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-215", "Initiative_ID": "AI-215", "Title": "AI-assisted prioritization of lease-operator visits in routeless operations", "Date": "2025-08-07", "Finding": "Automation, field sensors and AI prioritize lease-operator routes; management reports routeless operations across part of onshore production.", "Metric": "Approximately 40% of onshore production covered; this is a production-share denominator, not a share of wells or personnel.", "Source_IDs": "SW2-003", "Evidence": "Full 17-page company-hosted transcript reviewed.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-153", "Initiative_ID": "AI-153", "Title": "AI Center of Excellence and 2025 project portfolio", "Date": "2026-03-19", "Finding": "The proxy reports 20 AI projects developed in 2025. Separately, it describes reservoir-characterization AI tools integrated with geoscience platforms; their inclusion in the count and link to the Center of Excellence are not specified.", "Metric": "20 projects; each assigned $10–90 million in estimated annual benefit. No realized portfolio total disclosed.", "Source_IDs": "SW2-004", "Evidence": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-216", "Initiative_ID": "AI-216", "Title": "Real-time AI process optimization across gas plants", "Date": "2026", "Finding": "AI-driven optimizers allocate feed gas, reduce energy consumption and improve product recovery across multiple plants.", "Metric": "More than US$20 million in value delivered to date; value composition and cumulative period unstated.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-217", "Initiative_ID": "AI-217", "Title": "Neuron 5 asset-integrity and process-optimization platform", "Date": "2026", "Finding": "A centralized AI platform combines predictive analytics and first-principles models to detect anomalies and support gas-plant reliability and process optimization.", "Metric": "No Neuron 5-specific numerical result disclosed in the reviewed report section.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-218", "Initiative_ID": "AI-218", "Title": "Cantilever AI-enabled inspection management", "Date": "2026", "Finding": "An AI-enabled inspection platform supports non-intrusive inspections, forecasting and drone-enabled monitoring.", "Metric": "Safety, accuracy and efficiency improvements described without numerical results.", "Source_IDs": "SW2-005", "Evidence": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-219", "Initiative_ID": "AI-219", "Title": "Machine learning for well configurations integrated with simulation", "Date": "2025-11", "Finding": "Machine-learning well-configuration analysis is integrated with simulation systems.", "Metric": "R$350 million captured; period and definition of captured value unstated.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-220", "Initiative_ID": "AI-220", "Title": "AI predictive monitoring and integrated operations control rooms", "Date": "2025-11", "Finding": "AI enhances predictive monitoring and integrated operations control rooms (COI).", "Metric": "8.3 million BOE/year of losses avoided, described as approximately three days of production per year.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-221", "Initiative_ID": "AI-221", "Title": "Machine-learning revenue forecasting for the domestic market", "Date": "2025-11", "Finding": "Advanced machine-learning algorithms forecast domestic-market revenue.", "Metric": "51% greater forecast accuracy; accuracy measure and comparison baseline unspecified.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}, {"ID": "SW2-D-AI-146", "Initiative_ID": "AI-146", "Title": "AI industrial-flare control and Smart Tocha monitoring", "Date": "2025-11", "Finding": "Petrobras reports AI control of industrial flares in real time; earlier corporate material separately describes Smart Tocha automatic steam-flow adjustment.", "Metric": "10,000 tCO2/year reduction attributed to real-time AI industrial-flare control.", "Source_IDs": "SW2-006", "Evidence": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "Origin": "Second September research swarm"}, {"ID": "SW3-D-AI-222", "Initiative_ID": "AI-222", "Title": "Energy Optimizer for liquids-pipeline power decisions", "Date": "2024-10-07", "Finding": "AI supplies operational insights to control-center staff choosing how to transport liquids with less power.", "Metric": "No quantified energy, cost or emissions result.", "Source_IDs": "SW3-001", "Evidence": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "Origin": "Third September research swarm"}, {"ID": "SW3-D-AI-223", "Initiative_ID": "AI-223", "Title": "AI-assisted aerial right-of-way threat monitoring", "Date": "2024-10-07", "Finding": "AI helps review aerial pipeline-surveillance data for potentially damaging activity, hazardous conditions and loss of containment.", "Metric": "Qualitative detection and response improvements only.", "Source_IDs": "SW3-001", "Evidence": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "Origin": "Third September research swarm"}, {"ID": "SW3-D-AI-224", "Initiative_ID": "AI-224", "Title": "Companywide use of Copilot and approved AI tools", "Date": "2026-06-01", "Finding": "Employees use Copilot and other approved AI tools to reduce routine work and support decisions.", "Metric": "More than 90% employee adoption; active-use threshold and observation period unspecified.", "Source_IDs": "SW3-002", "Evidence": "Original operator article read in full; date from article URL; adoption definition limitation retained.", "Origin": "Third September research swarm"}, {"ID": "SW3-D-AI-225", "Initiative_ID": "AI-225", "Title": "AI/ML data-physics water-injection optimization across Heera assets", "Date": "2025-08-07", "Finding": "Use AI/ML data-physics study recommendations to optimize water-injection rates: pilot in Heera, implement recommendations in South Heera and later in North and Mid Heera wells, adjusting rates for field performance and regularly monitoring production-injection performance.", "Metric": "No quantified production gain, recovery uplift, injection reduction, accuracy, cost saving, or ROI disclosed.", "Source_IDs": "SW3-003", "Evidence": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "Origin": "Third September research swarm"}, {"ID": "SW3-D-AI-226", "Initiative_ID": "AI-226", "Title": "Real-Time Operations Center AI drilling oversight across ADNOC's 120+ rig fleet", "Date": "2026-08-04", "Finding": "Live rig data is consolidated into dashboards and AI insights for drilling teams to identify risks and coordinate decisions.", "Metric": "ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.", "Source_IDs": "SW3-004", "Evidence": "Full primary release opened and read; release body reviewed in full.", "Origin": "Third September research swarm"}, {"ID": "SW3-D-AI-227", "Initiative_ID": "AI-227", "Title": "Harvey legal AI selection for land, transactional, litigation and compliance work", "Date": "2025-08-14", "Finding": "Devon announced adoption of Harvey's legal AI platform to support its in-house team's litigation, land, transactional and compliance functions. The announcement does not specify individual tasks or workflows within those functions.", "Metric": "No workflow performance, usage or adoption metric reported.", "Source_IDs": "SW3-005", "Evidence": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "Origin": "Third September research swarm"}, {"ID": "SW4-D-AI-228", "Initiative_ID": "AI-228", "Title": "Visbreaker stability and sulphur prediction at La Rábida", "Date": "2026-02-26", "Finding": "AI models predict outlet stability and sulphur content in the La Rábida Visbreaker unit.", "Metric": "No accuracy, yield or financial result quantified.", "Source_IDs": "SW4-001", "Evidence": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-229", "Initiative_ID": "AI-229", "Title": "Real-time diesel and kerosene blending-property estimation at San Roque", "Date": "2026-02-26", "Finding": "AI estimates diesel and kerosene blending properties in real time at San Roque.", "Metric": "No accuracy, yield or financial result quantified.", "Source_IDs": "SW4-001", "Evidence": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-230", "Initiative_ID": "AI-230", "Title": "Mistral document-understanding assistant pilot for industrial centers", "Date": "2025-12-04", "Finding": "Planned industrial digital assistant to help staff understand documents and carry out business tasks.", "Metric": "No measured adoption or productivity result.", "Source_IDs": "SW4-002; SW4-001", "Evidence": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-231", "Initiative_ID": "AI-231", "Title": "FTW Technologies live AI well surveillance and production intelligence at Presidio", "Date": "2026-06", "Finding": "Presidio says a 24/7 AI surveillance tool is live in well surveillance and production intelligence and available to all operators and engineers. Users can query live production data in natural language. The underlying well-intelligence layer combines field pumper voice/text observations with well files, drilling and completion reports, production volumes, and geology. Operations staff remain central; an asset-intelligence group uses analytics to drive field efficiencies. The deck also describes machine-learning pump-by-exception routing: high-cash-flow wells receive daily visits and other wells are visited on exceptions; field crews remain responsible for action.", "Metric": "Company reports approximately 1% YTD production uplift toward a 3–5% 2026 target. No measurement cutoff, baseline, attribution method or independent validation is supplied. Separately, its pump-by-exception operating case reports 50% fewer well visits; scope is selected acquisition case studies, not necessarily the entire fleet.", "Source_IDs": "SW4-003", "Evidence": "Full SEC-hosted HTML presentation opened and reviewed, including legal disclaimer and AI slides.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-232", "Initiative_ID": "AI-232", "Title": "AI Builder flight-invoice extraction for airline GST processing", "Date": "2025-08-08", "Finding": "SLB's Flight Airline Automation Bot retrieves flight invoices from email, uses AI Builder to extract 22 key fields and tax details, validates the data, and compiles an Airline GST report for the tax team.", "Metric": "SLB reports processing more than 1,000 emails and 200 vendor invoices per month and effort equivalent to one full-time employee saved. The case separately says 22 invoice fields are extracted.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-233", "Initiative_ID": "AI-233", "Title": "AI-powered project discovery and recommendations across SLB facilities", "Date": "2025-08-08", "Finding": "SLB employees use an AI-powered project recommendation tool to find similar initiatives across global facilities, identify opportunities to consolidate work, and discover projects to replicate or join. Financial controllers validate project records; management reviews the resulting projects.", "Metric": "Within first 28 days of launch, SLB reports 800+ AI-powered matches and 150 similar projects inspiring facility initiatives. These are activity/claimed follow-on counts, not quantified savings or verified project consolidation.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-234", "Initiative_ID": "AI-234", "Title": "AI extraction of shipping documents for SLB export-duty recovery", "Date": "2025-08-08", "Finding": "SLB processes partner shipping documents for tools and parts across multiple countries to associate shipment details with SAP orders and entities and support export-duty recovery.", "Metric": "Microsoft page reports $10 million recovered in export duties; workflow description says it processes received documents. The source does not disclose the measurement period, baseline, or direct attribution method.", "Source_IDs": "SW4-004", "Evidence": "Full official Microsoft Learn page opened and reviewed.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-235", "Initiative_ID": "AI-235", "Title": "Closed-loop machine-learning gas-lift optimization across 1,300+ Permian wells", "Date": "2025-10-01", "Finding": "The well-by-well workflow remotely changes gas-lift injection-rate setpoints, performs multirate tests, automatically acquires well data, and iteratively uses the test response to select a production-optimizing setting. The abstract says a machine-learning model supplies recommendations on wells without downhole pressure gauges, while production data are quality-checked and uplift is model-tracked using productivity and steady-state tests.", "Metric": "Operator-authored SPE abstract reports about 2% average oil uplift. JPT’s October 2024 interview reports 1.87% in a three-month human-approval stage and about 2.2% in the next three-month, mostly automatic stage. These summaries are not additive; periods and cohorts are incompletely disclosed.", "Source_IDs": "SW4-005; SW4-006; SW3-006", "Evidence": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "Origin": "Fourth research swarm"}, {"ID": "SW4-D-AI-236", "Initiative_ID": "AI-236", "Title": "Machine-learning production forecasting across Dagang Oilfield blocks", "Date": "2026-02-04", "Finding": "A machine-learning production-forecast model was trialed in multiple Dagang Oilfield blocks to predict oilfield production.", "Metric": "CNPC article reports that forecast accuracy “improved by 15%” (预测准确率提升15%). The comparison baseline, calculation, validation set, forecast horizon, and whether this is a relative improvement or percentage-point change are unspecified; it is not an absolute 15% accuracy score.", "Source_IDs": "SW4-007", "Evidence": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "Origin": "Fourth research swarm"}, {"ID": "SW5-D-AI-237", "Initiative_ID": "AI-237", "Title": "Repsol legal team Harvey rollout and adoption update", "Date": "2026-03-30", "Finding": "Repsol’s in-house legal team uses Harvey to assist legal work; the original Repsol announcement listed legal-document analysis, contract/regulation comparison, clause suggestions, translation, legal queries and document summaries. The March 2026 source establishes rollout scope and adoption-support practices, but does not add new task-level workflows.", "Metric": "Harvey interview: Repsol tracks about 30 power users out of 150; approximately 6–7 meet both usage and qualitative-output definitions. These are adoption-support measures, not productivity outcomes. Repsol’s April 2024 issuer announcement estimated up to three hours per week of repetitive low-legal-content work saved per lawyer during implementation; this was an estimate, not an audited realized saving and predates the window.", "Source_IDs": "SW5-001; SW5-002", "Evidence": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-238", "Initiative_ID": "AI-238", "Title": "AI-assisted non-op revenue statement extraction and WolfePak export at Spearpoint Resources", "Date": "2026-09-05", "Finding": "Spearpoint’s small accounting team processes monthly non-operated revenue statements arriving in 27 purchaser formats. Joltly extracts structured line items at well/lease level—including production dates, volumes, prices, interests, taxes and deductions—then users review exceptions and export the results for WolfePak posting.", "Metric": "Spearpoint Controller Elle Ross estimates Joltly gives the team 20–30 hours per week back across three entities. From May through August 2026 the case reports 82 statements processed, 5,441 pages and 56,358 line items across 27 purchaser formats. Time estimate is customer-reported; no baseline time study, quality/exception rate or independent validation is supplied.", "Source_IDs": "SW5-003", "Evidence": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-239", "Initiative_ID": "AI-239", "Title": "Machine-learning surrogate accelerates INPEX CCS reservoir simulation", "Date": "2025-07-04", "Finding": "A high-speed machine-learning surrogate model provides rapid predictions of pressure, gas saturation and other reservoir parameters from geological and operating conditions, supporting reservoir-efficiency and pressure-rise evaluation plus optimization and sensitivity analysis for CCS decision-making.", "Metric": "The stated benefit is reduced simulation time and the ability to evaluate parameters quickly; no measured speedup, accuracy figure, validation result or project result is quantified.", "Source_IDs": "SW5-004; SW5-005; SW5-006", "Evidence": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-240", "Initiative_ID": "AI-240", "Title": "Ichthys AI tool searches operational reports and structures maintenance history", "Date": "2026-07-07", "Finding": "An Ichthys-specific generative-AI tool converts daily reports into a searchable database and answers natural-language questions with past operating performance and troubleshooting cases. A connected function structures maintenance history by categorizing work descriptions, failure details and affected components.", "Metric": "INPEX reports that searches formerly taking several hours now complete in seconds to minutes; the report graphic gives “seconds to minutes.” It says AI structured approximately 110,000 maintenance records and achieved over 90% classification accuracy. No accuracy metric definition, validation sample, baseline method, error rate or time-study method is provided.", "Source_IDs": "SW5-005; SW5-006", "Evidence": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-241", "Initiative_ID": "AI-241", "Title": "AI video safety alerts with human verification", "Date": "2026-05-20", "Finding": "Cameras identify selected unsafe conditions and notify trained safety personnel. Karimun examples include missing PPE and unsafe proximity to machinery.", "Metric": "11 operational sites reported for 2025. No attributable incident reduction or false-positive rate disclosed.", "Source_IDs": "SW5-007; SW5-008; SW5-009; SW5-010; SW5-011", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-242", "Initiative_ID": "AI-242", "Title": "AI-assisted drone surveys for dropped-object hazards", "Date": "2026-05-20", "Finding": "Drone imagery of elevated structures is processed by a trained AI algorithm to support dropped-object hazard surveys.", "Metric": "No measured detection accuracy, inspection productivity or avoided-loss result.", "Source_IDs": "SW5-007; SW5-010; SW5-011", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-243", "Initiative_ID": "AI-243", "Title": "Shape Reef Phase II safety-barrier and gas-leak modelling", "Date": "2025-08-13", "Finding": "Develop risk models for degraded safety barriers and gas-leak scenarios on offshore units.", "Metric": "No validated accuracy or incident-reduction result.", "Source_IDs": "SW5-012", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-244", "Initiative_ID": "AI-244", "Title": "FPSO maintenance data foundation and planned full-field AI integration", "Date": "2025-11-13", "Finding": "Unify engineering and operating data for maintenance prioritization and reliability analysis. A linked SLB alliance aims to extend insights from subsea systems through FPSO topsides.", "Metric": "No AI-attributed uptime or maintenance saving. Reported overall fleet uptime is not evidence of AI impact.", "Source_IDs": "SW5-013; SW5-014; SW5-015", "Evidence": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "Origin": "Fifth research swarm"}, {"ID": "SW5-D-AI-214", "Initiative_ID": "AI-214", "Title": "Undated operator corroboration of Intelligent Oilfield steam optimization", "Date": "Undated", "Finding": "Sentinel Peak identifies machine-learning and AI analysis of reservoir data for real-time steam management.", "Metric": "The operator estimates 200,000 metric tons CO2e per year enabled through intelligent-oilfield technologies. Period, baseline and AI-only contribution are unspecified.", "Source_IDs": "SW5-016", "Evidence": "Full original HTML page opened and read. It names Sentinel Peak's Intelligent Oilfield as an ML/AI-powered data system for real-time steam-management optimization and estimates 200,000 MT CO2e/year enabled through intelligent oilfield technologies. The page gives no publication date, calculation, baseline, or AI-only attribution split.", "Origin": "Fifth research swarm"}, {"ID": "SW6-D-AI-245", "Initiative_ID": "AI-245", "Title": "AI-assisted internal audit data review", "Date": "2026", "Finding": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "Metric": "No measured outcome disclosed. Faster review is an intended benefit.", "Source_IDs": "SW6-001; SW6-002", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}, {"ID": "SW6-D-AI-246", "Initiative_ID": "AI-246", "Title": "MOL MOVE AI customer-offer personalization", "Date": "2025", "Finding": "Personalize loyalty-program experiences and offers.", "Metric": "No AI-attributed outcome. Nearly five million app downloads describe the broader platform.", "Source_IDs": "SW6-003", "Evidence": "Original PDF opened; full relevant paragraph and surrounding pages read.", "Origin": "Sixth research swarm"}, {"ID": "SW6-D-AI-247", "Initiative_ID": "AI-247", "Title": "EVO talent-profile and role matching", "Date": "2026", "Finding": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "Metric": "No measured outcome disclosed.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}, {"ID": "SW6-D-AI-147", "Initiative_ID": "AI-147", "Title": "AI portfolio: 2025 investment and benefits disclosure", "Date": "2026", "Finding": "Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.", "Metric": "2025: US$1.88m reported AI investment, US$2.06m revenue increase and US$1.44m cost reduction (US$3.50m total benefits). Reported ROI: 7%, cumulative since 2021.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}, {"ID": "SW6-D-AI-148", "Initiative_ID": "AI-148", "Title": "CADE energy dispatch: Rubiales model development update", "Date": "2026", "Finding": "Consolidate electrical operating signals and develop dispatch recommendations for Rubiales operations.", "Metric": "No quantified AI-attributed result disclosed.", "Source_IDs": "SW6-004", "Evidence": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "Origin": "Sixth research swarm"}, {"ID": "SW7-D-AI-248", "Initiative_ID": "AI-248", "Title": "Agent-assisted drilling and completion data validation", "Date": "2025-09-29", "Finding": "UiPath interprets inconsistent well-data emails; a Databricks agent validates context and scores accuracy. Robots return errors to suppliers and deliver validated data to engineers. Award material also names WellView drilling-report integration.", "Metric": "Undated vendor case: 200+ emails/day, approximately five minutes saved per previously manual email, one reusable robot supporting 40+ processes. Undated award page: accuracy rose from about 70% to the high 90% range.", "Source_IDs": "SW7-001; SW7-002; SW7-003; SW7-004", "Evidence": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-097", "Initiative_ID": "AI-097", "Title": "AI-driven rod-lift optimization", "Date": "2026-02-25", "Finding": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "Metric": "Operator reports early approximately 25% improvement in rod-pump run times. The 99% denominator is rod-lift wells, not the entire well fleet.", "Source_IDs": "SW7-005; SW7-006; SW7-007; SW7-008", "Evidence": "Full source opened and read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-006", "Initiative_ID": "AI-006", "Title": "Autonomous artificial-lift optimization at operating wells", "Date": "2026-05-06", "Finding": "AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.", "Metric": "May 2026 CEO remarks report >850 fully autonomous lift wells without a new quantified productivity result. Earlier 2025 pilot uplift remains in the original profile and is not remeasured here.", "Source_IDs": "FU001", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-249", "Initiative_ID": "AI-249", "Title": "AI alerting from SCADA and decision-support-center data for operational / methane issues", "Date": "2025 (report; covers calendar 2024 unless noted)", "Finding": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "Metric": "No AI-attributable emissions, repair-time, uptime or financial result disclosed. The 100% number is monitoring-technology coverage, not AI coverage or emissions reduction.", "Source_IDs": "SW7-009", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-250", "Initiative_ID": "AI-250", "Title": "Hourly computer-vision checks of flare equipment function", "Date": "2025 (report; covers calendar 2024 unless noted)", "Finding": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "Metric": "No performance or emissions metric attributed to this model. Devon separately reports flaring intensity but does not attribute that outcome to AI.", "Source_IDs": "SW7-009", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-251", "Initiative_ID": "AI-251", "Title": "Veriforce machine-learning contractor incident-risk scoring", "Date": "2025 (2024 performance reporting)", "Finding": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "Metric": "No Diamondback-specific incidents avoided or safety improvement attributed to the model. Its corporate safety statistics are not attributed to Veriforce scoring.", "Source_IDs": "SW7-010", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-045", "Initiative_ID": "AI-045", "Title": "Machine learning / AI for artificial-lift choice and late-life well optimization", "Date": "2025-11-03", "Finding": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "Metric": "No result specifically attributed to this AI workflow. Nov 2025 letter states the intended gains; May 2026 letter says expected future uptime and LOE improvements. Reported corporate production and LOE are not causally linked to AI.", "Source_IDs": "SW7-011; SW7-012", "Evidence": "Original source opened; relevant passage read.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-166", "Initiative_ID": "AI-166", "Title": "Invoice AI and Snowflake-backed production-data agent rollout", "Date": "2025-04-29", "Finding": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "Metric": "Customer reports hundreds of administrative hours saved annually for one employee by the invoice application; not disaggregated to AI extraction alone. No production-agent outcome measured.", "Source_IDs": "RS085", "Evidence": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-195", "Initiative_ID": "AI-195", "Title": "Snowflake forecasting ML, drilling analytics and site alerts", "Date": "Undated supplemental case", "Finding": "Snowpark supports ML-based type-curve forecasts; Snowflake hosts drilling ML and Cortex Analyst questions. SCADA/site data produces alerts for operations-center investigation.", "Metric": "Estimated 1,000+ labor hours saved annually in one business unit; data monitoring covers 3,700 sites. Neither proves production uplift. Separately, 40% less data-related M&A time is a platform result.", "Source_IDs": "SW7-013", "Evidence": "Full official Snowflake customer case read; customer executive quotes present. Undated, vendor-authored, estimates/claims not independently verified.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-202", "Initiative_ID": "AI-202", "Title": "Montney AI surveillance, artificial lift and flow assurance", "Date": "2026-04-01", "Finding": "IOCaaS monitors Montney wells, gathering equipment and compressors for gas-lift opportunities, hydrate risk, liquid loading and equipment anomalies. Operators review alerts and act.", "Metric": "Vendor authors report the asset team calculated 3–4% production above forecast and ~5% lower LOE after four months; no hydrate outages in the evaluation. Approximately 6% future uplift is a projection.", "Source_IDs": "FU003", "Evidence": "Original full article opened/read, including the Montney deployment, architecture, timeline, outcomes, and authorship. The source gives operator/team attributions but does not provide underlying data or independent validation.", "Origin": "Seventh research swarm"}, {"ID": "SW7-D-AI-072", "Initiative_ID": "AI-072", "Title": "Guyana Neuro and DrillOps directional-drilling automation", "Date": "2026 (source URL; exact date unknown)", "Finding": "Neuro controls trajectory and downlink requests; DrillOps automates surface equipment and drilling tasks. Shore teams monitor operations.", "Metric": "SLB reports 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "Source_IDs": "SW7-014", "Evidence": "Original full page opened/read; deployment, named products, functions, 48-km path, autonomy share, and downlink metrics appear in final case-study section. SLB is the supplier and reports the performance claims.", "Origin": "Seventh research swarm"}, {"ID": "OCT4-AI-006", "Initiative_ID": "AI-006", "Title": "Autonomous artificial-lift optimization at operating wells", "Date": "2026-08-05", "Finding": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "Metric": "1,000 live wells as of August 2026; >2,000 additional wells prospective. Separate 2025 gas-lift pilot reported 2–3% uplift in May 2026; not a measured result for the full August fleet.", "Source_IDs": "SW8-001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "October 4 evidence review"}, {"ID": "OCT4-AI-007", "Initiative_ID": "AI-007", "Title": "AI agents for drilling and completions monitoring", "Date": "2026-08-05", "Finding": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "Metric": "Company reported 85% of broader $1bn program achieved", "Source_IDs": "SW8-001", "Evidence": "Issuer-hosted transcript reviewed", "Origin": "October 4 evidence review"}, {"ID": "OCT5-AI-140-1", "Initiative_ID": "AI-140", "Title": "Pearl GTL machine-vision hazard alerts and AI equipment health", "Date": "2026-05-13", "Finding": "Shell's FY2025 annual report says Pearl GTL uses machine-vision systems to detect hazards in real time and alert supervisors, while AI also monitors equipment health. It adds that the World Economic Forum's Global Lighthouse Network recognised this in September 2025.", "Metric": "", "Source_IDs": "OCT5-S009", "Source_URL": "https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section/promo.multi.stream/1779352356739/36306d968747b6079d3f800b2b1552a033856b5d/shell-integrated-annual-and-sustainability-report.pdf", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review"}, {"ID": "OCT5-AI-038-1", "Initiative_ID": "AI-038", "Title": "Qualcomm MoU for on-device edge AI over 450 MHz network", "Date": "2025-05-18", "Finding": "Aramco Digital and Qualcomm signed an MoU to explore edge AI on rugged devices, robots, drones, cameras and sensors connected through Aramco Digital's 450 MHz 5G industrial network. The register's AI-038 already lists Qualcomm as a partner; this adds the dated agreement source.", "Metric": "", "Source_IDs": "OCT5-S023", "Source_URL": "https://jpt.spe.org/aramco-signs-mous-worth-up-to-90-billion-for-lng-ai-emissions-reduction-and-asset-management-initiatives", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review"}, {"ID": "OCT5-AI-231-1", "Initiative_ID": "AI-231", "Title": "DOUG surveillance, weekend routing and plunger optimization: Q2 update", "Date": "2026-08-12", "Finding": "Presidio’s Q2 call distinguishes DOUG recommendations, weekend visit routing and automatic plunger-cycle adjustments. DOUG alerts field engineers; pumpers receive a prioritized map. Controllers on 24 wells continuously adjust cycle timing. Management says field judgment remains with staff; controller safeguards are unspecified.", "Metric": "Reported gas rates on the controller population: 2,680 to 2,946 Mcf/day, roughly 10%. Weekend production reportedly rose 2.5%. These overlap the existing 2.3% program-uplift claim and must not be added to it. Existing program revenue is not net cash.", "Source_IDs": "OCT5-S024", "Source_URL": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "Evidence": "Source reviewed 5 October 2026", "Origin": "October 5 research review", "Assessment": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage."}, {"ID": "OCT5-AI-054-1", "Initiative_ID": "AI-054", "Title": "Asset and capital planning", "Date": "2026-04-30", "Finding": "A Quantum portfolio company in the Piceance Basin uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.", "Metric": "Microsoft reports a 20,000-acre scenario took 20 minutes versus three weeks; leasehold utilization rose from 50% to 85%. Production runs evaluated 2.3m pad candidates, 3.5m well possibilities and over 50m allocations; the accumulation period is unstated.", "Source_IDs": "OCT5-S032", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-088-1", "Initiative_ID": "AI-088", "Title": "Generative AI for geology, drilling and safety", "Date": "2026 annual report covering 2025; exact publication day not confirmed", "Finding": "EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.", "Metric": "No AI-specific outcome or rollout count disclosed.", "Source_IDs": "OCT5-S033", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-089-1", "Initiative_ID": "AI-089", "Title": "ML production optimizers and remote Utica operations", "Date": "2026 annual report covering 2025; exact publication day not confirmed", "Finding": "EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary.", "Metric": "No AI-specific outcome or rollout count disclosed.", "Source_IDs": "OCT5-S034", "Source_URL": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-092-1", "Initiative_ID": "AI-092", "Title": "In-house AI for artificial lift, drilling and completions", "Date": "2026-02-23", "Finding": "Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.", "Metric": "No quantified AI-specific outcome disclosed.", "Source_IDs": "OCT5-S035", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-094-1", "Initiative_ID": "AI-094", "Title": "AI tunes frac-design factors in real time", "Date": "2026-02-23", "Finding": "Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.", "Metric": "No quantified AI-specific outcome disclosed.", "Source_IDs": "OCT5-S036", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-171-1", "Initiative_ID": "AI-171", "Title": "Pipeline IQ engineering-document verification agents", "Date": "Undated source; see chronology", "Finding": "PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.", "Metric": "AWS reports 85% less document-review time and more than 100 PIPER users. Early tests on documents up to 120 pages reported 95–100% accuracy. Over $18.7m in maintenance capital avoidance is expected over five years for PIPER and SCOUT combined; it is a forecast.", "Source_IDs": "OCT5-S037", "Source_URL": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-280-1", "Initiative_ID": "AI-280", "Title": "Dated media reference to the ALS assistant case", "Date": "2025-07-17", "Finding": "EPAM’s 17 July 2025 media summary links the undated original client case and repeats the pilot evaluation and planned integration. It supports an in-window public account; it does not establish the go-live date or independent verification.", "Metric": "", "Source_IDs": "OCT5-S039", "Source_URL": "https://www.epam.com/about/newsroom/in-the-news/2025/how-ai-accelerates-oil-production-and-saves-thousands-of-hours-of-support", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-146-1", "Initiative_ID": "AI-146", "Title": "P-52 flare-reduction application: platform scope added", "Date": "2026-05", "Finding": "Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.", "Metric": "", "Source_IDs": "OCT5-S040", "Source_URL": "https://transparencia.petrobras.com.br/documents/41711919/42223019/Relatorio%2Bde%2BGestao%2B-%2B2025.pdf/9ef35500-1f38-0cb6-539d-072b0569514f?download=true&t=1779977455000&version=1.0", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-195-1", "Initiative_ID": "AI-195", "Title": "Snowflake architecture: operating drilling ML versus dashboard development", "Date": "Undated; reviewed 2026-10-06", "Finding": "The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.", "Metric": "", "Source_IDs": "OCT5-S041", "Source_URL": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-217-1", "Initiative_ID": "AI-217", "Title": "CPAD predecessor: historical deployment and benefit boundaries", "Date": "2025 (2024 report; exact publication day not stated)", "Finding": "The issuer reports US$29m saved through performance/reliability improvement and more than US$90m avoided costs through early prediction. GenAI integration and expansion to static equipment are roadmap work.", "Metric": "US$29m reported savings and >US$90m avoided costs. These have no reconciled overlap or common measurement period.", "Source_IDs": "OCT5-S042", "Source_URL": "https://www.adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas-integrated-report-2024-english.ashx", "Evidence": "Source reviewed 6 October 2026", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-037-1", "Initiative_ID": "AI-037", "Title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "Date": "2026-03-10", "Finding": "Aramco reports SAR 19.9bn ($5.3bn) FY2025 Technology Realized Value: SAR 9.8bn ($2.6bn) attributed to AI-driven solutions and SAR 10.1bn ($2.7bn) to non-AI technologies. The report defines TRV as incremental financial benefit versus baseline from capital or operating expenditure avoided or revenue increase. This replaces the less precise CEO-call description “over half.”", "Metric": "Company-reported FY2025 aggregate: $5.3bn total TRV, including $2.6bn AI-driven and $2.7bn non-AI-driven. The separate FY2025 results presentation supplies the total-portfolio benefit mix; no AI-only allocation by benefit type is given.", "Source_IDs": "OCT5-S043", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-037-2", "Initiative_ID": "AI-037", "Title": "Saudi Aramco FY2025 results presentation — TRV value composition", "Date": "2026-03-10", "Finding": "The FY2025 results material breaks the $5.3bn company-wide TRV into $3.9bn capex avoidance, $1.0bn opex reduction and $0.4bn revenue enhancement. Its “around half from AI solutions” statement is portfolio-level; it does not allocate the category composition to the $2.6bn AI subtotal.", "Metric": "Portfolio-wide FY2025 TRV mix: $3.9bn capex avoidance + $1.0bn opex reduction + $0.4bn revenue enhancement = $5.3bn total. No AI-only value-type allocation.", "Source_IDs": "OCT5-S044", "Source_URL": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "Evidence": "Official operator FY2025 results presentation slide 10 independently reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-077-1", "Initiative_ID": "AI-077", "Title": "TotalEnergies and Mistral announce a three-year frontier subsurface AI program", "Date": "2026-09-15", "Finding": "TotalEnergies and Mistral announced a three-year joint program representing more than €100m investment to develop frontier AI models for exploration, reservoir characterization and development. A joint scientific laboratory is planned. This duplicates linked disclosure REF-021 in the current record; update the existing case summary only.", "Metric": "Announced joint-program investment >€100m over three years; no deployed-model outcome or return reported.", "Source_IDs": "OCT5-S045", "Source_URL": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-078-1", "Initiative_ID": "AI-078", "Title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "Date": "2026-02-11", "Finding": "Management says a signed AspenTech contract will deploy Inmation for real-time collection across 40 upstream and 16 refining/chemical sites; integrated power was already done. Cognite will transform, enrich, store and expose the data. Management aims to multiply asset data points by ten, with full deployment targeted by end-2027; a later AI layer is intended to improve production and availability. At least 500 engineers by 2027 is a critical-mass target for all three programs (digital HSE, digital plant and Integrated Power modelling).", "Metric": "Planned data-layer scope: 40 upstream + 16 refining/chemical sites; 10x data-point objective; end-2027 deployment target. India competency-center target: ≥500 engineers by 2027 across programs.", "Source_IDs": "OCT5-S046", "Source_URL": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-216-1", "Initiative_ID": "AI-216", "Title": "ADNOC Gas Q4/FY2025 Results Presentation", "Date": "2026-02-09", "Finding": "The FY2025 achievements table places “Closed Loop RTO deployed in Ruwais” and “$50m savings realized” within the AI-driven production-optimization row (separate 2026–30 opportunity: up to $300m). Existing AI-216 cites the 2025 Integrated Report claim of more than $20m value delivered to date across multiple plants. The presentation does not isolate RTO’s contribution or reconcile the two scopes.", "Metric": "Company-reported row-level $50m savings realization; period, baseline and calculation perimeter unstated. Separate existing claim: >$20m value to date across multiple plants; do not add or replace without reconciliation.", "Source_IDs": "OCT5-S047", "Source_URL": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-218-1", "Initiative_ID": "AI-218", "Title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "Date": "2025-11-02", "Finding": "ADNOC Gas announced a three-year AI/robotics collaboration with AIQ and Gecko Robotics using Cantilever to digitize inspection-to-decision. Year 1 is joint deployment/data modeling, years 2–3 scaled access and continuous optimization. The issuer forecasts over $300m maintenance and inspection cost savings over five years. ADNOC Gas says it had already used Gecko robots in 2023; the new element is the AI/Cantilever program. A 2026 results slide later says “Cantiliver GECKO - in all assets,” without asset denominator or completion definition.", "Metric": "Issuer forecast: >$300m maintenance/inspection cost savings over five years under a three-year phased program. No realized savings reported.", "Source_IDs": "OCT5-S048", "Source_URL": "https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-022-1", "Initiative_ID": "AI-022", "Title": "Introducing Vantage: A new lens on upstream operations", "Date": "2025-10-10", "Finding": "ExxonMobil says Vantage monitors >22,000 U.S. sites, manages/optimizes >1m b/d oil, and field/remote operations production optimization realized >29,000 b/d uplift. The article separately reports >20m cubic feet of flaring eliminated in 2024 and >6,000 driving hours avoided in 2024. These describe Vantage/field operations; the source does not attribute the outcomes to AI.", "Metric": "Company-reported operations context: >22,000 sites monitored; >1m b/d managed/optimized; >29,000 b/d uplift from field/remote optimization (period not stated); >20m cubic feet flaring eliminated and >6,000 driving hours avoided in 2024.", "Source_IDs": "OCT5-S049", "Source_URL": "https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-238-1", "Initiative_ID": "AI-238", "Title": "Joltly Pricing — Operators & Non-Op / Royalty Owners", "Date": "Undated; accessed 2026-10-06", "Finding": "Joltly describes a monthly platform fee plus usage tied to documents, ACH payments, mailed checks and enabled workflows, with no seat fees. Pricing cards show $750 Independent Operator starter and $1,500 Midsize Operator under a “Monthly Contracts” label. Integrations include direct QuickBooks and Quorum On-Demand connections and file exports for PakEnergy and Integra. The vendor claims AP can go live in under one week and revenue in about two weeks with a sample statement.", "Metric": "Public pricing cards: $750 and $1,500 monthly plan amounts; included volume/modules and usage costs are undefined. These are not Spearpoint contract terms.", "Source_IDs": "OCT5-S053", "Source_URL": "https://www.joltly.io/pricing", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT5-AI-238-2", "Initiative_ID": "AI-238", "Title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "Date": "2026-08-08", "Finding": "Joltly describes extraction of line-item JIBs and purchaser revenue statements, well/account mapping, exception flagging, line-level accounting output, deposit matching and posting revenue by well/product. This is vendor workflow explanation, not evidence of Spearpoint-specific control implementation.", "Metric": "No customer-specific additional result in this workflow article.", "Source_IDs": "OCT5-S054", "Source_URL": "https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas", "Evidence": "Relevant original primary source reviewed on 2026-10-06.", "Origin": "October 6 research review"}, {"ID": "OCT9-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "Date": "2026", "Finding": "Quantum Energy Cloud connects portfolio financial, operational and market data using Microsoft and Databricks. The 2026 report identifies Claude Enterprise as the firm interface succeeding QIQ.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "OCT9-AI-053", "Initiative_ID": "AI-053", "Title": "Claude Enterprise succeeds Quantum IQ", "Date": "2026", "Finding": "The 2026 sponsor report identifies Claude Enterprise as the current firm generative AI platform, succeeding Quantum IQ.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "OCT9-AI-054", "Initiative_ID": "AI-054", "Title": "QB Energy Automated Pad Development with Quantum", "Date": "2026", "Finding": "The 2026 Quantum report names QB Energy behind the Piceance planning case. Earlier Microsoft reporting left the operator unnamed.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "OCT9-AI-276", "Initiative_ID": "AI-276", "Title": "Vickery Energy StrataSource AI land-title research with Quantum", "Date": "2026", "Finding": "The 2026 Quantum report names Vickery Energy and StrataSource AI, used in West Virginia leasing. Earlier Microsoft reporting left the operator unnamed.", "Metric": "See source-reported detail in 9 October supplementary notes.", "Source_IDs": "OCT9-QCG-2026", "Source_URL": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "Evidence": "Primary sponsor report reviewed 9 October 2026; exact publication day unknown.", "Origin": "9 October company follow-up"}, {"ID": "OCT10-AI-231-0817", "Initiative_ID": "AI-231", "Title": "August presentation: achieved uplift, platform scope and 2026 roadmap", "Date": "2026-08-17", "Finding": "The August deck repeats 2.3% YTD AI-linked production uplift and a 3–5% 2026 target. It describes live surveillance and natural-language production queries. Its goal slide cites roughly 2,300 platform wells; fleet-wide agentic orchestration, budget agents and automatic acquisition screening appear in the 2026 deployment column, not the NOW column.", "Metric": "August deck: 2.3% YTD achieved, company-reported; 3–5% 2026 production-uplift target. Roughly 2,300 wells appears on the intelligence-platform GOAL slide; coverage date and active intervention population are not established. $2.5m annual revenue per 1% uplift at $30/boe is illustrative, not a realized return.", "Source_IDs": "OCT10-PRESIDIO-0817", "Source_URL": "https://ir.bypresidio.com/sec-filings/all-sec-filings/content/0001213900-26-090532/ea030229001ex99-1.htm", "Evidence": "Relevant original issuer presentation text reviewed 10 October 2026; results remain issuer-reported.", "Origin": "10 October integration review", "Assessment": "Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs."}, {"ID": "OCT10-AI-284", "Initiative_ID": "AI-284", "Title": "Daily lift recommendations across a 441-well Permian pilot", "Date": "2026", "Finding": "Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.", "Metric": "Authors report average 6% incremental oil on wells that adopted recommendations, plus more than 20% lower average field gas-lift use. The number of adopting wells is not disclosed.", "Source_IDs": "OCT10-AI-284-SWPSC", "Source_URL": "https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established."}, {"ID": "OCT10-AI-285", "Initiative_ID": "AI-285", "Title": "Pump-by-exception monitoring and control at a West Texas lease", "Date": "2026", "Finding": "Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.", "Metric": "Over eight months, authors report 43% lower electricity consumption and daily strokes; electricity expense of $660 to $376 per well-month; routine field time of 126 to about 18 hours per month (86% lower); and 26% lower hot-oiling cost. Production stayed around 19 BOPD versus an expected 4.25% annual decline. Authors calculate about $65,000 incremental profit at $65/bbl WTI and capital recovery in slightly over one month.", "Source_IDs": "OCT10-AI-285-SWPSC; OCT10-BURK-PDF", "Source_URL": "https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established."}, {"ID": "OCT10-AI-286", "Initiative_ID": "AI-286", "Title": "Autonomous liquid unloading on nine Haynesville gas wells", "Date": "2026-07-01", "Finding": "An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.", "Metric": "SLB’s JPT account reports 70–139% cumulative gas-production gains over 63–83 days on nine wells across eight pads, with daily gains up to 350 Mscf/d. More than 80 MMscf additional gas per well-year is an estimate, not a year of observed production.", "Source_IDs": "OCT10-AI-286-SWPSC; OCT10-EXPAND-JPT", "Source_URL": "https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments", "Evidence": "Original technical sources reviewed; operator/vendor-reported results", "Origin": "10 October 2026 field-case review", "Assessment": "Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established."}, {"ID": "OCT10-AI-089-UTICA", "Initiative_ID": "AI-089", "Title": "Utica production optimizer: related technology result, not AI attribution", "Date": "2026-08-04", "Finding": "EOG’s August 2026 deck reports 5% improved Utica base performance from its production-optimizer suite after the Encino acquisition. The deck does not attribute that result specifically to AI or machine learning.", "Metric": "5% improved Utica base performance; issuer-reported operating-technology result, not counted as an AI result.", "Source_IDs": "OCT10-EOG-UTICA", "Source_URL": "https://investors.eogresources.com/image/EOG_0826.pdf", "Evidence": "Original issuer deck reviewed; AI attribution not established", "Origin": "10 October 2026 field-case review", "Assessment": "Related operating-technology context only. The deck does not define base performance or disclose the well count, measurement period, method, costs or counterfactual. The annual report’s description of machine-learning tools does not establish AI causality for this 5% result."}, {"ID": "OCT12-FLY-PLATFORM-DISC", "Initiative_ID": "AI-287", "Title": "Existing internal AI app platform", "Date": "2026", "Finding": "Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-PLATFORM", "Source_URL": "https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750", "Evidence": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-HIRING-DISC", "Initiative_ID": "AI-287", "Title": "Land and accounting AI roles: intended automation work", "Date": "2026", "Finding": "Separately, Flywheel advertises land and accounting AI analyst roles. The accounting role specifies AI-assisted workflows, document extraction, API/data-pipeline integration, acquisition-data normalization and migration, retries, fallbacks and exception queues. It calls for human review of ownership or legal determinations. These are staffing responsibilities, not confirmation that every workflow is deployed. OpenAI/Azure OpenAI are examples of desired candidate experience, not confirmed vendors.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-ACCOUNTING; OCT12-FLY-FEED", "Source_URL": "https://www.linkedin.com/jobs/view/ai-analyst-at-flywheel-energy-llc-4472009913", "Evidence": "Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-CONTEXT-DISC", "Initiative_ID": "AI-287", "Title": "Quorum land software: adjacent technology context", "Date": "Undated", "Finding": "Quorum names Flywheel as an On Demand Land customer for more than five years. Its testimonial describes easier data movement and access for nontechnical users. Quorum’s newer AI features are general product capabilities; Flywheel-specific use is not established.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-QUORUM", "Source_URL": "https://www.quorumsoftware.com/solutions/upstream-on-demand/land/", "Evidence": "Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT12-FLY-INTERN-DISC", "Initiative_ID": "AI-287", "Title": "Summer 2027 prototype agenda: indexed hiring lead", "Date": "2026", "Finding": "A search-indexed Summer 2027 internship description proposes LLM, agentic and MCP integration prototypes, plus ML work on production forecasting and predictive maintenance. The direct posting expired/redirected during review; these prospective tasks are not operating field AI.", "Metric": "No quantified AI result disclosed.", "Source_IDs": "OCT12-FLY-INTERN", "Source_URL": "https://www.linkedin.com/jobs/view/ai-intern-summer-2027-at-flywheel-energy-llc-4461575049", "Evidence": "Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source.", "Origin": "12 October 2026 Flywheel research review", "Assessment": "The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."}, {"ID": "OCT14-AI-288", "Initiative_ID": "AI-288", "Title": "Safety cameras that flag unsafe conditions", "Date": "Undated", "Finding": "CNX says it uses AI video analytics to identify unsafe conditions, coach employees in real time and inform training. HSE++ also analyzes earlier incidents and safety observations to find trends. Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "Metric": "No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.", "Source_IDs": "OCT14-CNX-SAFETY", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "Evidence": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 targeted review", "Assessment": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch."}, {"ID": "OCT14-AI-289", "Initiative_ID": "AI-289", "Title": "Agents for risk assessment and alarm reviews", "Date": "2026-05-21", "Finding": "Cognite and ABB name Aker BP as the first customer signed on to implement an agentic offering for risk assessment and alarm rationalization. The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "Metric": "No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.", "Source_IDs": "OCT14-AKER-ABB", "Source_URL": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "Evidence": "Relevant original source text reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 targeted review", "Assessment": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed."}, {"ID": "OCT14B-AI-290", "Initiative_ID": "AI-290", "Title": "A model that diagnoses rod-pumped wells in Xinjiang", "Date": "2026-04-15", "Finding": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions. A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "Metric": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "Source_IDs": "OCT14B-CNPC-XINJIANG", "Source_URL": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain."}, {"ID": "OCT14B-AI-291", "Initiative_ID": "AI-291", "Title": "Finding methane-release precursors and prioritizing repairs", "Date": "Undated", "Finding": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory. Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "Metric": "CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.", "Source_IDs": "OCT14B-CNX-METHANE", "Source_URL": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority."}, {"ID": "OCT14B-AI-292", "Initiative_ID": "AI-292", "Title": "An invoice assistant with human exception review", "Date": "Undated", "Finding": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review. Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "Metric": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "Source_IDs": "OCT14B-HG-INVOICES", "Source_URL": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October 2026 second targeted review", "Assessment": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata."}, {"ID": "OCT14B-SANTOS-72", "Initiative_ID": "AI-139", "Title": "Production AI, virtual pressure measurement and drilling controls", "Date": "2026-05-26", "Finding": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "Metric": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "Source_IDs": "RS012", "Source_URL": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "Evidence": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "Origin": "14 October 2026 original slide review", "Assessment": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology."}, {"ID": "OCT14C-AI-293", "Initiative_ID": "AI-293", "Title": "Connecting rock samples to a geological knowledge platform", "Date": "2026-08-25", "Finding": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists. Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "Metric": "No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.", "Source_IDs": "OCT14C-ARAMCO-CORE", "Source_URL": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established."}, {"ID": "OCT14C-AI-072", "Initiative_ID": "AI-072", "Title": "Autonomous drilling in Guyana, with supplier-reported results", "Date": "Undated (2026 URL)", "Finding": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work. Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "Metric": "SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "Source_IDs": "SW7-014", "Source_URL": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained."}, {"ID": "OCT14C-AI-148", "Initiative_ID": "AI-148", "Title": "An operating energy desk, with a dispatch model still being built", "Date": "2026 (2025 annual report)", "Finding": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI. A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.", "Metric": "No measured energy, emissions, reliability, cost or production benefit is disclosed for these AI components.", "Source_IDs": "OCT14C-ECOPETROL-CADE", "Source_URL": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given."}, {"ID": "OCT14C-AI-143", "Initiative_ID": "AI-143", "Title": "Deployed exploration tools and a separate equipment-development agreement", "Date": "2026-07-08", "Finding": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools. The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "Metric": "No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.", "Source_IDs": "OCT14C-PETRONAS-TRICIPTA", "Source_URL": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed."}, {"ID": "OCT14C-AI-097", "Initiative_ID": "AI-097", "Title": "Rod-lift AI sits inside a broader artificial-lift improvement program", "Date": "2026-02-25", "Finding": "Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register. The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.", "Metric": "The broader program reports over 50% improvement in ESP replacement cycle times, over 25% improvement in failure rates since early 2025, and about 1,200 fewer workover rig days in 2025. These are not separately isolated ML effects.", "Source_IDs": "OCT14C-CHORD-LIFT", "Source_URL": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "Third October 14 targeted review", "Assessment": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted."}, {"ID": "WATCH14PM-AI-294", "Initiative_ID": "AI-294", "Title": "Employees build agents for research, rules and management meetings", "Date": "2026-09-01", "Finding": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions. SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "Metric": "Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.", "Source_IDs": "WATCH14PM-INPEX", "Source_URL": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method."}, {"ID": "WATCH14PM-AI-224", "Initiative_ID": "AI-224", "Title": "A field technician turns equipment reports into turbine alerts", "Date": "2026-09-09", "Finding": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration. The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "Metric": "More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.", "Source_IDs": "WATCH14PM-WILLIAMS", "Source_URL": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share."}, {"ID": "WATCH14PM-AI-158", "Initiative_ID": "AI-158", "Title": "Egyptian drilling teams expand an edge-based warning system", "Date": "2026-03-11", "Finding": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings. Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "Metric": "NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.", "Source_IDs": "WATCH14PM-NOV", "Source_URL": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "Evidence": "Relevant original source reviewed 14 October 2026; claims remain source-reported.", "Origin": "14 October evening research watch", "Assessment": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established."}, {"ID": "WATCH15AM-CNOOC", "Initiative_ID": "AI-145", "Title": "One oilfield connects reservoir, production and equipment data", "Date": "2026-04-09", "Finding": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans. More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.", "Metric": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.", "Source_IDs": "WATCH15AM-CNOOC", "Source_URL": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "Evidence": "Original company ESG report reviewed", "Origin": "15 October morning watch; existing program expanded", "Assessment": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved."}, {"ID": "WATCH17AM-AI-279", "Initiative_ID": "AI-279", "Title": "CNRL puts an annual savings figure on hydrogen optimization", "Date": "2025-11-07", "Finding": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations. The presentation names the task but does not explain the model, inputs, vendor or approval process.", "Metric": "The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.", "Source_IDs": "WATCH17AM-CNRL", "Source_URL": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "Evidence": "Original operator source read; results are not independently validated", "Origin": "17 October morning source review", "Assessment": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate."}, {"ID": "WATCH17AM-AI-061", "Initiative_ID": "AI-061", "Title": "Governed productivity and decision-support pilots", "Date": "2026-02-18", "Finding": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots. It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.", "Metric": "No measured outcome, cost or ROI disclosed.", "Source_IDs": "S40", "Source_URL": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "Evidence": "Original operator source read; results are not independently validated", "Origin": "17 October morning source review", "Assessment": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established."}, {"ID": "WATCH18AM-AI-295", "Initiative_ID": "AI-295", "Title": "SANAD helps drilling engineers find the procedure they need", "Date": "2026-09-29", "Finding": "KOC describes SANAD, a drilling knowledge assistant developed with SLB and piloted by its North & West Kuwait Drilling Group. The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision.", "Metric": "A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.", "Source_IDs": "WATCH18AM-KOC-DIGEST; WATCH18AM-KOC-INTERVIEW", "Source_URL": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "Evidence": "Original KOC publication read; later attributed engineer interview reviewed. Outcomes not independently validated.", "Origin": "18 October 2026 targeted source review", "Assessment": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085."}, {"ID": "WATCH18PM-AI-133", "Initiative_ID": "AI-133", "Title": "OMV tests field-development decisions with analytical agents", "Date": "2026-04-14", "Finding": "OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI. Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "Metric": "OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "Source_IDs": "WATCH18PM-OMV", "Source_URL": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "Evidence": "Original OMV article reviewed; early internal claims not independently validated", "Origin": "18 October evening source review", "Assessment": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132)."}, {"ID": "WATCH20UP-AI-016-BP", "Initiative_ID": "AI-016", "Title": "Bumerangue’s appraisal contract includes drilling automation", "Date": "2026-08-25", "Finding": "Halliburton’s August contract for bp’s Bumerangue appraisal campaign includes LOGIX automation and remote operations offshore Brazil. The integrated drilling and evaluation scope connects planning, execution and evaluation. The release says LOGIX will be deployed; it does not identify an LLM, individual control permissions or human override arrangements.", "Metric": "No number of wells, completed automated wells, operating comparator, quantified benefit, contract value or net ROI disclosed.", "Source_IDs": "WATCH20UP-BP", "Source_URL": "https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil", "Evidence": "Original supplier release read; planned deployment", "Origin": "20 October user-requested source review", "Assessment": "A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform."}, {"ID": "WATCH20PM-AI-296", "Initiative_ID": "AI-296", "Title": "Twelve wells moved from recommendations to automatic adjustments", "Date": "2026-07", "Finding": "U.S. Energy and ROAM-AI describe automatic ESP adjustments on 12 Permian wells. Engineers define the operating limits and handle exceptions. Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed.", "Metric": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "Source_IDs": "WATCH20PM-USEDC", "Source_URL": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "Evidence": "Original operator/vendor account read; results not independently validated", "Origin": "20 October evening Luna research; editor original-source review", "Assessment": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case."}, {"ID": "WATCH20PM-AI-148", "Initiative_ID": "AI-148", "Title": "Rubiales adds a dated operating footprint to CADE", "Date": "2026-05-04", "Finding": "Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.", "Metric": "The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.", "Source_IDs": "WATCH20PM-CADE", "Source_URL": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "Evidence": "Original operator announcement read; power coverage is not AI performance", "Origin": "20 October evening Luna research; editor original-source review", "Assessment": "No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population."}, {"ID": "WATCH21AM-AI-297", "Initiative_ID": "AI-297", "Title": "PTTEP describes an in-house engine for specialist agents", "Date": "2025-11-03", "Finding": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live.", "Metric": "No X.brain user count, adoption rate or measured return is disclosed. The same article separately reports LogSight interpretation falling from two days to one hour and DORA drilling savings described only as “millions of US dollars annually”. Those are product-specific claims, not demonstrated X.brain results.", "Source_IDs": "WATCH21AM-PTTEP", "Source_URL": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "Evidence": "Accessible original reviewed; issuer/vendor claims not independently validated", "Origin": "21 October morning Luna research; editor source and attribution review", "Assessment": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons."}, {"ID": "WATCH21AM-AI-054", "Initiative_ID": "AI-054", "Title": "Quantum’s planning gain includes a separate algorithm improvement", "Date": "2026-04-30", "Finding": "Microsoft says algorithm changes in an updated version of Quantum’s field-planning tool reduced solve time by 50–65%. The integer-linear-programming optimizer selects feasible pads and wells. A Copilot Studio agent explains outputs. The numerical improvement belongs to the optimization algorithm; the source does not attribute it to the language model.", "Metric": "50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.", "Source_IDs": "OCT5-S032", "Source_URL": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "Evidence": "Accessible original reviewed; issuer/vendor claims not independently validated", "Origin": "21 October morning Luna research; editor source and attribution review", "Assessment": "Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model."}, {"ID": "WATCH21EX-AI-086", "Initiative_ID": "AI-086", "Title": "Devon confirms ChatDVN is part of daily work", "Date": "2026-05-06", "Finding": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.", "Metric": "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.", "Source_IDs": "FU001", "Source_URL": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "Evidence": "Issuer-hosted corrected call transcript, relevant original page read", "Origin": "21 October additional Luna review; editor primary-source admission", "Assessment": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status."}, {"ID": "WATCH21PM-AI-052", "Initiative_ID": "AI-052", "Title": "Quantum adds a relational layer to its deal-evaluation data", "Date": "Unknown", "Finding": "Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.", "Metric": "Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.", "Source_IDs": "WATCH21PM-DBX", "Source_URL": "https://www.databricks.com/customers/quantum-capital-group", "Evidence": "Original vendor customer story reviewed; date unknown; data foundation, not proof of AI outcomes", "Origin": "21 October evening Luna research; editor original-source review", "Assessment": "No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution."}, {"ID": "WATCH22PM-AI-298", "Initiative_ID": "AI-298", "Title": "Vista uses AI monitoring to adjust completion designs", "Date": "2025-11-12", "Finding": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025.", "Metric": "Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.", "Source_IDs": "WATCH22PM-VISTA; WATCH22PM-VISTA-SEC", "Source_URL": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "Evidence": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted."}, {"ID": "WATCH22PM-AI-299", "Initiative_ID": "AI-299", "Title": "Saipem installs a predictive-maintenance pilot on one drillship", "Date": "2026-02-23", "Finding": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests.", "Metric": "No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.", "Source_IDs": "WATCH22PM-SAIPEM", "Source_URL": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "Evidence": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed."}, {"ID": "WATCH22PM-AI-300", "Initiative_ID": "AI-300", "Title": "Worley puts four internal tools into engineering and bid work", "Date": "2025-08-27", "Finding": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest.", "Metric": "Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.", "Source_IDs": "WATCH22PM-WORLEY", "Source_URL": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "Evidence": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established."}, {"ID": "WATCH22PM-AI-199", "Initiative_ID": "AI-199", "Title": "Technip Energies reports progress, without proving 55 AI deployments", "Date": "2026 (exact publication day unverified)", "Finding": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established.", "Metric": "The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.", "Source_IDs": "WATCH22PM-TEN", "Source_URL": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "Evidence": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "Origin": "22 October evening Luna research; editor original-source admission", "Assessment": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved."}, {"ID": "WATCH23AM-AI-301", "Initiative_ID": "AI-301", "Title": "Azule connects the well plan to drilling automation", "Date": "2026-10-23", "Finding": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola.", "Metric": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "Source_IDs": "WATCH23AM-SLB-Q3; WATCH23AM-AZULE", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate."}, {"ID": "WATCH23AM-AI-027", "Initiative_ID": "AI-027", "Title": "Eni adds a concrete on-premise subsurface workflow", "Date": "2026-10-23", "Finding": "SLB says Eni expanded petrophysics and geomechanics AI workflows using an on-premise Lumi platform.", "Metric": "No task-level benefit, asset population or full cost quantified in this update. Eni’s existing portfolio metrics and targets remain separately sourced; none is allocated to Lumi.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications."}, {"ID": "WATCH23AM-AI-302", "Initiative_ID": "AI-302", "Title": "BSP awards an on-premise software contract", "Date": "2026-10-23", "Finding": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work.", "Metric": "No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout."}, {"ID": "WATCH23AM-AI-226", "Initiative_ID": "AI-226", "Title": "SLB repeats ADNOC’s rig-center scope, without a new rollout count", "Date": "2026-10-23", "Finding": "SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs.", "Metric": "No additional rollout count or newly measured benefit. Earlier ADNOC claims retain their own source, period and limitations.", "Source_IDs": "WATCH23AM-SLB-Q3", "Source_URL": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "Evidence": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "Origin": "23 October morning bounded Luna review plus editor fresh earnings-release check", "Assessment": "Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet."}, {"ID": "WATCH23PM-AI-303", "Initiative_ID": "AI-303", "Title": "AI checks supplier documents. Petrobras staff make the decision.", "Date": "Unknown", "Finding": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them.", "Metric": "No accuracy, workload, completed-case count, cost, saving or ROI reported.", "Source_IDs": "WATCH23PM-PETROBRAS", "Source_URL": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "Evidence": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "Origin": "23 October evening bounded Luna research; editor original-source review", "Assessment": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established."}, {"ID": "WATCH24AM-AI-304", "Initiative_ID": "AI-304", "Title": "A maintenance alert only helps if a crew can act on it.", "Date": "2026-10-22", "Finding": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs.", "Metric": "The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.", "Source_IDs": "WATCH24AM-HAL-RELIABILITY; WATCH24AM-HAL-PRODUCT", "Source_URL": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "Evidence": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "Origin": "24 October morning editor original-source check alongside three bounded Luna tracks", "Assessment": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016."}, {"ID": "SLBTH-AI-009", "Initiative_ID": "AI-009", "Title": "What holds up in SLB’s data and AI investment thesis?", "Date": "2026-10-23", "Finding": "SLB has a credible route to earning more from AI: it can put specialized models into software and operating workflows that customers already use. The public evidence supports a platform, engineering and governed-access advantage. It does not support treating every customer’s downhole data as an unrestricted SLB-owned training pool. Menon describes a seismic model trained on public and SLB multiclient datasets, with customers able to fine-tune for their own assets. Lumi connects governed data; domain models handle technical formats; Tela provides interaction and workflow orchestration. SLB describes subscription upgrades, metered usage, tailored workflows and marketplace revenue sharing. These are commercial mechanisms, not disclosed AI-only earnings.", "Metric": "FY2025 Digital revenue was $2.660bn and adjusted EBITDA $928m (34.9% margin). Q3 2026 Digital revenue was $693m, adjusted EBITDA $228m and ARR $1.04bn. These are division-wide figures, not AI revenue or customer ROI. The CFO’s 2030 $1.8–2.0bn adjusted EBITDA and approximately $2bn ARR are ambitions, not results.", "Source_IDs": "SLBTH-Q3; SLBTH-FY25; SLBTH-MENON; SLBTH-BIGUET; SLBTH-TERMS; SLBTH-GENAI; SLBTH-MULLIN; SLBTH-RANNEN", "Source_URL": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "Evidence": "Original investor scripts, financial tables and public terms read", "Origin": "24 October editor SLB thesis review; no new initiative", "Assessment": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures."}, {"ID": "WATCH24PM-AI-305", "Initiative_ID": "AI-305", "Title": "PDO has moved ESP prediction from pilot to operating use", "Date": "2026-04-21", "Finding": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework.", "Metric": "No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.", "Source_IDs": "WATCH24PM-WFRD-Q1", "Source_URL": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "Evidence": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "Origin": "24 October evening editor original-source check alongside three bounded Luna tracks", "Assessment": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case."}, {"ID": "WATCH25AM-AI-306", "Initiative_ID": "AI-306", "Title": "PTTEP reports faster purchase-order processing with X.Bot", "Date": "Unknown", "Finding": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing.", "Metric": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "Source_IDs": "WATCH25AM-PTTEP-XBOT; WATCH25AM-PTTEP-OPEX", "Source_URL": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "Evidence": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "Origin": "25 October morning editor original-source check alongside three bounded Luna tracks", "Assessment": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot."}], "crosswalk": [{"Input_ID": "AI-001", "Origin": "Original register", "Canonical_ID": "AI-001", "Disposition": "Retained original profile"}, {"Input_ID": "AI-002", "Origin": "Original register", "Canonical_ID": "AI-002", "Disposition": "Retained original profile"}, {"Input_ID": "AI-003", "Origin": "Original register", "Canonical_ID": "AI-003", "Disposition": "Retained original profile"}, {"Input_ID": "AI-004", "Origin": "Original register", "Canonical_ID": "AI-004", "Disposition": "Retained original profile"}, {"Input_ID": "AI-005", "Origin": "Original register", "Canonical_ID": "AI-005", "Disposition": "Retained original profile"}, {"Input_ID": "AI-006", "Origin": "Original register", "Canonical_ID": "AI-006", "Disposition": "Retained original profile"}, {"Input_ID": "AI-007", "Origin": "Original register", "Canonical_ID": "AI-007", "Disposition": "Retained original profile"}, {"Input_ID": "AI-008", "Origin": "Original register", "Canonical_ID": "AI-008", "Disposition": "Retained original profile"}, {"Input_ID": "AI-009", "Origin": "Original register", "Canonical_ID": "AI-009", "Disposition": "Retained original profile"}, {"Input_ID": "AI-010", "Origin": "Original register", "Canonical_ID": "AI-010", "Disposition": "Retained original profile"}, {"Input_ID": "AI-011", "Origin": "Original register", "Canonical_ID": "AI-011", "Disposition": "Retained original profile"}, {"Input_ID": "AI-012", "Origin": "Original register", "Canonical_ID": "AI-012", "Disposition": "Retained original profile"}, {"Input_ID": "AI-013", "Origin": "Original register", "Canonical_ID": "AI-013", "Disposition": "Retained original profile"}, {"Input_ID": "AI-014", "Origin": "Original register", "Canonical_ID": "AI-014", "Disposition": "Retained original profile"}, {"Input_ID": "AI-015", "Origin": "Original register", "Canonical_ID": "AI-015", "Disposition": "Retained original profile"}, {"Input_ID": "AI-016", "Origin": "Original register", "Canonical_ID": "AI-016", "Disposition": "Retained original profile"}, {"Input_ID": "AI-017", "Origin": "Original register", "Canonical_ID": "AI-017", "Disposition": "Retained original profile"}, {"Input_ID": "AI-018", "Origin": "Original register", "Canonical_ID": "AI-018", "Disposition": "Retained original profile"}, {"Input_ID": "AI-019", "Origin": "Original register", "Canonical_ID": "AI-019", "Disposition": "Retained original profile"}, {"Input_ID": "AI-020", "Origin": "Original register", "Canonical_ID": "AI-020", "Disposition": "Retained original profile"}, {"Input_ID": "AI-021", "Origin": "Original register", "Canonical_ID": "AI-021", "Disposition": "Retained original profile"}, {"Input_ID": "AI-022", "Origin": "Original register", "Canonical_ID": "AI-022", "Disposition": "Retained original profile"}, {"Input_ID": "AI-023", "Origin": "Original register", "Canonical_ID": "AI-023", "Disposition": "Retained original profile"}, {"Input_ID": "AI-024", "Origin": "Original register", "Canonical_ID": "AI-024", "Disposition": "Retained original profile"}, {"Input_ID": "AI-025", "Origin": "Original register", "Canonical_ID": "AI-025", "Disposition": "Retained original profile"}, {"Input_ID": "AI-026", "Origin": "Original register", "Canonical_ID": "AI-026", "Disposition": "Retained original profile"}, {"Input_ID": "AI-027", "Origin": "Original register", "Canonical_ID": "AI-027", "Disposition": "Retained original profile"}, {"Input_ID": "AI-028", "Origin": "Original register", "Canonical_ID": "AI-028", "Disposition": "Retained original profile"}, {"Input_ID": "AI-029", "Origin": "Original register", "Canonical_ID": "AI-029", "Disposition": "Retained original profile"}, {"Input_ID": "AI-030", "Origin": "Original register", "Canonical_ID": "AI-030", "Disposition": "Retained original profile"}, {"Input_ID": "AI-031", "Origin": "Original register", "Canonical_ID": "AI-031", "Disposition": "Retained original profile"}, {"Input_ID": "AI-032", "Origin": "Original register", "Canonical_ID": "AI-032", "Disposition": "Retained original profile"}, {"Input_ID": "AI-033", "Origin": "Original register", "Canonical_ID": "AI-033", "Disposition": "Retained original profile"}, {"Input_ID": "AI-034", "Origin": "Original register", "Canonical_ID": "AI-034", "Disposition": "Retained original profile"}, {"Input_ID": "AI-035", "Origin": "Original register", "Canonical_ID": "AI-035", "Disposition": "Retained original profile"}, {"Input_ID": "AI-036", "Origin": "Original register", "Canonical_ID": "AI-036", "Disposition": "Retained original profile"}, {"Input_ID": "AI-037", "Origin": "Original register", "Canonical_ID": "AI-037", "Disposition": "Retained original profile"}, {"Input_ID": "AI-038", "Origin": "Original register", "Canonical_ID": "AI-038", "Disposition": "Retained original profile"}, {"Input_ID": "AI-039", "Origin": "Original register", "Canonical_ID": "AI-039", "Disposition": "Retained original profile"}, {"Input_ID": "AI-040", "Origin": "Original register", "Canonical_ID": "AI-040", "Disposition": "Retained original profile"}, {"Input_ID": "AI-041", "Origin": "Original register", "Canonical_ID": "AI-041", "Disposition": "Retained original profile"}, {"Input_ID": "AI-042", "Origin": "Original register", "Canonical_ID": "AI-042", "Disposition": "Retained original profile"}, {"Input_ID": "AI-043", "Origin": "Original register", "Canonical_ID": "AI-043", "Disposition": "Retained original profile"}, {"Input_ID": "AI-044", "Origin": "Original register", "Canonical_ID": "AI-044", "Disposition": "Retained original profile"}, {"Input_ID": "AI-045", "Origin": "Original register", "Canonical_ID": "AI-045", "Disposition": "Retained original profile"}, {"Input_ID": "AI-046", "Origin": "Original register", "Canonical_ID": "AI-046", "Disposition": "Retained original profile"}, {"Input_ID": "AI-047", "Origin": "Original register", "Canonical_ID": "AI-047", "Disposition": "Retained original profile"}, {"Input_ID": "AI-048", "Origin": "Original register", "Canonical_ID": "AI-048", "Disposition": "Retained original profile"}, {"Input_ID": "AI-049", "Origin": "Original register", "Canonical_ID": "AI-049", "Disposition": "Retained original profile"}, {"Input_ID": "AI-050", "Origin": "Original register", "Canonical_ID": "AI-050", "Disposition": "Retained original profile"}, {"Input_ID": "AI-051", "Origin": "Original register", "Canonical_ID": "AI-051", "Disposition": "Retained original profile"}, {"Input_ID": "AI-052", "Origin": "Original register", "Canonical_ID": "AI-052", "Disposition": "Retained original profile"}, {"Input_ID": "AI-053", "Origin": "Original register", "Canonical_ID": "AI-053", "Disposition": "Retained original profile"}, {"Input_ID": "AI-054", "Origin": "Original register", "Canonical_ID": "AI-054", "Disposition": "Retained original profile"}, {"Input_ID": "AI-055", "Origin": "Original register", "Canonical_ID": "AI-055", "Disposition": "Retained original profile"}, {"Input_ID": "AI-056", "Origin": "Original register", "Canonical_ID": "AI-056", "Disposition": "Retained original profile"}, {"Input_ID": "AI-057", "Origin": "Original register", "Canonical_ID": "AI-057", "Disposition": "Retained original profile"}, {"Input_ID": "AI-058", "Origin": "Original register", "Canonical_ID": "AI-058", "Disposition": "Retained original profile"}, {"Input_ID": "AI-059", "Origin": "Original register", "Canonical_ID": "AI-059", "Disposition": "Retained original profile"}, {"Input_ID": "AI-060", "Origin": "Original register", "Canonical_ID": "AI-060", "Disposition": "Retained original profile"}, {"Input_ID": "AI-061", "Origin": "Original register", "Canonical_ID": "AI-061", "Disposition": "Retained original profile"}, {"Input_ID": "AI-062", "Origin": "Original register", "Canonical_ID": "AI-062", "Disposition": "Retained original profile"}, {"Input_ID": "AI-063", "Origin": "Original register", "Canonical_ID": "AI-063", "Disposition": "Retained original profile"}, {"Input_ID": "AI-064", "Origin": "Original register", "Canonical_ID": "AI-064", "Disposition": "Retained original profile"}, {"Input_ID": "AI-065", "Origin": "Original register", "Canonical_ID": "AI-065", "Disposition": "Retained original profile"}, {"Input_ID": "AI-066", "Origin": "Original register", "Canonical_ID": "AI-066", "Disposition": "Retained original profile"}, {"Input_ID": "AI-067", "Origin": "Original register", "Canonical_ID": "AI-067", "Disposition": "Retained original profile"}, {"Input_ID": "AI-068", "Origin": "Original register", "Canonical_ID": "AI-068", "Disposition": "Retained original profile"}, {"Input_ID": "AI-069", "Origin": "Original register", "Canonical_ID": "AI-069", "Disposition": "Retained original profile"}, {"Input_ID": "REF-000", "Origin": "User-supplied reference", "Canonical_ID": "AI-037", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-001", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-001", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-002", "Origin": "User-supplied reference", "Canonical_ID": "AI-039", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-003", "Origin": "User-supplied reference", "Canonical_ID": "AI-070", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-004", "Origin": "User-supplied reference", "Canonical_ID": "AI-038", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-005", "Origin": "User-supplied reference", "Canonical_ID": "AI-071", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-006", "Origin": "User-supplied reference", "Canonical_ID": "AI-040", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-007", "Origin": "User-supplied reference", "Canonical_ID": "AI-040", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-008", "Origin": "User-supplied reference", "Canonical_ID": "AI-040", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-009", "Origin": "User-supplied reference", "Canonical_ID": "AI-040", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-010", "Origin": "User-supplied reference", "Canonical_ID": "AI-072", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-011", "Origin": "User-supplied reference", "Canonical_ID": "AI-073", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-012", "Origin": "User-supplied reference", "Canonical_ID": "AI-024", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-013", "Origin": "User-supplied reference", "Canonical_ID": "AI-074", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-014", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-014", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-015", "Origin": "User-supplied reference", "Canonical_ID": "AI-075", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-016", "Origin": "User-supplied reference", "Canonical_ID": "AI-076", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-017", "Origin": "User-supplied reference", "Canonical_ID": "AI-010", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-018", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-018", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-019", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-019", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-020", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-020", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-021", "Origin": "User-supplied reference", "Canonical_ID": "AI-077", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-022", "Origin": "User-supplied reference", "Canonical_ID": "AI-029", "Disposition": "Merged related disclosure", "Related_original_IDs": "AI-030"}, {"Input_ID": "REF-023", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-023", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-024", "Origin": "User-supplied reference", "Canonical_ID": "AI-078", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-025", "Origin": "User-supplied reference", "Canonical_ID": "AI-033", "Disposition": "Merged related disclosure", "Related_original_IDs": "AI-034; AI-035; AI-036"}, {"Input_ID": "REF-026", "Origin": "User-supplied reference", "Canonical_ID": "AI-027", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-027", "Origin": "User-supplied reference", "Canonical_ID": "AI-032", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-028", "Origin": "User-supplied reference", "Canonical_ID": "AI-031", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-029", "Origin": "User-supplied reference", "Canonical_ID": "AI-031", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-030", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-030", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-031", "Origin": "User-supplied reference", "Canonical_ID": "AI-079", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-032", "Origin": "User-supplied reference", "Canonical_ID": "AI-080", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-033", "Origin": "User-supplied reference", "Canonical_ID": "AI-081", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-034", "Origin": "User-supplied reference", "Canonical_ID": "AI-082", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-035", "Origin": "User-supplied reference", "Canonical_ID": "AI-041", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-036", "Origin": "User-supplied reference", "Canonical_ID": "AI-083", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-037", "Origin": "User-supplied reference", "Canonical_ID": "AI-084", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-038", "Origin": "User-supplied reference", "Canonical_ID": "AI-085", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-039", "Origin": "User-supplied reference", "Canonical_ID": "AI-086", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-040", "Origin": "User-supplied reference", "Canonical_ID": "AI-006", "Disposition": "Merged related disclosure", "Related_original_IDs": "AI-007"}, {"Input_ID": "REF-041", "Origin": "User-supplied reference", "Canonical_ID": "AI-006", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-042", "Origin": "User-supplied reference", "Canonical_ID": "AI-006", "Disposition": "Merged related disclosure", "Related_original_IDs": "AI-007"}, {"Input_ID": "REF-043", "Origin": "User-supplied reference", "Canonical_ID": "AI-007", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-044", "Origin": "User-supplied reference", "Canonical_ID": "AI-087", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-045", "Origin": "User-supplied reference", "Canonical_ID": "AI-088", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-046", "Origin": "User-supplied reference", "Canonical_ID": "AI-089", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-047", "Origin": "User-supplied reference", "Canonical_ID": "AI-045", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-048", "Origin": "User-supplied reference", "Canonical_ID": "AI-045", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-049", "Origin": "User-supplied reference", "Canonical_ID": "AI-090", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-050", "Origin": "User-supplied reference", "Canonical_ID": "AI-044", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-051", "Origin": "User-supplied reference", "Canonical_ID": "AI-044", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-052", "Origin": "User-supplied reference", "Canonical_ID": "AI-091", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-053", "Origin": "User-supplied reference", "Canonical_ID": "AI-092", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-054", "Origin": "User-supplied reference", "Canonical_ID": "AI-093", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-055", "Origin": "User-supplied reference", "Canonical_ID": "AI-094", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-056", "Origin": "User-supplied reference", "Canonical_ID": "AI-095", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-057", "Origin": "User-supplied reference", "Canonical_ID": "AI-096", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-058", "Origin": "User-supplied reference", "Canonical_ID": "AI-096", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-059", "Origin": "User-supplied reference", "Canonical_ID": "AI-097", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-060", "Origin": "User-supplied reference", "Canonical_ID": "AI-097", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-061", "Origin": "User-supplied reference", "Canonical_ID": "AI-098", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-062", "Origin": "User-supplied reference", "Canonical_ID": "AI-099", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-063", "Origin": "User-supplied reference", "Canonical_ID": "AI-100", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-064", "Origin": "User-supplied reference", "Canonical_ID": "AI-101", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-065", "Origin": "User-supplied reference", "Canonical_ID": "AI-102", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-066", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-066", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-067", "Origin": "User-supplied reference", "Canonical_ID": "AI-103", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-068", "Origin": "User-supplied reference", "Canonical_ID": "AI-104", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-069", "Origin": "User-supplied reference", "Canonical_ID": "AI-009", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-070", "Origin": "User-supplied reference", "Canonical_ID": "AI-008", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-071", "Origin": "User-supplied reference", "Canonical_ID": "AI-105", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-072", "Origin": "User-supplied reference", "Canonical_ID": "AI-047", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-073", "Origin": "User-supplied reference", "Canonical_ID": "AI-106", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-074", "Origin": "User-supplied reference", "Canonical_ID": "AI-107", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-075", "Origin": "User-supplied reference", "Canonical_ID": "AI-108", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-076", "Origin": "User-supplied reference", "Canonical_ID": "AI-109", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-077", "Origin": "User-supplied reference", "Canonical_ID": "AI-110", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-078", "Origin": "User-supplied reference", "Canonical_ID": "AI-111", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-079", "Origin": "User-supplied reference", "Canonical_ID": "AI-021", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-080", "Origin": "User-supplied reference", "Canonical_ID": "AI-020", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-081", "Origin": "User-supplied reference", "Canonical_ID": "AI-112", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-082", "Origin": "User-supplied reference", "Canonical_ID": "AI-113", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-083", "Origin": "User-supplied reference", "Canonical_ID": "AI-114", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-084", "Origin": "User-supplied reference", "Canonical_ID": "AI-115", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-085", "Origin": "User-supplied reference", "Canonical_ID": "AI-116", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-086", "Origin": "User-supplied reference", "Canonical_ID": "AI-117", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-087", "Origin": "User-supplied reference", "Canonical_ID": "AI-118", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-088", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-088", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-089", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-089", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-090", "Origin": "User-supplied reference", "Canonical_ID": "AI-119", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-091", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-091", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-092", "Origin": "User-supplied reference", "Canonical_ID": "AI-049", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-093", "Origin": "User-supplied reference", "Canonical_ID": "AI-120", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-094", "Origin": "User-supplied reference", "Canonical_ID": "AI-121", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-095", "Origin": "User-supplied reference", "Canonical_ID": "AI-122", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-096", "Origin": "User-supplied reference", "Canonical_ID": "AI-123", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-097", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-097", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-098", "Origin": "User-supplied reference", "Canonical_ID": "AI-051", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-099", "Origin": "User-supplied reference", "Canonical_ID": "AI-065", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-100", "Origin": "User-supplied reference", "Canonical_ID": "AI-065", "Disposition": "Merged related disclosure", "Related_original_IDs": ""}, {"Input_ID": "REF-101", "Origin": "User-supplied reference", "Canonical_ID": "AI-124", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-102", "Origin": "User-supplied reference", "Canonical_ID": "AI-125", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-103", "Origin": "User-supplied reference", "Canonical_ID": "AI-126", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-104", "Origin": "User-supplied reference", "Canonical_ID": "AI-127", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-105", "Origin": "User-supplied reference", "Canonical_ID": "AI-128", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-106", "Origin": "User-supplied reference", "Canonical_ID": "AI-129", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-107", "Origin": "User-supplied reference", "Canonical_ID": "AI-130", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-108", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-108", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-109", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-109", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-110", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-110", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-111", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-111", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-112", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-112", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-113", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-113", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-114", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-114", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-115", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-115", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-116", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-116", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-117", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-117", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-118", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-118", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-119", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-119", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-120", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-120", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-121", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-121", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-122", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-122", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-123", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-123", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-124", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-124", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-125", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-125", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-126", "Origin": "User-supplied reference", "Canonical_ID": "AI-131", "Disposition": "Added profile", "Related_original_IDs": ""}, {"Input_ID": "REF-127", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-127", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-128", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-128", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-129", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-129", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-130", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-130", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-131", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-131", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-132", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-132", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-133", "Origin": "User-supplied reference", "Canonical_ID": "AI-037", "Disposition": "Merged related disclosure", "Related_original_IDs": "AI-038"}, {"Input_ID": "REF-134", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-134", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-135", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-135", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-136", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-136", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}, {"Input_ID": "REF-137", "Origin": "User-supplied reference", "Canonical_ID": "CTX-REF-137", "Disposition": "Context / historical / date review", "Related_original_IDs": ""}], "reference_original": [{"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "AI drives over half of $5.3B technology value", "date": "2026-03", "theme": "Strategy & org", "summary": "Aramco reported $5.3B of Technology Realized Value (TRV) in 2025 from AI, digital and other technology, with CEO Amin Nasser saying over half was enabled by AI. Cumulative TRV since 2023 is $11.3B, and the company guided to an additional $3-5B in 2026.", "metrics": "$5.3B TRV 2025 (>50% AI-enabled); $11.3B cumulative since 2023; $3-5B additional TRV expected 2026", "vendors": "", "takeaway": "Publishing a single audited-style 'realized value' KPI for AI/tech turns AI from a cost center into a board-level P&L story; builders should instrument value capture from day one.", "source": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "source_name": "Aramco (CEO FY2025 media call remarks)", "id": 0}, {"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "Aramco launches Metabrain 250B-parameter industrial LLM", "date": "2024-03", "theme": "Enterprise GenAI rollout", "summary": "Aramco unveiled METABRAIN, a 250-billion-parameter generative AI model trained on about 7 trillion tokens of public and internal data spanning ~90 years of company history. It targets drilling-plan analysis, well recommendations and refined-product price/market forecasting, with a 1-trillion-parameter version planned.", "metrics": "250B parameters; 7T training tokens; ~90 years of data; 80+ AI professionals", "vendors": "Aramco Digital", "takeaway": "Proprietary operational history is the moat; NOCs with deep archives are building domain LLMs rather than relying only on general-purpose models.", "source": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm", "source_name": "Middle East AI News", "id": 1}, {"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "Aramco takes significant minority stake in HUMAIN", "date": "2025-10", "theme": "Partnership & investment", "summary": "PIF and Aramco signed a non-binding term sheet for Aramco to acquire a significant minority stake in HUMAIN, combining PIF and Aramco AI programs and assets under the PIF company. Nasser said it would strengthen Aramco's leadership in industrial AI applications.", "metrics": "Non-binding term sheet; PIF retains majority", "vendors": "HUMAIN, PIF, ALLaM", "takeaway": "Majors are consolidating AI compute, models and talent into shared vehicles rather than building everything in-house; expect sovereign AI stacks to become preferred suppliers in the Gulf.", "source": "https://www.aramco.com/en/news-media/news/2025/pif-and-aramco-agree-for-aramco-to-acquire-a-significant-minority-stake-in-humain", "source_name": "Aramco", "id": 2}, {"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "Aramco-Microsoft MoU on industrial AI and digital talent", "date": "2026-02", "theme": "Partnership & investment", "summary": "Aramco and Microsoft signed a non-binding MoU to accelerate industrial AI adoption on Azure with sovereign controls for Saudi data residency. Focus areas include operational efficiency, an industry alliance framework, industrial AI IP co-innovation, and workforce programs in AI engineering, cybersecurity, data governance and product management.", "metrics": "", "vendors": "Microsoft, Azure", "takeaway": "Sovereignty/data residency is a gating requirement for Gulf NOCs; enablement offerings must pair tooling with local-cloud deployment and talent programs.", "source": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "source_name": "Aramco", "id": 3}, {"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "date": "2025-06", "theme": "Workforce enablement", "summary": "Aramco says it processes ~10 billion data points daily and is using AI to reprocess old seismic data, optimize autonomous drilling and detect pipeline corrosion. It is training 6,000+ AI developers with Imperial College, Caltech and KAUST, and partnered with Groq on a regional AI inferencing center.", "metrics": "10B data points/day; 1,500 PB storage; 6,000+ AI developers being trained; 5 WEF Lighthouse awards", "vendors": "Groq, Qualcomm, Imperial College, Caltech, KAUST", "takeaway": "Pair large-scale technical upskilling with high-value 'old data, new answers' use cases like seismic reprocessing to show quick ROI.", "source": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai", "source_name": "Aramco Europe", "id": 4}, {"company": "ADNOC", "segment": "NOC", "hq": "UAE", "title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "date": "2025-11", "theme": "Workforce enablement", "summary": "ADNOC deployed generative AI company-wide in November 2023; by ADIPEC 2025 more than 40,000 employees had completed AI training, with utilization above 90% and over 70,000 productivity hours gained per month. At the same time ADNOC signed an agreement with Microsoft to co-develop and deploy AI agents across its operations.", "metrics": "40,000+ trained; >90% utilization; 70,000+ hours/month saved", "vendors": "Microsoft, AIQ, Masdar, XRG, Gecko Robotics", "takeaway": "Enterprise-wide training plus usage and hours-saved telemetry is the adoption playbook to copy; report utilization, not just licenses.", "source": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/", "source_name": "The National", "id": 5}, {"company": "ADNOC", "segment": "NOC", "hq": "UAE", "title": "AIQ wins $340M contract to deploy ENERGYai agents", "date": "2025-03", "theme": "Agentic AI", "summary": "AIQ announced a $340M, three-year contract to deploy ENERGYai, an agentic AI solution, across ADNOC's upstream value chain covering 28+ producing fields. The first operational version with five AI agents was slated for mid-2025, built on Azure, OSDU and OpenAI models.", "metrics": "$340M over 3 years; 28+ fields; 5 initial agents", "vendors": "AIQ, G42, Microsoft, OpenAI, OSDU", "takeaway": "Buying agentic AI as a multi-year outcome contract with a captive AI vendor is an alternative to piecemeal pilots.", "source": "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "source_name": "PR Newswire / AIQ", "id": 6}, {"company": "ADNOC", "segment": "NOC", "hq": "UAE", "title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "date": "2025-01", "theme": "Subsurface & exploration", "summary": "ADNOC and AIQ completed a 90-day trial of ENERGYai, a 70B-parameter LLM tuned on 50+ years of ADNOC data and drawing on over 15% of its onshore and offshore wells. The trial showed a 70% improvement in accuracy on major aspects of seismic interpretation plus better anomaly detection in reservoir monitoring.", "metrics": "70B-parameter LLM; 90-day PoC; >15% of wells; 70% seismic interpretation accuracy gain", "vendors": "AIQ", "takeaway": "Time-boxed PoCs with a hard accuracy metric on a core technical workflow make the case for scaling agentic tools.", "source": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "source_name": "ADNOC", "id": 7}, {"company": "ADNOC", "segment": "NOC", "hq": "UAE", "title": "ADNOC moves AI from pilots to enterprise-wide operations", "date": "2026-08", "theme": "Strategy & org", "summary": "ADNOC is embedding AI across exploration, drilling, production, maintenance and decision-making, aiming to become 'the world's most AI-enabled energy company.' Tools include ENERGYai agents, the Neuron 5 asset-performance platform for predicting equipment failure, and an internal AI Lab that cuts proof-of-concept-to-deployment time 3x.", "metrics": "3x faster PoC-to-deployment; months-to-days on technical tasks", "vendors": "AIQ, Microsoft, G42", "takeaway": "A dedicated internal AI lab with a mandate to shorten PoC-to-production is what turns pilots into operating change.", "source": "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "source_name": "Computer Weekly", "id": 8}, {"company": "ADNOC", "segment": "NOC", "hq": "UAE", "title": "AIQ and SLB co-develop ENERGYai for subsurface", "date": "2025-05", "theme": "Partnership & investment", "summary": "AIQ and SLB agreed to advance development and deployment of the ENERGYai agentic solution across ADNOC's subsurface operations, spanning geology, seismic and reservoir workflows. SLB brings its digital and agentic AI portfolio, including its Tela assistant.", "metrics": "", "vendors": "AIQ, SLB (Tela)", "takeaway": "OFS majors are becoming the channel for agentic subsurface AI; enablement builders should plan to integrate with SLB and Halliburton platforms.", "source": "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr", "source_name": "SLB", "id": 9}, {"company": "ExxonMobil", "segment": "Supermajor", "hq": "USA", "title": "First fully autonomous AI-drilled well offshore Guyana", "date": "2026-03", "theme": "Drilling & completions", "summary": "ExxonMobil completed what it called the industry's first fully closed-loop automated well placement with full rig automation in the Stabroek Block. VP of Exploration John Ardill said no driller touched the controls and AI handled drilling and geosteering through the reservoir section.", "metrics": "", "vendors": "Halliburton (LOGIX, EarthStar), Sekal (DrillTronics), Noble", "takeaway": "Closed-loop autonomy is arriving first in drilling, where outcomes are measurable and operations are repeatable; human oversight shifts to remote centers.", "source": "https://www.offshore-mag.com/drilling-completion/news/55380289/exxonmobil-achieves-milestone-automated-well-placement-offshore-guyana", "source_name": "Offshore Magazine", "id": 10}, {"company": "ExxonMobil", "segment": "Supermajor", "hq": "USA", "title": "AI identifies four new Guyana exploration opportunities", "date": "2026-08", "theme": "Subsurface & exploration", "summary": "ExxonMobil said it used AI, including deep learning, machine learning, HPC and advanced seismic imaging, applied to historical discoveries and drilling results to identify four new exploration opportunities in Guyana's Stabroek Block. It is expanding use of these tools to evaluate prospects that were previously hard to assess.", "metrics": "4 new opportunities identified", "vendors": "", "takeaway": "Reinterpreting legacy subsurface data with AI can surface new prospects in mature acreage; data readiness is the prerequisite.", "source": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "source_name": "World Oil", "id": 11}, {"company": "ExxonMobil", "segment": "Supermajor", "hq": "USA", "title": "AI plus process mining behind $15B cost-cut push", "date": "2024-11", "theme": "Back office & corporate", "summary": "ExxonMobil tied its $15B operating-cost reduction target for 2027 to AI combined with process intelligence run through its Global Business Solutions group. It uses Celonis process mining across procurement, supply chain, logistics, maintenance and AP/AR, and the CFO acknowledged that historically siloed, inconsistent processes had held back automation.", "metrics": "$15B cost-savings target by 2027", "vendors": "Celonis, AWS, Snowflake, SAP, Salesforce", "takeaway": "Standardize and mine processes before layering on AI agents; back-office AI ROI depends on clean end-to-end process data.", "source": "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts", "source_name": "Constellation Research", "id": 12}, {"company": "Chevron", "segment": "Supermajor", "hq": "USA", "title": "ApEX multi-agent GenAI searches 1M+ exploration files", "date": "2025-12", "theme": "Subsurface & exploration", "summary": "Chevron's in-house generative AI tool ApEX, launched in August 2024 and first used in Gulf of America deepwater exploration, lets explorationists query through a chat interface. Multiple specialized search agents comb more than a million files, cutting pre-drill studies from weeks of manual review to minutes.", "metrics": "1M+ files; weeks to minutes", "vendors": "In-house", "takeaway": "A multi-agent RAG over technical archives is the most repeatable subsurface GenAI win; specialize agents by data type (geospatial, assessments, rock properties).", "source": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c", "source_name": "Chevron via MarketScreener", "id": 13}, {"company": "Chevron", "segment": "Supermajor", "hq": "USA", "title": "Up to 20% workforce cut, technology and hubs cited", "date": "2025-02", "theme": "Strategy & org", "summary": "Chevron said it would cut 15-20% of its global workforce (up to ~9,000 roles) by end-2026 to save $2-3B. Levers included leveraging technology to raise productivity and changing where work is done via global capability centers such as Manila, Buenos Aires and its Bengaluru ENGINE center.", "metrics": "15-20% of workforce; ~9,000 roles; $2-3B savings by 2026", "vendors": "", "takeaway": "Majors are pairing technology-enabled productivity with centralized service hubs; AI enablement pitches should target work consolidated into these centers.", "source": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce", "source_name": "Inspectioneering", "id": 14}, {"company": "Chevron", "segment": "Supermajor", "hq": "USA", "title": "DataRobot agentic AI for autonomous robotic inspections", "date": "2026-06", "theme": "Agentic AI", "summary": "DataRobot is working with Chevron U.S.A. to apply agentic AI at the edge to Chevron's aerial and ground robot inspections under its 'Facilities and Operations of the Future' program. Agents generate mission plans and run a continuous 'Safe Start' condition assessment, replacing manual approval of each robotic mission.", "metrics": "", "vendors": "DataRobot, NVIDIA NIM", "takeaway": "Agentic AI adds most value where it removes human approval bottlenecks on already automated physical workflows, with governance built in.", "source": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "source_name": "DataRobot", "id": 15}, {"company": "Shell", "segment": "Supermajor", "hq": "UK", "title": "Shell extends C3 AI with agent-based root-cause analysis", "date": "2026-06", "theme": "Maintenance & reliability", "summary": "Shell signed a new multi-year agreement extending its C3 AI Reliability program, which has monitored equipment across its assets since 2018. The expansion adds AI agent-based root-cause analysis and remediation on top of anomaly detection, running on Microsoft Azure.", "metrics": "13,000+ pieces of equipment monitored; program since 2018", "vendors": "C3 AI, Microsoft Azure", "takeaway": "Mature predictive-maintenance estates are the natural base for agents: move from 'alert' to 'diagnose and recommend fix.'", "source": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations", "source_name": "Business Wire / C3 AI", "id": 16}, {"company": "Shell", "segment": "Supermajor", "hq": "UK", "title": "SLB and Shell co-develop agentic AI for upstream", "date": "2025-12", "theme": "Agentic AI", "summary": "SLB and Shell signed a strategic collaboration to develop agentic AI solutions for technical experts across subsurface, well construction and production. The work builds on SLB's Lumi data and AI platform and Shell's earlier deployment of Petrel across its global assets.", "metrics": "", "vendors": "SLB (Lumi, Petrel)", "takeaway": "Supermajors are co-developing agents with OFS vendors on shared data platforms; independent builders should interoperate with Lumi/OSDU instead of competing with them.", "source": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts", "source_name": "JPT (SPE)", "id": 17}, {"company": "Shell", "segment": "Supermajor", "hq": "UK", "title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "date": "2024-09", "theme": "Safety & emissions", "summary": "Shell and NVIDIA used Fourier Neural Operators in NVIDIA PhysicsNeMo to build surrogate models of CO2 storage that run about 100,000x faster than numerical simulators while keeping R-squared above 0.97. This enables screening tens of thousands of storage configurations for CCS site selection.", "metrics": "~100,000x speedup; R-squared >0.97; 1,000-year simulation horizon", "vendors": "NVIDIA PhysicsNeMo", "takeaway": "Physics-ML surrogates turn slow simulation into fast scenario screening, a strong pattern for subsurface and CCS workflows.", "source": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/", "source_name": "NVIDIA Technical Blog", "id": 18}, {"company": "BP", "segment": "Supermajor", "hq": "UK", "title": "bp renews Palantir for 5 years with AIP LLMs", "date": "2024-09", "theme": "Partnership & investment", "summary": "bp renewed its decade-long Palantir relationship with a five-year strategic agreement centered on Palantir's Artificial Intelligence Platform (AIP). AIP brings LLMs into bp's operations on top of digital twins fed by more than 2 million sensors to speed human decision-making.", "metrics": "5-year deal; 2M+ sensors", "vendors": "Palantir (Foundry, AIP)", "takeaway": "Operators with an existing ontology and digital-twin layer can add LLM copilots quickly; the data model is the durable asset.", "source": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/", "source_name": "Palantir IR", "id": 19}, {"company": "BP", "segment": "Supermajor", "hq": "UK", "title": "bp expands job cuts as AI drives efficiency", "date": "2025-08", "theme": "Strategy & org", "summary": "bp raised expected job losses to about 6,200 (roughly 15% of office-based staff) plus contractor cuts. CEO Murray Auchincloss said technology is helping improve capital productivity and drive cost reductions, with AI part of the overhaul toward a $5B cost-cut target by end-2027.", "metrics": "~6,200 roles; $900M savings H1 2025; $5B cost-cut target by 2027", "vendors": "", "takeaway": "AI is now explicitly linked to headcount plans at majors; change management and redeployment belong in the business case alongside productivity.", "source": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html", "source_name": "Yahoo Finance UK", "id": 20}, {"company": "TotalEnergies", "segment": "Supermajor", "hq": "France", "title": "TotalEnergies and Mistral launch €100M frontier subsurface AI program", "date": "2026-09", "theme": "Subsurface & exploration", "summary": "TotalEnergies and Mistral AI announced a three-year joint program worth more than €100M to build frontier, agentic AI models for geoscientists in exploration, reservoir characterization and development. A joint lab pairs TotalEnergies subsurface experts with Mistral ML researchers on about 10 petaflops of data.", "metrics": ">€100M over 3 years", "vendors": "Mistral AI", "takeaway": "European majors are backing sovereign model vendors to co-train domain models on proprietary subsurface data.", "source": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "source_name": "TotalEnergies", "id": 21}, {"company": "TotalEnergies", "segment": "Supermajor", "hq": "France", "title": "Mistral joint lab: assistant for 1,000 researchers", "date": "2025-06", "theme": "Partnership & investment", "summary": "TotalEnergies and Mistral AI set up a joint innovation lab. First use cases include an AI assistant for TotalEnergies' 1,000 researchers, decision support for industrial asset performance and CO2 reduction, customer energy-saving tools, and specialized assistants for support functions.", "metrics": "1,000 researchers; 300-person Digital Factory; 100+ digital solutions, 60 using ML/GenAI", "vendors": "Mistral AI", "takeaway": "Start with a bounded expert population (R&D) for a domain assistant, then extend to support functions.", "source": "https://totalenergies.com/news/press-releases/totalenergies-collaborate-mistral-ai-increase-application-artificial", "source_name": "TotalEnergies", "id": 22}, {"company": "TotalEnergies", "segment": "Supermajor", "hq": "France", "title": "Early adopter of Microsoft Copilot across the workforce", "date": "2024-02", "theme": "Enterprise GenAI rollout", "summary": "TotalEnergies said it was one of the first companies to adopt Microsoft's generative AI assistant to boost employee productivity. Its 300-person Digital Factory runs AI for production forecasting, predictive maintenance and renewables siting, with experts keeping the final say.", "metrics": "300 AI/digital experts in Digital Factory", "vendors": "Microsoft Copilot", "takeaway": "Horizontal copilots plus a central digital factory for vertical use cases is the standard two-track model among majors.", "source": "https://totalenergies.com/news/news/ia-expediting-energy-transition", "source_name": "TotalEnergies", "id": 23}, {"company": "TotalEnergies", "segment": "Supermajor", "hq": "France", "title": "Cognite industrial AI across all operated upstream assets", "date": "2025-09", "theme": "Production & operations", "summary": "TotalEnergies expanded its partnership with Cognite in a three-year plan to deploy industrial AI and an AI-ready data foundation across all its operated upstream assets worldwide, from drilling to production. OneTech president Namita Shah framed the unified data layer as the precondition for scaling AI solutions.", "metrics": "3-year program; all operated upstream assets", "vendors": "Cognite (Data Fusion)", "takeaway": "A contextualized industrial data layer comes before agents; budget for it explicitly.", "source": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale", "source_name": "TotalEnergies", "id": 24}, {"company": "Equinor", "segment": "Integrated", "hq": "Norway", "title": "AI saved Equinor $130M in 2025", "date": "2026-01", "theme": "Production & operations", "summary": "Equinor said AI created $130M in value and savings in 2025 and more than $330M since 2020. Predictive maintenance on 700+ rotating machines with 24,000 sensors accounts for $120M since 2020; AI interpreted 2 million km2 of seismic in 2025 at 10x capacity; and a well-planning tool saved $12M on Johan Sverdrup Phase 3.", "metrics": "$130M in 2025; $330M+ since 2020; 700+ machines, 24,000 sensors; 10x seismic interpretation; $12M Sverdrup saving", "vendors": "", "takeaway": "Publish value by use case; a few industrial ML workhorses (rotating equipment, seismic) still deliver most of the dollars while GenAI tools spread to employees.", "source": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "source_name": "Equinor", "id": 25}, {"company": "Eni", "segment": "Integrated", "hq": "Italy", "title": "HPC6 supercomputer and company-wide responsible-AI training", "date": "2025-06", "theme": "Workforce enablement", "summary": "Eni runs HPC6, launched November 2024 and ranked in the global TOP10, for AI-driven geoscience, safety (the 'Safety Pre-sense' tool) and other uses, and has opened HPC6 access to startups through Call4Innovators. From 2025 it rolled out a company-wide training program on responsible AI, regulation, ethics and data literacy, with advanced modules.", "metrics": "HPC6 in global TOP10; 10 Call4Innovators projects selected (Dec 2025)", "vendors": "HPE, AMD, CINECA, Plug and Play", "takeaway": "Owning compute is a differentiator for subsurface AI; pairing it with mandatory AI literacy helps meet EU AI Act obligations.", "source": "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html", "source_name": "Eni", "id": 26}, {"company": "Repsol", "segment": "Integrated", "hq": "Spain", "title": "Copilot to 25,000 staff; pilot saved 121 min/week", "date": "2026-02", "theme": "Enterprise GenAI rollout", "summary": "Repsol extended Copilot to about 25,000 employees after a four-month M365 Copilot pilot with 550 users showed 121 minutes saved per person per week and 16% higher work quality. It runs 60+ GenAI use cases and 34 agents, and has trained 5,000+ employees plus all senior management in prompting.", "metrics": "~25,000 employees; 121 min/week saved; +16% quality; 60+ use cases; 34 agents; 5,000+ trained; 200k+ lines via GitHub Copilot", "vendors": "Microsoft 365 Copilot, GitHub Copilot", "takeaway": "Run a measured pilot (time saved, quality, would-not-give-up rate) before mass licensing; it gives the board-ready business case.", "source": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml", "source_name": "Repsol", "id": 27}, {"company": "Repsol", "segment": "Integrated", "hq": "Spain", "title": "Multi-agent platform: 22 agents supervised by 50 staff", "date": "2025-10", "theme": "Agentic AI", "summary": "Repsol launched a multi-agent platform built with Accenture in which AI agents self-organize and coordinate tasks within workflows. The pilot ran three teams with 22 agents in the 700-person Digitalization and Services department, supervised by 50+ employees for over four months.", "metrics": "22 agents; 3 agent teams; 50+ supervisors; 4+ months", "vendors": "Accenture", "takeaway": "Pilot agents inside your own IT/digital function first, with explicit human supervisors, before pushing to operations.", "source": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "source_name": "Repsol", "id": 28}, {"company": "Repsol", "segment": "Integrated", "hq": "Spain", "title": "Accenture and NVIDIA to scale AI agents", "date": "2025-02", "theme": "Partnership & investment", "summary": "Repsol expanded its Digital Program by deploying AI agent systems with Accenture's AI Refinery and the NVIDIA AI platform, targeting process optimization, planning, forecasting, maintenance and customer personalization. It also plans to explore NVIDIA Omniverse digital twins.", "metrics": "650+ digital initiatives; 1,200+ people on digital projects; 50 tech partners", "vendors": "Accenture AI Refinery, NVIDIA AI Enterprise, NVIDIA Omniverse", "takeaway": "SI-led agent factories are a common way integrateds get to scale fast; independent tools need to plug into SI delivery models.", "source": "https://newsroom.accenture.com/news/2025/repsol-boosts-its-digital-strategy-by-collaborating-with-accenture-to-adopt-ai-agents", "source_name": "Accenture Newsroom", "id": 29}, {"company": "Petrobras", "segment": "NOC", "hq": "Brazil", "title": "ChatPetrobras GenAI rolled out to 110,000 workers", "date": "2023-12", "theme": "Enterprise GenAI rollout", "summary": "Petrobras built ChatPetrobras with Microsoft on Azure OpenAI Service (GPT models) and made it available through an internal portal to 110,000 employees and contractors. It supports reports, summaries, presentations and Q&A to cut time and cost in administrative and operational processes.", "metrics": "110,000 workers (employees + contractors)", "vendors": "Microsoft, Azure OpenAI", "takeaway": "A secure internal ChatGPT-style wrapper extended to contractors reaches the whole workforce quickly and cheaply.", "source": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br", "source_name": "Microsoft News Center Latam", "id": 30}, {"company": "Petronas", "segment": "NOC", "hq": "Malaysia", "title": "Petronas-Microsoft MoU on agentic AI and Copilot", "date": "2025-07", "theme": "Partnership & investment", "summary": "Petronas and Microsoft signed an MoU at Energy Asia 2025 to deepen AI and digital transformation, covering agentic AI, Microsoft Copilot, data analytics, cloud and cybersecurity across Petronas' value chain. It uses Microsoft's new Malaysia West cloud region.", "metrics": "", "vendors": "Microsoft Copilot, Azure (Malaysia West)", "takeaway": "In-country cloud regions are unlocking Copilot/agent rollouts at Asian NOCs.", "source": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "source_name": "Petronas", "id": 31}, {"company": "Petronas", "segment": "NOC", "hq": "Malaysia", "title": "Agentic AI added to myPROdata upstream platform", "date": "2026-09", "theme": "Agentic AI", "summary": "Petronas, through Malaysia Petroleum Management, partnered with Iraya Energies to add agentic AI to its myPROdata upstream data platform so it can autonomously analyze and connect G&G and engineering data for investors. The aim is to speed E&P evaluation and support RM50-60B a year of upstream investment.", "metrics": "RM50-60B/yr investment target; discovery-to-first-oil goal cut from 100 to 50 months", "vendors": "Iraya Energies", "takeaway": "Agentic AI is being pointed outward at investors and partners; data-room agents are a monetizable NOC use case.", "source": "https://www.oedigital.com/news/542771-petronas-turns-to-ai-to-accelerate-upstream-investment", "source_name": "Offshore Engineer", "id": 32}, {"company": "PetroChina / CNPC", "segment": "NOC", "hq": "China", "title": "CNPC releases 300B-parameter Kunlun energy model", "date": "2025-05", "theme": "Enterprise GenAI rollout", "summary": "CNPC released an upgraded 300B-parameter Kunlun large model, which it described as the first large model approved by regulators in China's energy and chemical sector (approved August 2024). The industry LLM scaled from 70B to 300B parameters, alongside larger vision (4.4B) and multimodal (80B) models.", "metrics": "LLM 70B to 300B; vision 300M to 4.4B; multimodal 16B to 80B", "vendors": "China Mobile, Huawei, iFlytek", "takeaway": "Chinese NOCs are building sovereign domain foundation models; Western vendors are largely shut out of that stack.", "source": "https://www.cnpc.com.cn/en/nr2025/202506/e9b44b36040f452796711916fbf49135.shtml", "source_name": "CNPC", "id": 33}, {"company": "CNOOC", "segment": "NOC", "hq": "China", "title": "Offshore predictive maintenance cuts failures ~20%", "date": "2026-04", "theme": "Maintenance & reliability", "summary": "CNOOC deployed AI predictive maintenance across offshore platforms including Bohai Oilfield, using real-time sensor data from subsea pumps and turbines. China Daily reported roughly a 20% cut in critical equipment failure rates; CNPC's Changqing field reported ~30% shorter exploration cycles and 10%+ higher per-well output.", "metrics": "~20% fewer critical equipment failures (CNOOC); ~30% shorter exploration cycle, >10% per-well output (CNPC Changqing)", "vendors": "", "takeaway": "Rotating-equipment predictive maintenance remains the most consistently quantified AI win across all regions.", "source": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "source_name": "China Daily", "id": 34}, {"company": "Woodside Energy", "segment": "Integrated", "hq": "Australia", "title": "Lumina enterprise AI platform runs 50+ agents", "date": "2025-12", "theme": "Agentic AI", "summary": "Woodside's enterprise AI platform Lumina supports more than 50 AI agents handling day-to-day tasks across business functions, overseen by an AI Centre of Excellence and a senior-leader AI Council. Microsoft Copilot covers personal productivity and ChatGPT Enterprise is used as a sandbox for testing custom agents before they move into Lumina (date approximate; undated company page).", "metrics": "50+ AI agents in production", "vendors": "Microsoft Copilot, ChatGPT Enterprise", "takeaway": "A layered stack (Copilot for everyone, ChatGPT Enterprise as sandbox, an in-house agent platform for production) plus a governance council is a template worth copying.", "source": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "source_name": "Woodside Energy", "id": 35}, {"company": "Santos", "segment": "Integrated", "hq": "Australia", "title": "Xecta AI digital twin expanded to East Australia and PNG", "date": "2025-07", "theme": "Production & operations", "summary": "After a successful rollout in the Cooper Basin, Santos signed a five-year agreement with Xecta to build an AI-driven digital twin for production optimization across its Eastern Australia and Papua New Guinea operations. This builds on its 2023 Kelvin.ai agreement for autonomous control of wellsite and facility operations.", "metrics": "5-year agreement", "vendors": "Xecta, Kelvin.ai, SIG ML", "takeaway": "Mid-size integrateds scale AI by proving value in one basin and then signing enterprise-wide multi-year deals.", "source": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta", "source_name": "Energy News Bulletin", "id": 36}, {"company": "OMV", "segment": "Integrated", "hq": "Austria", "title": "OMV AI Hub: 2,500 monthly GPT users, 10,000 trained", "date": "2026-03", "theme": "Enterprise GenAI rollout", "summary": "OMV launched the AI Hub on Azure in 2025 as a single entry point for GenAI chatbots, assistants and agents; 2,500+ employees use its in-house GPT assistant monthly and 10,000+ joined GenAI workshops. AI projects in use doubled to 50+, with tools such as a Norwegian AI companion over 2.5 million pages of geological data and 90-day-ahead well-failure prediction.", "metrics": "2,500+ monthly GPT users; 10,000+ trained; 50+ AI projects live; 40+ in development; 210+ ideas", "vendors": "Microsoft Azure, SAP, Salesforce, SLB, TerraAI", "takeaway": "A single AI front door with a visible pipeline (ideas to development to live) makes adoption measurable.", "source": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "source_name": "OMV Annual Report 2025", "id": 37}, {"company": "Kuwait Oil Company", "segment": "NOC", "hq": "Kuwait", "title": "AI Innovation Center delivers G Agent drilling win", "date": "2025-09", "theme": "Drilling & completions", "summary": "KOC opened its first AI Innovation Center in August 2025 with Microsoft, Halliburton and Ghaia.ai. Its first result is an AI-powered rig drilling project on Ghaia.ai's G Agent platform, a human-agent collaboration mesh that improves productivity and planning of drilling rig operations.", "metrics": "", "vendors": "Microsoft, Halliburton, Ghaia.ai (G Agent)", "takeaway": "Co-located innovation centers with a hyperscaler, an OFS firm and an agent startup can get a first production use case within weeks.", "source": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/", "source_name": "Kuwait Times", "id": 38}, {"company": "Devon Energy", "segment": "Independent E&P", "hq": "US", "title": "ChatDVN internal AI platform plus AI-101 training", "date": "2025-07", "theme": "Workforce enablement", "summary": "Devon built ChatDVN, a proprietary platform combining AI tools, machine learning and LLMs, and paired it with a company-wide AI-101 course. CTO Trey Lowe cited productivity gains among engineers, software developers and geoscientists, and use cases such as gas-lift optimization and simul-frac planning done in hours. Devon is standardizing wellsite sensors and cameras so AI can scale, including planned video-to-data analytics in 2026.", "metrics": "About 1,000 employees completed AI-101 by end of June 2025; 15-30% productivity gain for engineers, developers and geoscientists; 5% oil uplift at one Delaware gas-lift site", "vendors": "ChatDVN (in-house)", "takeaway": "A branded internal assistant plus a mandatory foundational course gets adoption started faster than rolling out tools on their own.", "source": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "source_name": "American Oil & Gas Reporter", "id": 39}, {"company": "Devon Energy", "segment": "Independent E&P", "hq": "US", "title": "AI-enabled lift and maintenance drive $1B optimization plan", "date": "2026-02", "theme": "Production & operations", "summary": "On the Q4 2025 call, Devon said it is taking AI-enabled artificial lift optimization and advanced analytics beyond pilots, and adding condition-based maintenance and faster drilling cycle times. Management said investments in AI and platforms built over the prior year were 'coming to fruition' in production. The work sits inside a $1B pre-tax business optimization program.", "metrics": "85% of $1B optimization target captured by YE2025; more than 100 active workstreams; downtime cut from about 7% to under 5%; base production about 5,000 BOE/d ahead of plan", "vendors": "", "takeaway": "Track AI as named workstreams inside a cost program with a dollar target, so the impact shows up in guidance and not only in anecdotes.", "source": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q4 2025 call transcript)", "id": 40}, {"company": "Devon Energy", "segment": "Independent E&P", "hq": "US", "title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "date": "2026-05", "theme": "Production & operations", "summary": "Devon's Smart Gas Lift program pairs AI models with a physics-based calculation to set gas-lift injection rates in a closed loop directly at the wells. A 2025 pilot delivered a 2-3% uplift. Devon then moved to full implementation in the Delaware and is piloting similar AI models on ESPs and rod pumps.", "metrics": "Pilot uplift 2-3%; more than 850 wells live, heading to about 1,500; full-scale uplift said to exceed the pilot", "vendors": "", "takeaway": "Closed-loop setpoint control on artificial lift, a physics model plus ML, is a proven first 'agentic' use case in production ops that scales well by well.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 41}, {"company": "Devon Energy", "segment": "Independent E&P", "hq": "US", "title": "AI speeds Coterra integration; in-house plus vendor model", "date": "2026-08", "theme": "Agentic AI", "summary": "CTO Trey Lowe said Devon 'leaned in really hard' on enterprise AI after ChatGPT launched and is merging its operations-focused AI stack with Coterra's ML well-prediction models. Closed-loop gas lift went from a few wells to about 1,000, and was extended to legacy Coterra wells. Service-company partners supply pieces of the solution, but Lowe called 'empowering our employees' the 'secret sauce'. AI tools also benchmark every rig and crew against a best-performance curve to find D&C gaps.", "metrics": "Closed-loop gas-lift wells rose from 'a few' to a couple of hundred (March 2026) to about 1,000; more than 2,000 more Permian candidates; new data tables built in 2 days instead of a couple of months", "vendors": "ChatDVN, Smart Gas Lift (in-house); unnamed service-provider partners", "takeaway": "In M&A, a mature AI and data stack becomes a synergy engine: onboarding the acquired staff onto the tools is itself a value lever.", "source": "https://www.fool.com/earnings/call-transcripts/2026/08/11/devon-energy-dvn-q2-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Q2 2026 call transcript)", "id": 42}, {"company": "Devon Energy", "segment": "Independent E&P", "hq": "US", "title": "AI agents for well design and completions sequencing", "date": "2026-05", "theme": "Drilling & completions", "summary": "At IADC Drilling Onshore, Devon's VP of D&C described ChatDVN plus the proprietary InnDrill and InnFrac data applications, which were built over about 2.5 years. Uses include well-design optimization agents, AI-driven completions sequencing, basin-wide pre-drill performance prediction, and supply-chain and contract optimization. These tools support the post-Coterra synergy plan.", "metrics": "$350M of the $1B merger synergy target comes from D&C capital optimization", "vendors": "ChatDVN, InnDrill, InnFrac (in-house)", "takeaway": "Owning the D&C data layer (InnDrill/InnFrac) is what makes design agents and pre-drill prediction possible; build the data apps before the agents.", "source": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873", "source_name": "Drilling Contractor (IADC)", "id": 43}, {"company": "Coterra Energy", "segment": "Independent E&P", "hq": "US", "title": "ML well-prediction models fed capital planning", "date": "2026-08", "theme": "Subsurface & exploration", "summary": "According to Devon's CTO after the merger closed, Coterra had invested heavily in AI and machine-learning models for well prediction that informed its capital plans. Those models are now being combined with Devon's operations-focused AI systems.", "metrics": "", "vendors": "", "takeaway": "Well-performance prediction models used in capital allocation are a high-value, low-integration-risk AI use case that survives M&A.", "source": "https://www.fool.com/earnings/call-transcripts/2026/08/11/devon-energy-dvn-q2-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Devon Q2 2026 call transcript)", "id": 44}, {"company": "EOG Resources", "segment": "Independent E&P", "hq": "US", "title": "Generative AI across geology, drilling and safety", "date": "2025-11", "theme": "Enterprise GenAI rollout", "summary": "EOG said it uses 'quite a bit of generative AI' to organize geologic data and find hidden trends, alongside real-time drilling optimization, predictive maintenance and some autonomous field operations. The work builds on almost 10 years of smart centralized gas-lift technology and newer deep-learning tools that capture human observation and experiential learning. EOG said it is 'not quite to true agentic intelligence' and continues to favor building apps in-house next to field operations.", "metrics": "About 10 years of ML-driven gas-lift optimization", "vendors": "In-house apps", "takeaway": "Tie in-house software to proprietary field sensors and downhole tools, and capture expert tacit knowledge as training data.", "source": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q3 2025 call transcript)", "id": 45}, {"company": "EOG Resources", "segment": "Independent E&P", "hq": "US", "title": "ML 'production optimizers' and 24-hour control room", "date": "2026-02", "theme": "Production & operations", "summary": "EOG described its proprietary production optimizers program, which uses machine learning to optimize base production for better run time and lower cost. EOG is also rolling automation across acquired Utica operations so wells can be managed remotely from its 24-hour control room.", "metrics": "", "vendors": "Production optimizers (in-house)", "takeaway": "Standardize automation and remote monitoring on acquired assets right away so the ML optimizers have data to work on.", "source": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "source_name": "The Motley Fool (Q4 2025 call transcript)", "id": 46}, {"company": "Diamondback Energy", "segment": "Independent E&P", "hq": "US", "title": "ML and AI in field ops cut downtime post-Endeavor", "date": "2026-05", "theme": "Production & operations", "summary": "Diamondback management said that, after the Endeavor merger, layering machine learning onto its data streams and putting AI into field operations was reducing downtime. They cited this as a big part of the Q1 2026 production beat, together with completion-design and workover optimization.", "metrics": "", "vendors": "", "takeaway": "Combining merged teams' playbooks with ML on unified data streams can produce a production beat within a few quarters.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 47}, {"company": "Diamondback Energy", "segment": "Independent E&P", "hq": "US", "title": "Moving to 'pump by exception' with AI and automation", "date": "2026-08", "theme": "Production & operations", "summary": "CFO Jere Thompson said AI 'is helping Diamondback in the office today' and that AI and automation will be very big drivers of a shallower or cheaper-to-maintain production base as the company moves to a pump-by-exception model. Operations leadership said AI and automation are applied first to day-to-day artificial-lift optimization and downtime management. CEO Kaes Van't Hof called it the 'first inning' and said even a 1% improvement in decline rate matters.", "metrics": "", "vendors": "", "takeaway": "Framing AI value as lower base decline and sustaining capex makes the investor case concrete, even when adoption is early.", "source": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Q2 2026 call transcript)", "id": 48}, {"company": "ConocoPhillips", "segment": "Independent E&P", "hq": "US", "title": "CEO: embedding AI into core processes at scale", "date": "2026-05", "theme": "Strategy & org", "summary": "Answering a shareholder question at the 2026 annual meeting, CEO Ryan Lance called AI 'an increasingly important enabler'. He said it builds on long-standing analytics and that the company is focused on embedding AI into core business processes to cut manual effort, speed decision cycles and improve consistency across geographically diverse operations. He said all applications run under an enterprise framework for governance, security and responsible use.", "metrics": "", "vendors": "", "takeaway": "Enterprise framing (governance first, scale proven use cases, redeploy people to higher-value work) is the template large E&Ps are signaling to shareholders.", "source": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "source_name": "ConocoPhillips (2026 Annual Meeting transcript)", "id": 49}, {"company": "Occidental Petroleum", "segment": "Independent E&P", "hq": "US", "title": "AI center of excellence and Gulf of America AI", "date": "2025-02", "theme": "Strategy & org", "summary": "Oxy said it created an AI center of excellence to align all intercompany AI initiatives. Its Gulf of America operations use AI for supply chain management, asset integrity and reservoir characterization. Oxy also cited large new seismic datasets, including Algeria's largest-ever seismic acquisition, as fuel for AI applications.", "metrics": "", "vendors": "", "takeaway": "A central AI CoE that aligns business-unit initiatives, applied first to a complex asset, is a common starting operating model.", "source": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "source_name": "The Motley Fool (Q4 2024 call transcript)", "id": 50}, {"company": "Occidental Petroleum", "segment": "Independent E&P", "hq": "US", "title": "AI subsurface project targets complex Gulf reservoirs", "date": "2025-08", "theme": "Subsurface & exploration", "summary": "CEO Vicki Hollub said Oxy is building internal AI capabilities, with a project aimed at the highly complex Gulf of America subsurface, and hoped to begin acting on results within 1-2 years. She said AI is already helping with Permian and other onshore subsurface work. Oxy also credited teams 'using AI' in its operations center with record artificial-lift reliability and workover performance.", "metrics": "", "vendors": "", "takeaway": "Point AI at the hardest subsurface problems where incremental recovery is worth the most, and report ops-center reliability wins alongside.", "source": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q2 2025 call transcript)", "id": 51}, {"company": "Occidental Petroleum", "segment": "Independent E&P", "hq": "US", "title": "AI now part of workforce efficiency agenda", "date": "2026-05", "theme": "Workforce enablement", "summary": "On the Q1 2026 call, Oxy leadership said 'AI is taking on a larger role across all disciplines' and named workforce efficiency as a priority. Richard Jackson said Oxy has been doing 'quite a bit of work on workforce', including AI and re-examining processes and priorities. In Q2 2026, Oxy listed AI with CO2, power and water as core technology and partnership areas behind its $4B incremental cash-flow target.", "metrics": "$4B incremental sustainable cash flow target by 2030 (AI one of several levers)", "vendors": "", "takeaway": "Expect AI to be tied explicitly to workforce redesign and process re-engineering, not only to technical use cases.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 52}, {"company": "Ovintiv", "segment": "Independent E&P", "hq": "US", "title": "In-house AI across production, drilling and completions", "date": "2025-11", "theme": "Production & operations", "summary": "CEO Brendan McCracken said Ovintiv is 'just at the very front end' of applying AI, and is active in three areas: production operations (uptime, artificial-lift optimization), drilling times and costs, and completions. For the NuVista acquisition, Ovintiv planned to connect wells to its Grand Prairie operations control center and use in-house AI tools and algorithms to optimize lift setpoints, with remote shut-in and restart in minutes. It listed AI as both a near-term and a long-term synergy bucket.", "metrics": "", "vendors": "Operations control center; in-house AI algorithms", "takeaway": "Put AI and automation in the acquisition synergy case and connect acquired wells to the control center on day one.", "source": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q3 2025 call transcript)", "id": 53}, {"company": "Ovintiv", "segment": "Independent E&P", "hq": "US", "title": "DRIVE Center AI optimizes rigs from day one", "date": "2026-02", "theme": "Drilling & completions", "summary": "Weeks after closing an acquisition, Ovintiv had linked all producing wells to its operations control center and the drilling rigs to its DRIVE Center, an optimization tool that uses AI to improve drilling performance. Management said this lets synergies start immediately.", "metrics": "About $1M per well synergy savings built into guidance", "vendors": "DRIVE Center (in-house)", "takeaway": "A centralized AI drilling-optimization center makes acquired rigs productive in weeks, which strengthens bid economics.", "source": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q4 2025 call transcript)", "id": 54}, {"company": "Ovintiv", "segment": "Independent E&P", "hq": "US", "title": "AI tunes 70-odd frac design inputs in real time", "date": "2026-05", "theme": "Drilling & completions", "summary": "McCracken called AI paired with Ovintiv's private dataset 'the big technology frontier'. The company develops in-house algorithms for production operations centers (uptime, artificial lift) and for frac design, tuning 'the 70 odd input design factors' of each frac job in real time.", "metrics": "About 70 frac design input factors optimized in real time", "vendors": "In-house algorithms", "takeaway": "A proprietary operating dataset plus in-house models is presented as a moat that competitors cannot copy.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 55}, {"company": "EQT", "segment": "Independent E&P", "hq": "US", "title": "AI tools to target service rebids and procurement", "date": "2026-02", "theme": "Back office & corporate", "summary": "EQT said it will aggressively rebid oilfield services and that AI tools will help pinpoint those efforts. It expects low-single-digit cost reductions on the procurement side, and said it is focused on 'the procurement and the backside' as well as the field.", "metrics": "Low-single-digit % procurement savings targeted", "vendors": "", "takeaway": "Procurement analytics (spend categorization, rebid targeting) is a quick back-office AI win with measurable savings.", "source": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q4 2025 call transcript)", "id": 56}, {"company": "Matador Resources", "segment": "Independent E&P", "hq": "US", "title": "Vendor-led AI: 'fast follower' approach", "date": "2026-02", "theme": "Partnership & investment", "summary": "CEO Joe Foran said Matador's most effective path has been working with its vendors, who 'most often' already have the AI programs and are ahead. He described 'baby steps' taken deliberately with partners as a win-win.", "metrics": "", "vendors": "Unnamed oilfield vendors", "takeaway": "Mid-caps can reach AI value by partnering with service vendors instead of building platforms, but need their own governance.", "source": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q4 2025 call transcript)", "id": 57}, {"company": "Matador Resources", "segment": "Independent E&P", "hq": "US", "title": "AI committee; AI used for lateral targeting", "date": "2026-05", "theme": "Drilling & completions", "summary": "Matador formed a cross-functional committee, led by Glenn Stetson with Jordan Ellington, to adopt AI in a 'controlled, organized fashion'. Uses include AI-driven analytics that surface actionable alerts to cut downtime, real-time frac monitoring (pressures, water and sand volumes, logistics) as simul- and trimul-frac expand, and AI-assisted geosteering to stay in the best part of the target zone.", "metrics": "More than 36 drilling records set in Q1 2026 (MaxCom-supported)", "vendors": "MaxCom (in-house real-time center)", "takeaway": "A small named AI committee with a clear 'fast follower' mandate is a lightweight governance model for mid-cap operators.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 58}, {"company": "Chord Energy", "segment": "Independent E&P", "hq": "US", "title": "AI rod-pump optimization arrests base decline", "date": "2026-05", "theme": "Production & operations", "summary": "Chord adjusted its artificial-intelligence rod-pump controls to maximize productivity from older Bakken wells, alongside lowering pumps deeper and adding workover rigs. Management showed a positive impact on base production and a halt in decline for the targeted well group.", "metrics": "", "vendors": "", "takeaway": "Retuning AI lift controllers for legacy wells is a cheap, short-cycle barrel source when prices call for it.", "source": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "source_name": "The Motley Fool (Q1 2026 call transcript)", "id": 59}, {"company": "Chord Energy", "segment": "Independent E&P", "hq": "US", "title": "AI rod-pump program deployed field-wide", "date": "2026-08", "theme": "Production & operations", "summary": "Chord said AI optimization of its rod-pump program is now 'pretty broadly implemented' across its Williston field, where nearly every well ends up on rod pump. The system sets pump loading to reduce wear and improve production, and Chord is looking for other places to apply similar computer-driven optimization.", "metrics": "", "vendors": "", "takeaway": "Standardized lift type (rod pump) across a field makes one AI optimizer broadly deployable.", "source": "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/", "source_name": "The Motley Fool (Q2 2026 call transcript)", "id": 60}, {"company": "Canadian Natural Resources", "segment": "Independent E&P", "hq": "Canada", "title": "Formal AI policy: no AI in production-critical processes", "date": "2026-03", "theme": "Strategy & org", "summary": "CNRL's 2025 Annual Information Form says it uses AI 'responsibly' to widen employee access to knowledge, deliver data-driven insights and improve efficiency through 'incremental, cost-justified deployments'. It will not rely on AI for time-sensitive or production-critical processes until controls meet necessary standards. Employees must stay involved in and accountable for AI outputs, and AI risk is reported to board committees.", "metrics": "", "vendors": "", "takeaway": "A conservative, human-in-the-loop policy disclosed in securities filings shows how cost-focused operators govern GenAI.", "source": "https://www.sec.gov/Archives/edgar/data/1017413/000101741326000018/cnq-20251231.htm", "source_name": "SEC (CNRL Form 40-F / 2025 AIF)", "id": 61}, {"company": "Suncor Energy", "segment": "Independent E&P", "hq": "Canada", "title": "Autonomous haul fleet nears 140 trucks", "date": "2025-07", "theme": "Production & operations", "summary": "Suncor runs Canada's largest autonomous haul truck fleet at its Base Plant oil sands mines, with nearly 120 Komatsu 980E trucks in 2025 and a plan for 140 by year-end. CEO Rich Kruger said the technology 'makes us a better operator'. The program was announced in 2018 with an expected loss of about 400 heavy-equipment operator jobs over six years.", "metrics": "About 120 autonomous trucks (target 140 by YE2025); about 400 operator positions affected (2018 plan)", "vendors": "Komatsu (AHS)", "takeaway": "Autonomy at fleet scale changes the workforce, so a transition plan needs to be part of the business case.", "source": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/", "source_name": "International Mining", "id": 62}, {"company": "Phillips 66", "segment": "Refining", "hq": "US", "title": "Machine learning shortens refinery turnarounds", "date": "2026-02", "theme": "Maintenance & reliability", "summary": "CEO Mark Lashier said Phillips 66 is taking work out of turnarounds and using tools 'like machine learning' to plan and execute them. This shortens duration and financial impact, and the company cut its turnaround spend forecast substantially mid-2025 and still beat it.", "metrics": "Turnaround spend forecast cut substantially mid-2025, then beaten", "vendors": "", "takeaway": "ML-driven turnaround scoping (moving work outside the outage window) is a direct availability and cost lever for refiners.", "source": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "source_name": "Phillips 66 IR (LSEG Q4 2025 transcript)", "id": 63}, {"company": "Phillips 66", "segment": "Refining", "hq": "US", "title": "Company AI program 'normalizing' AI for front line", "date": "2026-08", "theme": "Workforce enablement", "summary": "Lashier said Phillips 66 has implemented an AI program that is 'normalizing the use of AI inside the company', with dedicated staff finding use cases and extending them across the organization. Front-line engineers use AI to extract data and deploy fixes in 'days and weeks' instead of months or years. Refining leadership said AI helps it find opportunities hidden in 'the fog of too much data' across about 200 cost-out projects.", "metrics": "About 200 cost-reduction projects in flight; refining cost cut by more than $1/bbl, heading to $1.50/bbl (AI one of several contributors)", "vendors": "", "takeaway": "A central use-case team that replicates front-line wins across sites, described as 'human-centered', is a scalable refinery AI operating model.", "source": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "source_name": "Phillips 66 IR (S&P Global Q2 2026 transcript)", "id": 64}, {"company": "Valero Energy", "segment": "Refining", "hq": "US", "title": "'Cautiously optimistic' on AI; data quality first", "date": "2025-10", "theme": "Maintenance & reliability", "summary": "Valero's refining SVP Greg Bram said Valero is evaluating several AI and ML applications and has deployed some to improve availability. He stressed that success depends on high-quality operating data, which Valero began standardizing 10-15 years ago. Valero also uses robotics for tank cleaning, drones for inspections, and control-room consolidation.", "metrics": "", "vendors": "", "takeaway": "A decade of consistent operating-data practices is the prerequisite for refinery AI; start with availability and inspection use cases.", "source": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q3 2025 call transcript)", "id": 65}, {"company": "Marathon Petroleum", "segment": "Refining", "hq": "US", "title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "date": "2024", "theme": "Production & operations", "summary": "Marathon Petroleum reported that Imubit's closed-loop AI optimization, layered on its existing advanced process controls, cut suboptimal coker drum cycles and the associated giveaway. The result was presented at the 2024 AFPM Summit. MPC is also a strategic investor in Flyscan, an AI aerial pipeline leak-detection company.", "metrics": "25% reduction in suboptimal coke-drum cycles", "vendors": "Imubit, Flyscan", "takeaway": "Reinforcement-learning optimizers layered on existing APC can deliver measurable unit gains without replacing controls.", "source": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes", "source_name": "Imubit", "id": 66}, {"company": "Williams", "segment": "Midstream", "hq": "US", "title": "New CIO; AI model out-trades 10-year gas trader", "date": "2025-08", "theme": "Trading & commercial", "summary": "CEO Chad Zamarin said Williams hired a new chief information officer with 'a lot of passion around AI' to drive cost efficiency and business transformation, including AI in project execution and analytics in market intelligence. In a month-long internal test marketing gas in Dallas-Fort Worth, young data-analytics staff building AI models beat a 10-year Sequent trader. Williams is also applying data to maintenance strategy.", "metrics": "AI modeler scored 96% of perfect-foresight execution vs. 94% for the veteran trader", "vendors": "", "takeaway": "Head-to-head pilots against top human performers are a persuasive way to build the internal case for AI in trading.", "source": "https://www.fool.com/earnings/call-transcripts/2025/08/05/williams-wmb-q2-2025-earnings-call-transcript/", "source_name": "The Motley Fool (Q2 2025 call transcript)", "id": 67}, {"company": "Kinder Morgan", "segment": "Midstream", "hq": "US", "title": "Strategic investment in Flyscan AI leak detection", "date": "2025-01", "theme": "Safety & emissions", "summary": "Kinder Morgan joined Flyscan Systems' strategic investor group alongside Marathon Petroleum, Enbridge and Hatch. Flyscan uses remote sensing and AI to detect liquid hydrocarbon leaks and automate right-of-way inspection from patrol aircraft, and is expanding into natural-gas detection. KM COO James Holland called the technology 'state-of-the-art'.", "metrics": "", "vendors": "Flyscan Systems", "takeaway": "Midstream operators are using corporate venture stakes to shape AI inspection tools for their own networks.", "source": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html", "source_name": "PR Newswire", "id": 68}, {"company": "SLB", "segment": "Oilfield services", "hq": "United States", "title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "date": "2025-06", "theme": "Partnership & investment", "summary": "SLB said its Lumi data and AI platform would be available on Mistral Compute, Mistral AI's new European AI infrastructure. Mistral's LLMs have been the primary generative models inside Lumi since the partnership began in early 2024, and this deal lets the platform be hosted entirely within European borders for data residency.", "metrics": "Lumi launched Sept 2024 (SLB Digital Forum, Monaco)", "vendors": "Mistral AI, Lumi, Delfi", "takeaway": "Where a buyer needs data sovereignty, offering models and hosting that stay in-region wins deals, so a sovereign or on-prem deployment option should be designed in from the start.", "source": "https://www.slb.com/news-and-insights/newsroom/updates/2025/slb%E2%80%99s-flagship-digital-platform-to-be-available-on-new-mistral-compute-ai-infrastructure", "source_name": "SLB", "id": 69}, {"company": "SLB", "segment": "Oilfield services", "hq": "United States", "title": "Tela agentic AI assistant launched for upstream", "date": "2025-11", "theme": "Agentic AI", "summary": "SLB launched Tela, an agentic AI assistant built on the Lumi platform that combines LLMs with domain foundation models. SLB says it can interpret well logs, predict drilling issues and optimize equipment, working alongside humans or on its own. It is available across SLB's digital platforms, in the cloud or on premises.", "metrics": "Five-step agent loop: observe, plan, generate, act, learn", "vendors": "Lumi, Delfi", "takeaway": "The largest service company is now selling agents as a layer across its whole software estate, so independent agent vendors need either deep workflow niches or to interoperate with Tela.", "source": "https://www.businesswire.com/news/home/20251103326704/en/SLB-Unveils-Groundbreaking-New-Agentic-AI-Technology-for-the-Energy-Industry", "source_name": "Business Wire", "id": 70}, {"company": "SLB", "segment": "Oilfield services", "hq": "United States", "title": "SLB and NVIDIA build 'AI Factory for Energy'", "date": "2026-03", "theme": "Partnership & investment", "summary": "SLB expanded its NVIDIA collaboration into three areas: modular DSX AI-factory data centers, an 'AI Factory for Energy' reference environment with domain generative models and industrial-scale agentic AI on SLB platforms, and accelerated computing across SLB software. SLB will make modular data-center components at its Louisiana technology center.", "metrics": "3.1 million sq ft Louisiana manufacturing center; partnership dates to 2008", "vendors": "NVIDIA, Lumi, Delfi", "takeaway": "Oilfield services companies are bundling compute infrastructure with domain AI, so pure-software AI vendors should partner with them for delivery rather than compete on infrastructure.", "source": "https://www.energy-pedia.com/news/general/slb-industrializes-ai-for-the-energy-industry-with-nvidia-203316", "source_name": "energy-pedia", "id": 71}, {"company": "SLB", "segment": "Oilfield services", "hq": "United States", "title": "SLB Digital Marketplace lists ~200 AI and digital products", "date": "2026-06", "theme": "Agentic AI", "summary": "SLB launched a curated marketplace where customers can find and deploy AI agents, domain models, skills, data connectors and apps, including Tela AI skills and agents. Every listing is certified against SLB standards for security and interoperability, and ISVs can apply to publish through the SLB partner program.", "metrics": "~200 products at launch; 30+ partners", "vendors": "Tela, Lumi, Delfi, Ocean store, ISVs", "takeaway": "This is effectively an app store for O&G agents, so startups can use it as a distribution channel, but SLB controls certification and the customer relationship.", "source": "https://www.businesswire.com/news/home/20260615078195/en/SLB-Launches-Digital-Marketplace-to-Scale-AI-and-Digital-Innovation-Across-Energy", "source_name": "Business Wire", "id": 72}, {"company": "SLB / Tachyus", "segment": "Oilfield services", "hq": "United States", "title": "SLB acquires Tachyus physics-ML reservoir startup", "date": "2026-05", "theme": "Partnership & investment", "summary": "SLB agreed to acquire Houston-based Tachyus, whose Aqueon platform combines machine learning with reservoir physics for waterflood management, pressure forecasting and production optimization. SLB plans to fold it into Delfi and Lumi. Terms were not disclosed.", "metrics": "Aqueon deployed on 7,500+ wells; customers include ConocoPhillips, Shell, Equinor, PDO", "vendors": "Tachyus Aqueon, Delfi, Lumi, ConocoPhillips, Shell, Equinor, Petroleum Development Oman", "takeaway": "Startups with proven physics-plus-ML products and named major customers are being bought by oilfield services companies, which makes an exit to SLB, Halliburton or Baker Hughes a realistic plan.", "source": "https://jpt.spe.org/slb-expands-digital-portfolio-with-acquisition-of-ai-firm-tachyus", "source_name": "JPT (SPE)", "id": 73}, {"company": "SLB / S&P Global Energy", "segment": "Oilfield services", "hq": "United States", "title": "SLB buys Kingdom, Petra upstream software for agentic AI", "date": "2026-04", "theme": "Partnership & investment", "summary": "SLB agreed to acquire S&P Global Energy's geoscience and petroleum-engineering software (Kingdom, Petra, Harmony Enterprise), which is widely used by US onshore operators. The two companies will co-develop AI models on S&P's upstream data using Lumi and Tela. Closing is expected in H2 2026 or early 2027.", "metrics": "~110,000 users across ~4,000 organizations (reported); data spans 113 countries", "vendors": "Kingdom, Petra, Harmony Enterprise, Lumi, Tela, S&P Global Energy", "takeaway": "SLB is buying the desktop tools independents already use in order to push agentic AI to US onshore operators, the same mid-market segment many AI startups target.", "source": "https://www.orphan-idle-wells.com/news/slb-acquires-the-software-running-us-oil-fields", "source_name": "Orphan & Idle Wells (SLB release summary)", "id": 74}, {"company": "Microsoft", "segment": "Big tech", "hq": "United States", "title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "date": "2025-11", "theme": "Enterprise GenAI rollout", "summary": "ADNOC, Masdar, XRG and Microsoft signed an agreement to co-develop and deploy AI agents across ADNOC's value chain and to supply energy for Microsoft's data centers. The deal builds on ADNOC's enterprise-wide Microsoft Copilot rollout, which began in November 2023.", "metrics": "40,000+ employees AI-trained; Copilot utilization >90%; 70,000+ hours/month productivity gained", "vendors": "Microsoft Copilot, Azure, ADNOC, Masdar, XRG", "takeaway": "Hyperscalers trade power-supply deals for AI footprint inside NOCs, and ADNOC's Copilot adoption numbers are a useful benchmark for workforce-enablement pitches.", "source": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "source_name": "ADNOC", "id": 75}, {"company": "Halliburton", "segment": "Oilfield services", "hq": "United States", "title": "Octiv Auto Frac: first fully automated frac with Coterra", "date": "2025-01", "theme": "Drilling & completions", "summary": "Halliburton and Coterra launched Octiv Auto Frac on the ZEUS electric frac platform. It automates stage execution, making thousands of pumping decisions from the operator's design without a human at the controls, and Coterra extended it to its remaining Permian completion programs.", "metrics": "17% increase in stage efficiency on initial rollout", "vendors": "Coterra Energy, ZEUS, Sensori", "takeaway": "Autonomy in completions reaches the field through service companies that own the equipment and control loop, which leaves software-only players to compete on design, analytics and planning.", "source": "https://worldoil.com/news/2025/1/6/halliburton-coterra-launch-first-fully-automated-hydraulic-fracturing-program/", "source_name": "World Oil", "id": 76}, {"company": "Halliburton", "segment": "Oilfield services", "hq": "United States", "title": "ZEUS IQ closed-loop intelligent frac with Chevron", "date": "2025-06", "theme": "Drilling & completions", "summary": "Halliburton debuted ZEUS IQ in May 2025, describing it as the industry's first closed-loop fracturing platform, combining Octiv Auto Frac with Sensori subsurface monitoring. With Chevron it ran feedback-driven completions in Colorado's DJ Basin, where the job plan was adjusted in real time without human intervention.", "metrics": "", "vendors": "Chevron, Octiv, Sensori", "takeaway": "The frontier has moved from alerts to closed-loop control, so AI builders need live subsurface feedback and actuation partners to compete.", "source": "https://www.chevron.com/newsroom/2025/q2/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "source_name": "Chevron", "id": 77}, {"company": "Halliburton / Sekal", "segment": "Oilfield services", "hq": "United States", "title": "Halliburton acquires Sekal drilling-automation software", "date": "2026-04", "theme": "Partnership & investment", "summary": "Halliburton bought Sekal from Sumitomo Corporation, combining Sekal's DrillTronics with its LOGIX automation and remote operations. In 2025 the two had delivered the first automated on-bottom drilling system, for Equinor on the Norwegian Continental Shelf.", "metrics": "DrillTronics used on 1,300+ wells; up to 25% reduction in well delivery time (reported)", "vendors": "Sekal DrillTronics, LOGIX, Equinor, Sumitomo", "takeaway": "Oilfield services companies are buying automation IP they first proved through partnerships, so a joint pilot with a major service company is a common route to acquisition.", "source": "https://finance.yahoo.com/sectors/energy/articles/halliburton-strengthens-automation-strategic-sekal-133400448.html", "source_name": "Yahoo Finance / Zacks", "id": 78}, {"company": "Baker Hughes / C3 AI", "segment": "Oilfield services", "hq": "United States", "title": "BakerHughesC3.ai JV renewed through 2028", "date": "2025-05", "theme": "Partnership & investment", "summary": "C3 AI and Baker Hughes renewed and expanded their joint venture through June 2028 to sell enterprise AI to oil, gas and chemicals, integrated with Cordant. Baker Hughes will also expand internal use of C3 sourcing, inventory and sustainability apps.", "metrics": "JV has generated >$500M revenue from O&G and chemicals since 2019; 5th amendment", "vendors": "C3 Agentic AI Platform, Cordant, Shell, Eni, QatarEnergy LNG, Petronas, ExxonMobil, LyondellBasell, Flint Hills", "takeaway": "Distribution through an oilfield services company moved more than $500M of AI software, so a channel JV with a service company can be worth more than direct sales.", "source": "https://c3.ai/c3-ai-and-baker-hughes-renew-and-expand-joint-venture-agreement/", "source_name": "C3 AI", "id": 79}, {"company": "Baker Hughes", "segment": "Oilfield services", "hq": "United States", "title": "Cordant built on Azure AI Foundry with Microsoft", "date": "2025-02", "theme": "Partnership & investment", "summary": "Baker Hughes signed an MOU with Microsoft to speed development and adoption of its Cordant industrial software suite, which is built with Azure AI Foundry. It targets predictive maintenance, operations optimization, safety and energy efficiency, and later releases (25.2.1, 26.1) added AI-driven risk intelligence.", "metrics": "", "vendors": "Microsoft Azure AI Foundry, Cordant", "takeaway": "Incumbents are standardizing on hyperscaler agent frameworks, so building on the same foundry or model stack makes integration and co-selling easier.", "source": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial", "source_name": "Baker Hughes", "id": 80}, {"company": "Baker Hughes", "segment": "Oilfield services", "hq": "United States", "title": "Leucipa adds genAI assistant, AI lift optimizer for Repsol", "date": "2025-06", "theme": "Production & operations", "summary": "Baker Hughes and Repsol agreed to deploy new Leucipa capabilities, including a generative AI virtual assistant. Repsol is piloting the AI-powered Leucipa Lift Optimizer on plunger-lift wells to reduce the time production engineers spend on individual wells, following an October 2024 MOU on AI field automation.", "metrics": "", "vendors": "Repsol, Leucipa", "takeaway": "Production-engineer time on lift optimization is a proven, fundable use case that majors now co-develop with oilfield services companies.", "source": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "source_name": "Baker Hughes", "id": 81}, {"company": "Weatherford", "segment": "Oilfield services", "hq": "United States", "title": "Weatherford Industrial Intelligence portfolio launched", "date": "2025-10", "theme": "Production & operations", "summary": "At FWRD 2025, Weatherford combined ForeSite, Centro, PetroVisor and new tools into 'Industrial Intelligence'. New pieces include the StrataCore production foundation model, the MetaFlow AI virtual flow meter, Canvas no-code data science, and the WFRD Software Launchpad SaaS hub.", "metrics": "Operates in ~75 countries", "vendors": "ForeSite, Centro, PetroVisor, StrataCore, Canvas", "takeaway": "Even mid-sized oilfield services companies now pitch production foundation models and no-code tools, so 'foundation model for production' is no longer a differentiator.", "source": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html", "source_name": "GlobeNewswire", "id": 82}, {"company": "NOV", "segment": "Oilfield services", "hq": "United States", "title": "NOV and AIQ pilot AI drilling on Max platform", "date": "2026-01", "theme": "Drilling & completions", "summary": "NOV and AIQ agreed to combine NOV's Max industrial IoT and edge-computing platform with AIQ's AI to design and test AI-driven drilling optimization. Work will run through proof-of-concept and pilot deployments in live operations.", "metrics": "", "vendors": "AIQ, NOV Max", "takeaway": "Equipment makers with edge platforms are looking for AI partners, so an AI builder can reach the rig without owning hardware by plugging into their edge layer.", "source": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "source_name": "NOV", "id": 83}, {"company": "Nabors / Corva", "segment": "Oilfield services", "hq": "Bermuda / United States", "title": "RigCLOUD Powered by Corva drilling intelligence", "date": "2025-04", "theme": "Drilling & completions", "summary": "Nabors expanded its alliance with Corva by integrating Corva's AI analytics into its RigCLOUD edge and cloud platform, sold as 'RigCLOUD Powered by Corva'. Corva also offers Corva Copilot, an AI assistant for drilling and completions.", "metrics": "", "vendors": "Corva, Corva Copilot, RigCLOUD", "takeaway": "Drilling contractors white-label startup analytics rather than building their own, which gives startups distribution across a whole rig fleet.", "source": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/", "source_name": "World Oil", "id": 84}, {"company": "Helmerich & Payne", "segment": "Oilfield services", "hq": "United States", "title": "FlexRobotics rig-floor robotics deployed for supermajor", "date": "2026-02", "theme": "Drilling & completions", "summary": "H&P said it deployed its FlexRobotics rig-floor automation system on a rig for a supermajor in the Permian Basin. The system combines robotic arms, automation software and safety controls to keep people out of high-risk zones during drilling and tripping.", "metrics": "", "vendors": "FlexRobotics", "takeaway": "Physical automation on land rigs is starting to scale, and it will generate new data streams and workflows for software AI to use.", "source": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results", "source_name": "Business Wire", "id": 85}, {"company": "Patterson-UTI", "segment": "Oilfield services", "hq": "United States", "title": "Cortex ML autodriller and REX monitoring drive rig margins", "date": "2025-07", "theme": "Drilling & completions", "summary": "Patterson-UTI credits its Cortex automation platform (machine-learning autodriller), REX cloud early-alert monitoring and its Digital Performance Center with differentiated rig performance. On completions it uses Vertex frac automation and FleetStream real-time data feeds.", "metrics": "", "vendors": "Cortex, REX, Vertex, FleetStream, EOS Completions Platform", "takeaway": "Contractors market their digital and ML stack as a margin driver, so AI vendors should frame their value in the contractor's terms, such as revenue per rig and performance pricing.", "source": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "source_name": "SEC (Patterson-UTI 8-K)", "id": 86}, {"company": "Liberty Energy", "segment": "Oilfield services", "hq": "United States", "title": "Liberty launches Forge LLM for frac asset orchestration", "date": "2025-10", "theme": "Drilling & completions", "summary": "Liberty launched Forge, its own large language model for intelligent asset orchestration, which works with StimCommander, its AI-driven automated rate and pressure control software. The company reported fleet-level automation gains in its Q3 2025 results.", "metrics": "StimCommander: 65% faster time to target injection rate; 5-10% hydraulic efficiency gain; digiPrime maintenance savings >30%", "vendors": "Forge, StimCommander, digiPrime", "takeaway": "Frac companies are building proprietary LLMs to orchestrate their own equipment fleets, which leaves little room for outside vendors in pressure pumping itself.", "source": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "source_name": "Liberty Energy", "id": 87}, {"company": "Palantir", "segment": "Big tech", "hq": "United States", "title": "APA extends multi-year Palantir deal, adds AIP", "date": "2024-09", "theme": "Enterprise GenAI rollout", "summary": "APA signed a multi-year, multimillion-dollar extension of its 2021 Palantir enterprise agreement, adding the AIP platform. Palantir already supports APA's operational planning, supply chain, maintenance, production optimization and contract management, and AIP is used to flag anomalies in contracts and invoices against source systems.", "metrics": "Multi-year, multimillion-dollar", "vendors": "Palantir AIP, Foundry", "takeaway": "Palantir expands inside E&Ps from an operational data layer into back-office genAI such as invoice and contract checks, so land-and-expand from one ontology is the model to beat.", "source": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html", "source_name": "StockTitan (Business Wire release)", "id": 88}, {"company": "AWS", "segment": "Big tech", "hq": "United States", "title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "date": "2024-08", "theme": "Production & operations", "summary": "bpx energy, which runs bp's US onshore assets, built Xpert with the AWS Generative AI Innovation Center. Xpert is a Bedrock-based chat tool that turns employee questions into SQL or Python, runs the code against bpx production data and answers in text, and is intended to extend to mobile and voice for field staff.", "metrics": "bpx targeting 30% volume growth by 2025 vs 2022", "vendors": "Amazon Bedrock, AWS Generative AI Innovation Center, bpx energy", "takeaway": "Text-to-SQL over production data is a common first genAI use case for E&Ps, and hyperscaler innovation teams build it at little or no cost, so vendors need more defensible depth.", "source": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/", "source_name": "AWS for Industries blog", "id": 89}, {"company": "Snowflake", "segment": "Big tech", "hq": "United States", "title": "Snowflake launches Energy Solutions with 30+ partner apps", "date": "2026-01", "theme": "Enterprise GenAI rollout", "summary": "Snowflake launched Energy Solutions for its AI Data Cloud, bundling governance, 30+ partner-built apps, industry datasets, Snowflake Intelligence and Cortex AI. Named energy customers include ExxonMobil and Expand Energy.", "metrics": "30+ partner apps; 13,900+ Snowflake customers overall", "vendors": "Snowflake Cortex AI, Snowflake Intelligence, ExxonMobil, Expand Energy, CARTO, Itron", "takeaway": "Data platforms are becoming the distribution layer for O&G AI apps, so listing as a partner app on Snowflake or Databricks puts a vendor next to data the operator already governs.", "source": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "source_name": "Snowflake", "id": 90}, {"company": "Databricks", "segment": "Big tech", "hq": "United States", "title": "Data Intelligence Platform for Energy with genAI", "date": "2024-04", "theme": "Enterprise GenAI rollout", "summary": "Databricks launched an energy-specific Data Intelligence Platform bringing generative AI to the sector, with partners including AVEVA, Seeq, Deloitte and Capgemini. Shell's digital innovation VP said Databricks had accelerated its analytics and AI.", "metrics": "10,000+ Databricks orgs globally", "vendors": "Shell, TotalEnergies, Chevron Phillips Chemical, Cosmo Energy, Wood Mackenzie, AVEVA, Seeq", "takeaway": "Majors' AI and ML work often runs on Databricks, so building natively on its lakehouse (Unity Catalog, model serving) lowers the barrier to buying.", "source": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html", "source_name": "PR Newswire", "id": 91}, {"company": "Cognite", "segment": "Startup / scaleup", "hq": "Norway", "title": "Aker BP expands AI-first strategy with Cognite Atlas AI", "date": "2025-09", "theme": "Agentic AI", "summary": "Aker BP deepened its Cognite partnership to deploy agent-based applications built with Cognite Atlas AI on Cognite Data Fusion. Use cases include root-cause analysis, document automation, barrier management, equipment registration, predictive maintenance and production chemistry.", "metrics": "Root-cause-analysis agents cut engineering time >70%", "vendors": "Aker BP, Cognite Atlas AI, Cognite Data Fusion", "takeaway": "A contextualized knowledge graph underneath is what makes agents credible in operations, and time saved per engineering task is the proof metric buyers respond to.", "source": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/", "source_name": "World Oil", "id": 92}, {"company": "Cognite", "segment": "Startup / scaleup", "hq": "Norway", "title": "Cognite Atlas AI major release, low-code agent workbench", "date": "2025-09", "theme": "Agentic AI", "summary": "Cognite shipped a major Atlas AI release, which it describes as a low-code industrial agent workbench on real-time OT, IT and engineering data. Customers named include Aker BP, ADNOC Offshore, Hess and HMH, with SLB among its partners, and the company says it adds new customers weekly.", "metrics": "New customers onboarded weekly (company claim)", "vendors": "Aker BP, ADNOC Offshore, Hess, HMH, Celanese, Idemitsu, SLB, Radix", "takeaway": "A low-code agent builder where operators assemble their own agents is the competing model to custom agent services.", "source": "https://www.businesswire.com/news/home/20250909488860/en/Cognite-Atlas-AI-Drives-Customer-Momentum-with-New-Major-Release-Accelerates-Adoption-of-Agentic-AI-to-Unlock-More-Industrial-Value", "source_name": "Business Wire", "id": 93}, {"company": "Cognite", "segment": "Startup / scaleup", "hq": "Norway", "title": "Cognite integrates NVIDIA accelerated computing for agents", "date": "2025-10", "theme": "Maintenance & reliability", "summary": "Cognite integrated its Atlas AI and Data Fusion platform with NVIDIA accelerated computing and software. The first use case is real-time anomaly detection in time-series data with Celanese, and Cognite later integrated NVIDIA NV-Tesseract for forecasting.", "metrics": "", "vendors": "NVIDIA, Celanese, NV-Tesseract", "takeaway": "Time-series anomaly detection at scale is the core agent workload in operations, and partnering with a GPU vendor adds credibility and lowers compute cost.", "source": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "source_name": "Cognite", "id": 94}, {"company": "Novi Labs", "segment": "Startup / scaleup", "hq": "United States", "title": "Novi Labs raises $35M for AI-first energy analytics", "date": "2025-06", "theme": "Partnership & investment", "summary": "Novi Labs raised $35M from Invictus Growth Partners to speed its product roadmap, expand go-to-market and fund acquisitions. Its ML well forecasts feed upstream capital-allocation decisions.", "metrics": "$35M; >$50B/yr of capital allocation decisions run through the platform", "vendors": "Invictus Growth Partners, Shell, ExxonMobil, Chevron, Devon (reported users)", "takeaway": "Growth equity is backing vertical AI analytics for upstream investing, where accurate forecasts on proprietary data are the moat.", "source": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "source_name": "Novi Labs", "id": 95}, {"company": "Novi Labs", "segment": "Startup / scaleup", "hq": "United States", "title": "Novi acquires Turing Analytics and RBN Energy", "date": "2026-05", "theme": "Partnership & investment", "summary": "Novi Labs acquired Calgary-based Turing Analytics (EVA, an April 2026 deal) and Houston-based RBN Energy (a May 2026 deal). The acquisitions connect upstream analytics with Canadian coverage and midstream and commodity-market intelligence in one decision platform.", "metrics": "RBN reaches 120,000+ monthly users", "vendors": "Turing Analytics (EVA), RBN Energy", "takeaway": "Well-funded vertical AI players are consolidating data and audience assets, so owning proprietary content becomes a competitive lever.", "source": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "source_name": "Novi Labs", "id": 96}, {"company": "Enverus", "segment": "Software incumbent", "hq": "United States", "title": "Instant Analyst genAI with cited answers", "date": "2024-02", "theme": "Trading & commercial", "summary": "Enverus launched Instant Analyst, generative AI over its Intelligence Vault research that returns natural-language answers with document citations to limit hallucinations. It runs on Amazon Bedrock and later expanded to other products, including Courthouse records.", "metrics": "Searches tens of thousands of proprietary research docs", "vendors": "Amazon Bedrock, Enverus Intelligence Vault, PRISM, Tenaris, Ironwood Midstream", "takeaway": "Answers with citations over proprietary data are what buyers now expect as a minimum for O&G genAI.", "source": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/", "source_name": "Enverus", "id": 97}, {"company": "Enverus", "segment": "Software incumbent", "hq": "United States", "title": "Blackstone to acquire Enverus (~$6.5B reported)", "date": "2025-08", "theme": "Partnership & investment", "summary": "Blackstone funds agreed to acquire Enverus, the energy data analytics and SaaS platform, from Hellman & Friedman and Genstar Capital. Reuters reported the deal at roughly $6.5B.", "metrics": "~$6.5B (reported)", "vendors": "Blackstone, Hellman & Friedman, Genstar", "takeaway": "Private equity pays a premium for proprietary energy data that genAI workflows depend on, so owning data is what drives valuation.", "source": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "source_name": "Blackstone", "id": 98}, {"company": "Enverus", "segment": "Software incumbent", "hq": "United States", "title": "Enverus AI launched as 'AI operating system for energy'", "date": "2025-12", "theme": "Enterprise GenAI rollout", "summary": "Enverus unveiled Enverus AI, a governed system built on its energy datasets. It covers asset lifecycle management, document analysis, title verification, field safety monitoring and M&A deal screening.", "metrics": "8,000+ customer companies; claims '10X workforce impact'", "vendors": "Enverus", "takeaway": "Title, document and deal-screening workflows are being built into the incumbent data vendor's product, so specialist startups there face bundling pressure.", "source": "https://www.enverus.com/newsroom/enverus-ai-powering-the-next-era-of-energy-intelligence/", "source_name": "Enverus", "id": 99}, {"company": "Enverus", "segment": "Software incumbent", "hq": "United States", "title": "Enverus ONE governed AI platform with workflow 'Flows'", "date": "2026-04", "theme": "Agentic AI", "summary": "Enverus launched Enverus ONE, which pairs frontier models with Astra, its own energy model, as an execution layer for energy workflows. The first 'Flows' are AFE Evaluation, Current Production Valuation, Project Siting and QuickStart, running in a SOC 2 Type II private environment.", "metrics": "8,000+ companies in 50 countries; 2.7 PB data; 350M+ courthouse records; $500B+ annual transactions tracked", "vendors": "Astra, frontier LLMs", "takeaway": "Packaging agents as named, outcome-specific workflows (such as AFE evaluation) on proprietary data is the product pattern to borrow.", "source": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "source_name": "Enverus", "id": 100}, {"company": "Collide", "segment": "Startup / scaleup", "hq": "United States", "title": "Collide raises $5M seed for O&G genAI platform", "date": "2025-04", "theme": "Partnership & investment", "summary": "Houston-based Collide closed a $5M seed round led by Mercury Fund, with operator angels Bryan Sheffield, Billy Quinn and David Albin. It serves public, private and PE-backed E&Ps and grew from an industry community platform.", "metrics": "$5M seed; 6,000+ community users growing 11% MoM (at raise)", "vendors": "Mercury Fund", "takeaway": "Building a practitioner community first gave Collide a distribution base before selling enterprise AI.", "source": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "source_name": "Business Wire", "id": 101}, {"company": "Collide", "segment": "Startup / scaleup", "hq": "United States", "title": "RIGGS energy LLM automates RRC filings", "date": "2026-05", "theme": "Back office & corporate", "summary": "Collide rolled out RIGGS, an energy LLM that returns cited answers from authoritative sources. With Winn Resources it automated Texas Railroad Commission regulatory filings, and Collide says RIGGS beat general frontier models on a petroleum-engineering exam subset.", "metrics": "Filing time cut from hours to 20 min for 50 wells (>95%); 67.5% on 40-question PE exam subset vs Grok 4 62.5%, Claude Sonnet 4.5 52.5%", "vendors": "Winn Resources", "takeaway": "Regulatory filings are a narrow, measurable back-office wedge for small operators, and domain benchmarks against frontier models are becoming a marketing tool.", "source": "https://energycapitalhtx.com/collide-riggs-ai-platform", "source_name": "Energy Capital HTX", "id": 102}, {"company": "Collide", "segment": "Startup / scaleup", "hq": "United States", "title": "Battalion Oil makes strategic investment in Collide", "date": "2026-09", "theme": "Partnership & investment", "summary": "Public E&P Battalion Oil made an undisclosed strategic investment in Collide and is deploying its Riggs AI-native operations system across its upstream operations. Battalion plans to unify more than 100 TB of operational data, with Riggs reaching its production engineers around day 90.", "metrics": "100+ TB data; ~90-day time to production engineers; Collide cites $20B addressable market", "vendors": "Battalion Oil, Riggs", "takeaway": "Operators investing in their AI vendors aligns incentives and creates reference customers, and a stated 90-day time-to-value target is a useful benchmark.", "source": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "source_name": "GlobeNewswire", "id": 103}, {"company": "Applied Computing", "segment": "Startup / scaleup", "hq": "United Kingdom", "title": "$20M Series A for Orbital O&G foundation model", "date": "2026-07", "theme": "Production & operations", "summary": "London-based Applied Computing raised a $20M Series A led by KBR, with Databricks Ventures participating. Its Orbital foundation model combines time-series, physics and language modeling to predict plant state, find anomalies and simulate fixes.", "metrics": "$20M Series A; 'double-digit millions' ARR within ~18 months; facilities use <8% of sensor data", "vendors": "KBR (INSITE 3.0), Databricks Ventures, Wipro", "takeaway": "Strategic investors (an engineering firm and a data platform) give distribution, and a hybrid physics-time-series-language model is attracting funding.", "source": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/", "source_name": "TechCrunch", "id": 104}, {"company": "Seeq", "segment": "Software incumbent", "hq": "United States", "title": "Seeq Intelligence with Agent Q and custom agents", "date": "2026-03", "theme": "Agentic AI", "summary": "Seeq launched Seeq Intelligence, adding Agent Q (a natural-language AI analyst), build-your-own agents for multi-step workflows, secure connections to customer systems and unstructured-document extraction. It is included in the Seeq Enterprise package, and ExxonMobil and Marathon Petroleum have presented at Seeq events.", "metrics": "", "vendors": "ExxonMobil, Marathon Petroleum, PBF", "takeaway": "Time-series analytics incumbents are bundling agents into existing enterprise tiers rather than charging separately, which compresses what standalone agent vendors can charge.", "source": "https://www.prnewswire.com/news-releases/introducing-seeq-intelligence-bridging-industrial-ai-and-human-expertise-for-smarter-operational-decisions-302702464.html", "source_name": "PR Newswire", "id": 105}, {"company": "AVEVA", "segment": "Software incumbent", "hq": "United Kingdom", "title": "AVEVA embeds agentic AI, MCP across CONNECT", "date": "2026-05", "theme": "Agentic AI", "summary": "At AVEVA World 2026, AVEVA announced agentic twin-builder mapping, an AI-enriched industrial knowledge graph, and MCP integrations that let LLMs read operational data. It also announced a CONNECT major release planned for Q1 2027 and deeper Microsoft, Snowflake and ServiceNow integrations. Its Industrial AI Assistant runs on Azure OpenAI.", "metrics": "3,500+ delegates; 5,000 partners", "vendors": "Microsoft Azure OpenAI, Snowflake, ServiceNow, Cognizant, Crosser", "takeaway": "Industrial historians now expose data through MCP, so AI builders should plan to connect via MCP to PI and CONNECT rather than build custom connectors.", "source": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "source_name": "AVEVA", "id": 106}, {"company": "Emerson / AspenTech", "segment": "Software incumbent", "hq": "United States", "title": "Emerson completes AspenTech buyout for industrial AI", "date": "2025-03", "theme": "Partnership & investment", "summary": "Emerson bought the remaining AspenTech shares at $265 per share in cash, making AspenTech a wholly owned subsidiary. The deal combines automation hardware with optimization, APM and subsurface software aimed at autonomous operations.", "metrics": "Minority stake ~$7.2B; total AspenTech investment ~$17B (reported)", "vendors": "Emerson, AspenTech", "takeaway": "Automation vendors are consolidating industrial software to own the path to autonomy, which leaves independent AI players to become partners or acquisition targets.", "source": "https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "source_name": "Emerson", "id": 107}, {"company": "McKinsey & Company", "segment": "Survey & research", "hq": "United States", "title": "AI could unlock $230B a year in upstream", "date": "2026-08", "theme": "Adoption data", "summary": "McKinsey sized AI value across the upstream value chain: $65B a year near-term with today's technology, $125B as proven tools scale, and $230B at full potential with autonomous operations. The figures are net of more than $30B a year in implementation costs. Value is concentrated: the top 10 use cases account for nearly half of it, and production optimization and drilling are the largest pools.", "metrics": "$65B / $125B / $230B annual value; >$30B/yr implementation cost; top 10 use cases ~50% of value, top 60 ~95%; <20% of deployments track AI value KPIs; +$35B/yr reserves value from exploration; OFS revenue displacement $17B near-term to $60B", "vendors": "McKinsey", "takeaway": "Fewer than 20% of deployments track AI value, so building value measurement into the enablement program is part of the product.", "source": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value", "source_name": "Oil & Gas Journal", "id": 108}, {"company": "Rystad Energy", "segment": "Survey & research", "hq": "Norway", "title": "Digital and AI: a $500B upstream opportunity", "date": "2026-05", "theme": "Adoption data", "summary": "Rystad estimates that digitalization and AI could create close to $500B of cumulative value for E&Ps from 2026 to 2030, split roughly equally between cost reduction and production gains. E&Ps already investing are expected to capture an extra $80B a year by 2030 compared with 2025. Rystad points to ADNOC ($500M of AI value in 2023) and Equinor as proof points.", "metrics": "~$500B cumulative 2026-2030; +$80B/yr by 2030 vs 2025; five priority areas: drilling optimization, autonomous robotics, predictive maintenance, reservoir management, logistics", "vendors": "Rystad Energy, ADNOC, Equinor", "takeaway": "Frame use-case portfolios around the few value pools the analysts agree on (drilling, maintenance, reservoir) instead of spreading effort thin.", "source": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas", "source_name": "Rystad Energy", "id": 109}, {"company": "Rystad Energy", "segment": "Survey & research", "hq": "Norway", "title": "Digitalization could save $320B by 2030", "date": "2025-12", "theme": "Adoption data", "summary": "Rystad's supply-chain team estimated $320B of oil and gas savings from digitalization through 2030 and called the figure modest. It named high upfront hardware and software costs, cybersecurity requirements and aging infrastructure as the main barriers, especially for smaller operators and service firms.", "metrics": "$320B savings through 2030; SLB digital targeting ~35% margins in 2025; Viridien digital/data revenue $787M (+17%)", "vendors": "Rystad Energy, SLB, Halliburton, NOV, Baker Hughes, Viridien", "takeaway": "Smaller operators face cost and cyber barriers, so modular, low-infrastructure enablement offerings are an underserved gap.", "source": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/", "source_name": "World Oil", "id": 110}, {"company": "Boston Consulting Group", "segment": "Survey & research", "hq": "United States", "title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "date": "2025-08", "theme": "Strategy & org", "summary": "BCG's CEO briefing argues AI could raise O&G EBIT by 30-70% within five years (about +30% upstream, +50% refining, +70% fuel retail). It splits AI plays into Deploy (GenAI productivity tools, 10-15% productivity), Reshape (end-to-end workflow redesign) and Invent (new business models). Its playbook calls for executive sponsorship, data and platform upgrades, and upskilling of key roles with AI literacy and change management.", "metrics": "+30-70% EBIT potential in 5 yrs; Deploy/Reshape/Invent chart shows 72%/47%/22% of O&G companies with initiatives; top GenAI everyday uses: knowledge management 69%, code 67%, meeting summaries 62% (N=188); leaders put 80% of AI investment into Reshape/Invent", "vendors": "BCG, Cognite, SparkCognition, Microsoft, OpenAI (ChatGPT Enterprise), SLB (Delfi), Aker BP, bp", "takeaway": "Seat licences (Deploy) are table stakes. Most of the value sits in redesigning a few core workflows, so enablement has to go past prompt training.", "source": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf", "source_name": "BCG", "id": 111}, {"company": "Bain & Company", "segment": "Survey & research", "hq": "United States", "title": "Energy execs: fewer than 20% scaling AI with results", "date": "2026-02", "theme": "Adoption data", "summary": "In Bain's 2026 survey of energy and natural resources executives, about two-thirds said their company is experimenting with AI or running pilots without the outcomes it wanted. Only about a quarter have scaled applications with measurable impact, and fewer than 10% are running a companywide AI transformation. Nearly half named unclear outcomes or no link to business value as the biggest impediment, followed by shortages of technical expertise and poor data governance.", "metrics": "~2/3 stuck in experiments/pilots; ~1/4 scaling with measurable impact; <10% companywide transformation; <20% scaling with results; ~50% cite unclear value link as top barrier; 800+ executives surveyed", "vendors": "Bain & Company", "takeaway": "The top barrier is unclear business value, not technology, so tie each use case to an owner and a P&L metric before building it.", "source": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/", "source_name": "Bain & Company", "id": 112}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "O&G execs: 85% say reskilling decides success; 29% do it", "date": "2025-03", "theme": "Workforce enablement", "summary": "EY's Future of Energy Survey (1,020 respondents, Nov-Dec 2024) found a large gap between intent and action on workforce reskilling. 85% of oil and gas executives say reskilling will determine success over the next five years, but only 29% are retraining existing employees. 58% of oil and gas companies say their culture hinders adoption of digital technology.", "metrics": "85% say reskilling critical vs 29% retraining; 81% plan more reliance on vendors/contractors; 58% say culture impedes digital adoption; 92% of energy cos plan digital investment vs 27% retraining", "vendors": "EY", "takeaway": "The market gap is in doing the reskilling, not in convincing people it matters.", "source": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey", "source_name": "EY", "id": 113}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "Energy leaders struggle to link AI to productivity", "date": "2025-12", "theme": "Adoption data", "summary": "Wave 4 of EY's US AI Pulse Survey found that energy leaders see AI as a catalyst for transformation but have trouble proving its impact. 56% of energy senior leaders struggle to link productivity gains directly to AI, compared with 35% across all industries. 72% say more training is needed to communicate AI productivity gains accurately. EY also notes that employee job-security concerns are holding back experimentation.", "metrics": "56% energy vs 35% all-industry struggle to attribute gains; 72% need more training (vs 50%); 78% see gains as catalyst (vs 55%); 72% report rising responsible-AI interest (+25 pts since Jul 2024)", "vendors": "EY", "takeaway": "Energy lags other sectors on measuring AI gains, so offering an attribution method alongside training is a differentiator.", "source": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption", "source_name": "EY", "id": 114}, {"company": "EY", "segment": "Survey & research", "hq": "United Kingdom", "title": "96% of O&G execs cite data and vendor challenges", "date": "2025-01", "theme": "Adoption data", "summary": "EY reports that 96% of oil and gas executives see managing alliance and vendor relationships, along with data accessibility, as a challenge for AI and emerging technology. 96% also cite connecting data across organizational silos. EY says the hard part starts when a proof of concept is handed to the team responsible for scaling it.", "metrics": "96% cite vendor/alliance management and data accessibility; 96% cite cross-silo data connection", "vendors": "EY", "takeaway": "Plan the pilot-to-production handoff (data plumbing, vendor orchestration, ownership) as part of any use case, not as an afterthought.", "source": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company", "source_name": "EY", "id": 115}, {"company": "Deloitte", "segment": "Survey & research", "hq": "United States", "title": "AI heading past 50% of O&G IT spend by 2029", "date": "2025-10", "theme": "Adoption data", "summary": "Deloitte's 2026 Oil and Gas Industry Outlook projects that AI and GenAI will grow from less than 20% of US O&G IT spending today to more than 50% by 2029. About half of AI investment goes to process optimization. Early adopters of AI-driven maintenance report up to 40% fewer equipment failures and $10M a year in savings. The outlook expects GenAI, agentic AI and real-time analytics to move from pilots to enterprise deployment in 2026.", "metrics": "AI <20% of IT spend now to >50% by 2029; up to 40% fewer failures; $10M/yr savings; 140+ hours downtime prevented, 1.6% uptime protected (one case); 66% of workforce in mechanically intensive roles", "vendors": "Deloitte", "takeaway": "As budgets swing toward AI, the scarce input becomes people who can use it, which is a direct opening for enablement.", "source": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html", "source_name": "Deloitte Insights", "id": 116}, {"company": "Deloitte", "segment": "Survey & research", "hq": "United States", "title": "1.2M US energy and chemicals workers need upskilling", "date": "2025-05", "theme": "Workforce enablement", "summary": "Deloitte's workforce study says more than 60% of the US energy and chemicals workforce (about 1.2 million people) will need upskilling in new technologies, process operations and analytics, and another 534,000 will need reskilling. It expects about one-third of the tasks analyzed to be led by machines and agentic AI with humans in the loop.", "metrics": "1.2M workers (>60%) need upskilling; 534K need reskilling; ~1/3 of tasks human-in-the-loop with AI leading; workforce to grow 4.1% by 2033", "vendors": "Deloitte", "takeaway": "Upskilling demand is workforce-wide, not only for engineers, so programs need role-based tracks that include field and operations staff.", "source": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html", "source_name": "Deloitte via PR Newswire", "id": 117}, {"company": "KPMG", "segment": "Survey & research", "hq": "Netherlands", "title": "Energy CEOs: only 40% actively reskilling for AI", "date": "2025-11", "theme": "Workforce enablement", "summary": "KPMG's 2025 Global Energy, Natural Resources and Chemicals CEO Outlook found that CEOs are confident in AI's potential but lag on workforce action. 80% agree their leadership understands AI's disruptive potential, but only 40% are actively reskilling or upskilling the roles AI affects. 82% believe AI can support emissions reduction and energy efficiency.", "metrics": "40% actively reskilling AI-affected roles; 80% say leadership understands AI disruption; 82% see AI supporting emissions reduction", "vendors": "KPMG", "takeaway": "CEO awareness has outrun action on skills, so position enablement as how the AI strategy gets executed.", "source": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html", "source_name": "KPMG", "id": 118}, {"company": "DNV", "segment": "Survey & research", "hq": "Norway", "title": "Only 12% of O&G firms use AI in live operations", "date": "2024-10", "theme": "Adoption data", "summary": "DNV's 2024 Energy Industry Insights research found that oil and gas AI adoption is still low even though 62% say digitalization is central to strategy. 47% were at the planning or pilot stage, 12% used AI in live operations, and only 3% had advanced, highly integrated use. DNV says trust is what the industry needs to move beyond this low base.", "metrics": "12% live ops; 3% advanced/integrated; 47% planning/pilot; 62% digitalization central to strategy; 20% cite regulatory barriers; AI-in-O&G market ~$3B (2024) to $5.2B (2029, Mordor)", "vendors": "DNV, Mordor Intelligence", "takeaway": "Assurance (showing that AI is safe, secure and validated) is itself an adoption lever in a safety-critical industry.", "source": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "source_name": "DNV", "id": 119}, {"company": "DNV", "segment": "Survey & research", "hq": "Norway", "title": "60% plan AI deployment; human-AI interaction a key risk", "date": "2025-09", "theme": "Safety & emissions", "summary": "At SPE Offshore Europe 2025, DNV reported that 60% of 1,160 surveyed energy companies planned to deploy AI-driven applications in 2025, up from 47% in 2024. It proposed a systemic safety framework that combines safety engineering with AI assurance across the lifecycle, and it flagged human-AI interaction as a primary risk area.", "metrics": "60% planning AI deployment in 2025 (vs 47% in 2024); 1,160 companies surveyed", "vendors": "DNV, SPE", "takeaway": "Teach how to supervise AI (when to override, how to verify), not just how to use it.", "source": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/", "source_name": "Oilfield Technology", "id": 120}, {"company": "DNV Cyber", "segment": "Survey & research", "hq": "Norway", "title": "65% of energy firms see cyber as top risk", "date": "2025-01", "theme": "Safety & emissions", "summary": "DNV's Energy Cyber Priority 2025 surveyed 375 energy professionals. 65% say their leadership views cybersecurity as the greatest risk to the business, and 71% say their organization is more vulnerable to OT cyber events than ever. 57% admit their OT defences lag their IT defences, and the report names embracing AI and embedding new workforce skills as core challenges.", "metrics": "65% cyber = greatest risk; 71% more OT-vulnerable (vs 64% 2023); 57% OT defences lag IT; 67% expect more OT security investment; n=375", "vendors": "DNV Cyber", "takeaway": "Any agentic or OT-adjacent AI program has to be designed with OT security from the start; cyber is a gating objection.", "source": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/", "source_name": "DNV", "id": 121}, {"company": "International Energy Agency", "segment": "Survey & research", "hq": "France", "title": "IEA: oil and gas an early AI adopter", "date": "2025-04", "theme": "Adoption data", "summary": "The IEA's Energy and AI special report describes oil and gas as an early AI adopter in exploration, production, maintenance, safety and methane detection. Oil and gas companies ran 24 of the world's 500 fastest supercomputers in 2024, up from 11 in 2000, and their computing capacity grew about 70% a year. The IEA names missing data access, digital infrastructure, skills, and digital and physical security concerns as the main barriers.", "metrics": "O&G supercomputers in TOP500: 11 (2000) to 24 (2024); ~70%/yr compute capacity growth", "vendors": "IEA", "takeaway": "Even the IEA ranks skills alongside data and security as a core barrier, which supports treating enablement as infrastructure.", "source": "https://www.iea.org/reports/energy-and-ai", "source_name": "IEA", "id": 122}, {"company": "PwC", "segment": "Survey & research", "hq": "United Kingdom", "title": "20% of companies capture 74% of AI value", "date": "2026-04", "theme": "Adoption data", "summary": "PwC's 2026 AI Performance Study of 1,217 senior executives across 25 sectors found that 20% of organizations capture about three-quarters of AI's economic value. The rest are largely stuck in pilot mode. Leaders use AI for growth and business-model reinvention rather than productivity alone. The study is cross-industry, not specific to oil and gas.", "metrics": "74% of value to 20% of firms; n=1,217 across 25 sectors", "vendors": "PwC", "takeaway": "AI value is winner-take-most, so the case for enablement is competitive positioning, not just efficiency.", "source": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html", "source_name": "PwC", "id": 123}, {"company": "Novi Labs", "segment": "Survey & research", "hq": "United States", "title": "Reservoir engineer AI use rises to 56%", "date": "2025-06", "theme": "Subsurface & exploration", "summary": "Novi Labs surveyed more than 150 US reservoir engineers and found AI adoption rose from 43% in 2024 to 56% in 2025. Well forecasting use rose from 58% to 75%. 84% reported problems with the quality and accessibility of upstream public data, and teams using proprietary data reported much higher success than those relying only on public sources.", "metrics": "Adoption 43% to 56%; well-forecasting use 58% to 75%; 84% cite public-data quality issues; 73% vs 41% success (proprietary vs public data); hybrid build/buy success 72%; confidence in type curves 39% to 23%", "vendors": "Novi Labs", "takeaway": "Engineers now own data quality (74%), so enablement should teach data stewardship alongside AI tools.", "source": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/", "source_name": "Novi Labs", "id": 124}, {"company": "Shell", "segment": "Supermajor", "hq": "United Kingdom", "title": "Shell's DIY citizen developer program passes 4,000 builders", "date": "2023-06", "theme": "Workforce enablement", "summary": "Shell's DIY (Do IT Yourself) program uses Microsoft Power Platform to let non-IT staff build low-code apps and automations, with a dedicated center of excellence in digital engineering. The original target was 500 trained developers, and the program now has more than 4,000 active citizen developers. It is supported by embedded coaches, bootcamps, persona-based learning paths, hackathons and gamification.", "metrics": "Target 500 to >4,000 active citizen developers; lifting-approval app cut approval time ~60% (2.5 hrs to 1 hr); furnace app covers 20 units x 64 burners", "vendors": "Microsoft (Power Platform)", "takeaway": "Shell's playbook (CoE, embedded coaches, persona-based paths, hackathons) transfers directly from low-code to GenAI and agent building.", "source": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk", "source_name": "Microsoft Customer Stories", "id": 125}, {"company": "Aker BP", "segment": "Independent E&P", "hq": "Norway", "title": "Aker BP: Copilot for all staff plus champion network", "date": "2025-06", "theme": "Workforce enablement", "summary": "Aker BP started with 300 Copilot licences and committed to giving every employee one in 2025. It created an AI adoption champion network with accountability at executive-committee level. Training includes help-desk support, AI-fundamentals learning paths and hands-on prompt-engineering sessions, and 'Digital and AI competence' became one of five core competencies for all employees. Staff and small teams are building their own agents, such as knowledge agents.", "metrics": "300 initial licences to all employees; 96% active users over four weeks (Viva dashboard); AI competence is 1 of 5 core competencies", "vendors": "Microsoft", "takeaway": "Putting AI competence into the formal competency model and making executives accountable for champions is what moves usage to 96%.", "source": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "source_name": "Microsoft Customer Stories", "id": 126}, {"company": "Society of Petroleum Engineers", "segment": "Industry body", "hq": "United States", "title": "SPE Forum: AI in Upstream E&P 2030", "date": "2026-04", "theme": "Workforce enablement", "summary": "SPE held a closed forum in Santa Fe (April 7-10, 2026), following its 2024 AI Forum, on how AI is reshaping upstream engineering. Sessions covered which skills preserve the human edge, whether AI can acquire tacit knowledge, how to build trust in models, and how to design human-AI teaming and ethical AI use by 2030.", "metrics": "", "vendors": "SPE", "takeaway": "The profession's own agenda is about trust, tacit knowledge and human-AI teaming, so enablement content should address these as well as tools.", "source": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond", "source_name": "SPE", "id": 127}, {"company": "The Open Group OSDU Forum", "segment": "Industry body", "hq": "United States", "title": "OSDU pushes an AI-ready open data platform", "date": "2025-12", "theme": "Subsurface & exploration", "summary": "At The Open Group Houston Summit 2025, the OSDU Forum focused heavily on AI and analytics and on making the open, vendor-neutral data platform AI-ready by simplifying data consumption. Deliverables included a community implementation that encodes the standard, a data model for seismic interpretation workflows, and first steps toward certification.", "metrics": "", "vendors": "The Open Group, OSDU Forum", "takeaway": "Subsurface AI enablement depends on standardized data foundations such as OSDU; data literacy belongs in the curriculum.", "source": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/", "source_name": "The Open Group", "id": 128}, {"company": "Rice University / Chevron", "segment": "Industry body", "hq": "United States", "title": "Chevron Digital Scholars earn Rice data-science master's", "date": "2021-08", "theme": "Workforce enablement", "summary": "Chevron sponsored employees as Digital Scholars to complete a full-time Master of Computer Science at Rice University, concentrating in data science and machine learning. The first cohort had 21 scholars. Scholars completed D2K Lab capstones on real Chevron problems, such as modeling reservoir productivity decline, and returned as digital change agents.", "metrics": "First cohort: 21 scholars; multiple cohorts completed", "vendors": "Chevron, Rice University (D2K Lab)", "takeaway": "Deep, degree-level programs for a small cadre of domain experts complement broad AI literacy and create in-house translators.", "source": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation", "source_name": "Rice University D2K Lab", "id": 129}, {"company": "US Department of Energy", "segment": "Government", "hq": "United States", "title": "Genesis Mission targets AI for subsurface energy", "date": "2026-03", "theme": "Subsurface & exploration", "summary": "DOE's Genesis Mission, launched by executive order in November 2025, aims to double US science and engineering productivity by bringing AI into all 17 national labs. One of its 26 challenges applies AI to model subsurface environments for cost-effective development of unconventional oil and gas, geothermal and coalbed methane. NETL, DOE's hydrocarbon lab, is participating.", "metrics": "26 challenges; 17 national labs; >$5B announced for Genesis (White House, Jul 2026)", "vendors": "DOE, NETL, national laboratories", "takeaway": "Federal subsurface AI programs are a source of open models, datasets and training partners for US operators.", "source": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and", "source_name": "US Department of Energy", "id": 130}, {"company": "North Sea Transition Authority", "segment": "Government", "hq": "United Kingdom", "title": "NSTA funds AI chatbot for North Sea data", "date": "2025-10", "theme": "Subsurface & exploration", "summary": "The UK regulator received £107,000 from the Regulators' Pioneer Fund for an AI-supported data discovery project. It includes an enhanced geospatial viewer and an AI chatbot to help users navigate NSTA Open Data and the National Data Repository. The regulator holds more than a petabyte of data, and the workstreams were due for assessment in April 2026.", "metrics": "£107K grant; >1 PB of data held; assessment Apr 2026", "vendors": "NSTA, Regulators' Pioneer Fund", "takeaway": "Regulators are building AI access layers over public subsurface data; enablement programs should teach staff to use them.", "source": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/", "source_name": "NSTA", "id": 131}, {"company": "Havtil (Norwegian Ocean Industry Authority)", "segment": "Government", "hq": "Norway", "title": "Norway regulator flags AI black-box and data risks", "date": "2024-03", "theme": "Safety & emissions", "summary": "Havtil said AI use on the Norwegian shelf is early-stage, with testing in condition monitoring, maintenance planning, autonomous cranes, automated drilling and ICT security. It warned about poor training data, models that miss rare events, the black-box problem, and workers relying on screens instead of field observation. Havtil says companies must own the risk, and a DNV follow-up for Havtil (2025) piloted an ontology-grounded AI tool for dropped-object incident data that still needed extensive human review.", "metrics": "", "vendors": "Havtil, DNV, AIICE JIP", "takeaway": "Regulators expect companies to own AI risk, so enablement needs explainability, verification habits and escalation rules.", "source": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/", "source_name": "Havtil", "id": 132}, {"company": "Saudi Aramco", "segment": "NOC", "hq": "Saudi Arabia", "title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "date": "2025-01", "theme": "Workforce enablement", "summary": "Aramco's CEO said technology delivered $4B of value in 2024, double the 2023 figure, with AI embedded across operations. The company has trained hundreds of AI specialists and more than 6,000 AI-enabled professionals, and it has developed 300+ AI use cases. It is also building the METABRAIN industrial LLM on 90 years of company data. Separately, in late 2025 Aramco agreed to take a significant minority stake in HUMAIN, the PIF-backed national AI company.", "metrics": "$4B tech value 2024 (vs $2B 2023); 6,000+ AI-enabled professionals; 300+ use cases; METABRAIN LLM", "vendors": "Aramco, Aramco Digital, HUMAIN, PIF", "takeaway": "National programs link NOC workforce training to sovereign AI infrastructure; expect the same pattern in other Gulf NOCs.", "source": "https://www.middleeastainews.com/p/aramco-ai--drives-4-billion-value", "source_name": "Middle East AI News", "id": 133}, {"company": "Journal of Petroleum Technology (SPE)", "segment": "Cautionary", "hq": "United States", "title": "Why O&G AI pilots fail to scale", "date": "2026-08", "theme": "Strategy & org", "summary": "An ADNOC Offshore reservoir engineer writing in JPT cites the finding that about 95% of GenAI pilots fail to produce measurable business value. He attributes this to curated pilot data that fails in production, organizational misalignment and skill gaps. He argues that a hybrid physics-informed model that engineers trust will scale further than a more elegant model that sits unused.", "metrics": "~95% of GenAI pilots show no measurable value; 54% believe workforce lacks skills; 48% of traditional energy workforce over 45; 57% say OT defences lag IT", "vendors": "SPE, ADNOC Offshore", "takeaway": "Scaling is a governance and change-management problem, so budget for data plumbing and trust-building, not just models.", "source": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas", "source_name": "JPT", "id": 134}, {"company": "Enabled Emissions Campaign / Nature study", "segment": "Cautionary", "hq": "United States", "title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "date": "2026-08", "theme": "Safety & emissions", "summary": "A peer-reviewed Nature study reported by Inside Climate News estimates that AI tools that make extraction faster and cheaper add 0.47-1.8 Gt of CO2 a year. That is 1.2-4.8% of 2024 global energy-related emissions and up to 13 times data-center emissions. The authors call these enabled emissions and want them recognized, measured and governed.", "metrics": "0.47-1.8 GtCO2/yr; 1.2-4.8% of energy emissions; up to 13x data-center emissions; +470B bbl potentially accessible", "vendors": "Enabled Emissions Campaign, Microsoft, Google, Chevron, IEA", "takeaway": "Expect ESG scrutiny of O&G AI programs; framing that covers emissions, safety and methane use cases reduces reputational risk.", "source": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/", "source_name": "Inside Climate News", "id": 135}, {"company": "Microsoft (shareholders and ex-employees)", "segment": "Cautionary", "hq": "United States", "title": "Investors and ex-staff challenge Microsoft's O&G AI work", "date": "2025-12", "theme": "Partnership & investment", "summary": "Former Microsoft employees who quit in 2024 founded Enabled Emissions. At the December 2025 AGM, a shareholder proposal backed by As You Sow challenged Microsoft's sale of AI products to oil and gas producers. Critics note that Microsoft holds about 60% of O&G cloud contracts and that Google stopped building custom upstream AI in 2020.", "metrics": "~60% of O&G cloud contracts (2022 study); XTO target +50K boe/d cited", "vendors": "Microsoft, ExxonMobil (XTO), Chevron, Shell, As You Sow, Google", "takeaway": "Vendor reputational pressure can shift tool availability and terms, so avoid single-vendor lock-in in enablement stacks.", "source": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products", "source_name": "As You Sow", "id": 136}, {"company": "ConocoPhillips", "segment": "Cautionary", "hq": "United States", "title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "date": "2025-09", "theme": "Strategy & org", "summary": "ConocoPhillips said it will cut 20-25% of its roughly 13,000-person workforce under a restructuring called Competitive Edge, advised by BCG, citing rising per-barrel costs. The cuts were not attributed to AI, but they came alongside similar reductions at Chevron and bp. EY's surveys find that employee job-security concerns inhibit AI adoption and experimentation.", "metrics": "20-25% cut (~2,600-3,250 roles); ~13,000 headcount; costs $13/bbl vs $11 in 2021", "vendors": "ConocoPhillips, BCG", "takeaway": "When layoffs coincide with AI rollouts, enablement programs need explicit messaging that AI is for augmentation, or adoption will stall.", "source": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/", "source_name": "Offshore Technology", "id": 137}], "meta": {"title": "AI in Oil & Gas", "start": "2024-09-26", "end": "2026-09-26", "consolidated_on": "2026-09-27", "original_inputs": 69, "reference_inputs": 138, "source_expansion": 4, "profile_count": 306, "company_count": 138, "reviewed_profiles": 263, "pending_profiles": 8, "context_count": 40, "source_count": 592, "disclosure_count": 355, "scope_note": "Selected public-source research, not a global census. Initial screening cohort: 26 September 2024–26 September 2026. Broad classification review: 8 October 2026. Targeted field/source updates: 10 October. Flywheel platform and recruitment review: 12 October. Later publication does not imply every profile was re-audited. Source review is not independent outcome validation. Historical, undated and year-only sources retain their precision. Targeted 14 October review adds CNX safety and an Aker BP implementation announcement; an anonymous failure note is separate from named adoption counts. Earlier source reviews retain their dates. Second October 14 pass: Xinjiang diagnostics, CNX methane, HG invoice processing and Santos source detail. Numerical claims are source-reported; potential benefits remain separate. Third October 14 pass: Aramco core-lab knowledge, clarified deployment stages and existing supplier-reported drilling results. No new named failed deployment verified.", "audited_profiles": 187, "audit_baseline": 187, "round2_added_profiles": 12, "expanded_dossiers": 305, "expanded_profiles": 256, "companies_screened": 177, "new_profiles": 64, "research_on": "2026-10-14", "followup_added_profiles": 3, "followup_enriched_profiles": 1, "eqt_followup_on": "2026-09-29", "gap_review_on": "2026-09-29", "gap_review_added": 10, "gap_review_enriched": 3, "swarm2_added": 9, "swarm2_enriched": 2, "swarm3_added": 6, "swarm3_enriched": 0, "swarm4_added": 9, "swarm4_enriched": 0, "swarm5_added": 8, "swarm5_enriched": 1, "swarm6_added": 3, "swarm6_enriched": 2, "swarm7_added": 4, "swarm7_enriched": 7, "oct4_stage_reviews": 20, "oct4_added_profiles": 0, "oct5_added_profiles": 24, "oct6_added_profiles": 8, "oct5_updated_profiles": 20, "classification_on": "2026-10-08", "classification_note": "Record_type, Evidence_grade, Attribution, Outcome_type, Stage_norm, Scale_note, Segment_norm, Cost_cash_disclosed, Financial_benefits_disclosed, Outcome_counted, Single_source_vendor and related review fields are analyst classifications generated by tools/build_data.py with per-profile overrides.", "initial_screening_window": {"start": "2024-09-26", "end": "2026-09-26"}, "historical_research_on": "2026-10-04", "publication_updated_on": "2026-10-26", "last_register_classification_on": "2026-10-08", "last_substantive_register_review_on": "2026-10-08", "consistency_review_on": "2026-10-10", "supplemental_research_review_on": "2026-10-14", "review_population_counts": {"access_or_verification_limited": 21, "no_expanded_review_recorded": 22, "source_review_recorded": 263}, "review_population_definition": "Mutually exclusive categories based on Research_verification across all current profiles. Source review is not independent validation.", "pending_profiles_definition": "Record_type pending_review; overlaps the source-review categories and must not be added to reviewed_profiles.", "latest_targeted_register_update_on": "2026-10-25", "latest_targeted_supplemental_review_on": "2026-10-26"}, "research": {"packets": [{"key": "chevron-apex", "existing_ids": ["AI-074"], "company": "Chevron", "title": "ApEX multi-agent exploration search", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployed since Aug 2024; update disclosed Nov 2025", "practical_relevance": "A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.", "business_problem": "Exploration screening requires geoscientists to locate and reconcile relevant information across large, heterogeneous subsurface archives.", "workflow": "Explorationists query a chat interface; multiple specialized search agents retrieve relevant materials such as seismic interpretations, logs, maps, reports and prior plans for expert review.", "technology_data": "In-house multi-agent generative AI; described corpus exceeds 1 million files and 4 TB. Specific model, RAG configuration and integration systems not disclosed.", "integration_human_role": "Explorationists formulate queries, inspect the supporting materials and retain prospect/economic judgment.", "deployment_scope": "First used in Gulf of America deepwater exploration; company says it can support decisions around acreage investment. Further assets unreported.", "partners": [], "disclosed_results": [{"metric": "Corpus size", "value": ">1 million files / 4 TB", "period": "2025 disclosure", "attribution": "Company-reported", "source_ids": ["RS001"]}], "economics": "Chevron says ApEX is designed to improve speed and information access; no time saved, exploration success, or return metric disclosed.", "timeline": [{"date": "2024-08", "event": "ApEX launched and first used in Gulf deepwater exploration.", "source_ids": ["RS001"]}, {"date": "2025-11-12", "event": "Chevron disclosed current corpus size and workflow.", "source_ids": ["RS001"]}], "limitations": ["Source is company description; no independent accuracy or decision-impact evidence."], "assessment": "Analyst inference: A concrete named workflow with large corpus scale, although outcome measures and technical validation are absent.", "sources": [{"id": "RS001", "title": "A smarter way to prospect for oil and gas — Chevron ApEX", "url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "publisher": "Chevron", "date": "2025-11-12", "type": "company announcement", "locator": "How does ApEX work? / Advantages in exploration", "review_status": "full text reviewed", "supports": ["ApEX launched in August 2024 and was first used for deepwater exploration in the Gulf of America.", "The in-house multi-agent GenAI system searches more than one million files (4 TB) for explorationists, combining data from surveys, well logs, maps and technical reports.", "Use supports prospect screening and investment decisions by making subsurface materials easier to retrieve; experts remain the users and decision-makers."], "accessed": "2026-09-27", "original_id": "majors-001"}], "field_sources": {"business_problem": ["RS001"], "workflow": ["RS001"], "technology_data": ["RS001"], "integration_human_role": ["RS001"], "deployment_scope": ["RS001"], "economics": ["RS001"]}, "verification": "verified primary", "corrections_to_existing": "AI-074 relied on a secondary article/vendor context; this official 2025 disclosure verifies ApEX and adds technical/workflow specifics.", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-001", "source_ids": ["RS001"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-074"]}, {"key": "bp-bpx-driving-safety", "existing_ids": [], "company": "bp", "title": "Real-time driver hazard alerts in Permian operations", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Implemented across bpx Permian operations in 2024", "practical_relevance": "A frontline safety workflow with a reported operating-period signal: bp says the AI alerts reduced two driver-risk events in bpx Permian operations, while not disclosing fleet denominators or independent validation.", "business_problem": "Driving hazards create collision and injury exposure for onshore field staff.", "workflow": "An AI system assesses driving hazards and risks in real time and provides in-vehicle alerts intended to prompt preventative action.", "technology_data": "bp does not identify model, sensor inputs, vendor, coverage, or alert-performance metrics.", "integration_human_role": "Drivers receive alerts and must respond; precise escalation and supervision procedures are not disclosed.", "deployment_scope": "bpx-operated Permian Basin operations; extent by vehicle or fleet not stated.", "partners": [], "disclosed_results": [{"metric": "Driver distraction events", "value": "89% reduction", "period": "December 2023 to December 2024", "attribution": "Company-reported", "source_ids": ["RS002"]}, {"metric": "Tailgating events", "value": "75% reduction", "period": "December 2023 to December 2024", "attribution": "Company-reported", "source_ids": ["RS002"]}], "economics": "No financial economics disclosed. bp reports reductions in two driving-risk event types; this is not a quantified cost, injury, or collision reduction.", "timeline": [{"date": "2024", "event": "bpx implemented the AI technology across Permian operations.", "source_ids": ["RS002"]}], "limitations": ["Direct report PDF fetch returned 403; substantive passage was reviewed through the official indexed report excerpt.", "bp provides percentage changes without absolute event counts, fleet denominator, or independent validation.", "Do not generalize beyond bpx Permian operations or infer collision/injury reduction."], "assessment": "Analyst inference: in-vehicle alerts can place risk signals inside a driver’s response loop; the company-reported event reductions suggest a measurable operational feedback loop, but denominator and validation limits prevent estimating fleet-wide safety impact.", "sources": [{"id": "RS002", "title": "bp Sustainability Report 2024", "url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "publisher": "bp", "date": null, "type": "annual report", "locator": "Advancing safety through AI, bpx energy", "review_status": "excerpt reviewed", "supports": ["bp states bpx implemented AI technology across Permian operations in 2024 to assess driving hazards and risks in real time and provide in-vehicle alerts.", "The indexed official report passage attributes 89% fewer driver-distraction events and 75% fewer tailgating events between Dec 2023 and Dec 2024 to the technology; direct PDF retrieval returned 403."], "accessed": "2026-09-27", "original_id": "RS002"}], "field_sources": {"business_problem": ["RS002"], "workflow": ["RS002"], "technology_data": ["RS002"], "integration_human_role": ["RS002"], "deployment_scope": ["RS002"], "economics": ["RS002"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-002", "source_ids": ["RS002"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-136"]}, {"key": "exxon-autonomous-drilling", "existing_ids": ["AI-072"], "company": "ExxonMobil", "title": "AI advisory and closed-loop autonomous drilling in Guyana", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "In use in Guyana; details updated on official corporate page within window", "practical_relevance": "Operationally consequential workflow with closed-loop autonomy claimed by company; no independently validated results or control details.", "business_problem": "Offshore drilling teams must select operating parameters while managing safety, consistency and rate of penetration.", "workflow": "Proprietary advisory recommends drilling parameters and can control the drilling process in closed loop; company says automation removes repetitive rig tasks.", "technology_data": "AI parameters system; exact sensors, control architecture and boundaries are undisclosed.", "integration_human_role": "System can operate without human intervention according to ExxonMobil; precise guardrails and override conditions not specified.", "deployment_scope": "Guyana operations; offshore deployment indicated.", "partners": [], "disclosed_results": [], "economics": "Company states benefits but no recent quantified performance result found.", "timeline": [{"date": "2025-09-27", "event": "Company digital technologies page reviewed; described system in current Guyana use.", "source_ids": ["RS003"]}], "limitations": ["The online page lacks a clear publication date; review date is in window, and it is not proof that launch occurred in-window. Existing AI-072 relies on secondary reporting and its exact 2026 claims need primary corroboration."], "assessment": "Analyst inference: Operationally consequential workflow with closed-loop autonomy claimed by company; no independently validated results or control details.", "sources": [{"id": "RS003", "title": "Digital technologies — ExxonMobil", "url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "publisher": "ExxonMobil", "date": null, "type": "company announcement", "locator": "Autonomous drilling", "review_status": "full text reviewed", "supports": ["ExxonMobil describes a proprietary AI drilling advisory system in use in Guyana that determines drilling parameters and enables closed-loop control without human intervention."], "accessed": "2026-09-27", "original_id": "majors-003"}], "field_sources": {"business_problem": ["RS003"], "workflow": ["RS003"], "technology_data": ["RS003"], "integration_human_role": ["RS003"], "deployment_scope": ["RS003"], "economics": ["RS003"]}, "verification": "verified primary", "corrections_to_existing": "Official source supports the general deployed advisory/autonomy workflow, not the existing record’s specific 2026 milestone language.", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-003", "source_ids": ["RS003"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-072"]}, {"key": "total-cognite-industrial-ai", "existing_ids": ["AI-078"], "company": "TotalEnergies", "title": "Industrial data and AI foundation across operated upstream", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Three-year worldwide scale-up announced 2025-09-26", "practical_relevance": "Large, concrete industrial data foundation and deployment scope; business outcome remains an objective.", "business_problem": "Industrial operations need contextualized, accessible sensor and engineering data to monitor assets, improve decisions and deploy AI solutions.", "workflow": "Cognite platform is to unify/contextualize industrial information, enable asset visualization, critical equipment monitoring and faster AI application adoption from drilling through production.", "technology_data": "Cognite industrial data and AI platform; company says industrial data is made AI-ready; detailed schemas/models and use-case list not disclosed.", "integration_human_role": "Enables decisions for field and remote teams; human operational decision authority remains implied, not specified.", "deployment_scope": "Planned across all TotalEnergies-operated upstream assets worldwide over three years; results deck Feb 2026 shows 40 upstream sites, 16 R&C and 500 integrated power sites on larger data platform.", "partners": ["Cognite"], "disclosed_results": [], "economics": "No realized production, availability or savings result quantified; target uplift is stated but not measured.", "timeline": [{"date": "2025-09-26", "event": "Three-year scaling partnership announced.", "source_ids": ["RS004"]}, {"date": "2026-02", "event": "Investor deck reports site footprint and data contextualization scale.", "source_ids": ["RS005"]}], "limitations": ["Cognite partner included as co-publisher of announcement; deployment scale is planned."], "assessment": "Analyst inference: Large, concrete industrial data foundation and deployment scope; business outcome remains an objective.", "sources": [{"id": "RS004", "title": "Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI", "url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "publisher": "TotalEnergies / Cognite", "date": "2025-09-26", "type": "company announcement", "locator": "PDF pp. 1–2", "review_status": "full text reviewed", "supports": ["Three-year rollout of Cognite industrial data and AI platform across all TotalEnergies-operated upstream assets worldwide, drilling through production.", "The initiative contextualizes industrial data, supports dynamic asset visualization and monitoring of critical equipment, and aims to accelerate applications."], "accessed": "2026-09-27", "original_id": "majors-005"}, {"id": "RS005", "title": "2025 Results and 2026 Objectives Presentation", "url": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf", "publisher": "TotalEnergies", "date": "2026-02", "type": "investor presentation", "locator": "PDF slides 28–29", "review_status": "full text reviewed", "supports": ["Reported data platform footprint: 40 upstream, 16 refining and chemicals, and 500 integrated power sites; platform captures 10x real-time data points and contextualizes 35 million points.", "The deck names three AI strategic programs and a new global competency center in India, with target production and availability uplift but no quantified realized uplift."], "accessed": "2026-09-27", "original_id": "majors-006"}], "field_sources": {"business_problem": ["RS004"], "workflow": ["RS004", "RS005"], "technology_data": ["RS004", "RS005"], "integration_human_role": ["RS004"], "deployment_scope": ["RS004", "RS005"], "economics": ["RS005"]}, "verification": "verified primary", "corrections_to_existing": "AI-078 announcement dated exactly at window end; 2026 deck corroborates subsequent footprint.", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-004", "source_ids": ["RS004", "RS005"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-078"]}, {"key": "total-mistral-lab", "existing_ids": ["AI-029"], "company": "TotalEnergies", "title": "Joint AI innovation lab for researcher and industrial decision support", "department": "Research & innovation", "secondary_departments": [], "category": "Internal adoption", "stage": "Joint AI innovation lab announced 2025-06-12; named uses planned for design/testing", "practical_relevance": "The announced lab will design and test AI workflows for researchers, industrial asset decision support, customers and renewable production; production performance and financial returns are not yet disclosed.", "business_problem": "Researchers, operations teams and customers need AI support tailored to energy and geoscience knowledge.", "workflow": "Initial use cases: assistant for 1,000 researchers; industrial decision-support for asset performance/CO2; customer energy-saving support.", "technology_data": "Mistral generative AI; detailed architecture, model validation and data governance not disclosed.", "integration_human_role": "AI outputs are intended to support employees and customers; expert review details are not given.", "deployment_scope": "Innovation lab; enterprise scope and proof-of-concept status unspecified.", "partners": ["Mistral AI"], "disclosed_results": [], "economics": "Not disclosed for the Mistral lab; the 2026 €100m frontier-model investment is a distinct AI-077 initiative and is not attributed to this lab.", "timeline": [{"date": "2025-06-12", "event": "Joint innovation lab announced with researcher, industrial and customer use cases.", "source_ids": ["S18"]}], "limitations": [], "assessment": "Analyst inference: Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.", "sources": [{"id": "S18", "title": "TotalEnergies to Collaborate with Mistral AI to Increase the Application of Artificial Intelligence", "url": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "publisher": "TotalEnergies", "date": "2025-06-12", "type": "company announcement", "locator": "A shared AI innovation lab", "review_status": "full text reviewed", "supports": ["Joint lab use cases include an assistant for 1,000 researchers, decision-support for industrial performance and CO2 emissions, and customer energy-saving support.", "TotalEnergies reported Digital Factory had developed over 100 solutions, 60 using ML through generative AI; these are aggregate portfolio counts, not outcome metrics."], "accessed": "2026-09-27", "original_id": "majors-007"}], "field_sources": {"business_problem": ["S18"], "workflow": ["S18"], "technology_data": ["S18"], "integration_human_role": ["S18"], "deployment_scope": ["S18"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "Separate 2025 lab use cases from 2026 frontier-model investment; do not collapse into one deployment.", "related_existing_ids": ["AI-030"], "cohort": "majors", "packet_id": "majors-005", "source_ids": ["S18"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-029"]}, {"key": "woodside-maint-intel", "existing_ids": ["AI-042"], "company": "Woodside", "title": "Woodside’s maintenance estimate has a traceable source", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Operating at North West Shelf; potential maintenance-hours reduction", "practical_relevance": "Were the recommended intervals adopted, and did reliability hold while maintenance hours fell?", "business_problem": "Maintenance intervals must preserve reliability while avoiding unnecessary preventive work and model-analysis delays.", "workflow": "Analyzes maintenance records, identifies equipment failure modes and feeds statistical models that recommend optimal intervals against reliability targets.", "technology_data": "Maint Intel compares maintenance history with reliability targets to recommend intervals. A separate January article identifies AWS support and the North West Shelf launch.", "integration_human_role": "Maintenance teams use recommended intervals; automatic execution is not established.", "deployment_scope": "Tested at Angel and launched at North West Shelf. January article: 5 January 2026. AI page last updated 27 May 2026; original publication date and estimate period unknown.", "partners": ["AWS"], "disclosed_results": [{"metric": "Model processing time", "value": "5 days to under 2 hours", "period": "Maint Intel project", "attribution": "Company-reported", "source_ids": ["RS006"]}], "economics": "Company-reported model processing time: five days to under two hours. Separately, the AI page identifies potential 15% lower annual maintenance hours at Angel over five years; not an achieved reduction.", "timeline": [{"date": "2026-01-05", "event": "Maint Intel launched at North West Shelf following Angel testing.", "source_ids": ["RS006"]}, {"date": "Updated 2026-05-27; publication day unknown", "review_on": "2026-10-18", "event": "Woodside’s AI page supports the previously unresolved 15% maintenance-hours estimate at Angel. The AI page describes potential 15% lower annual maintenance hours over five years, not hours already avoided. The January article reports model processing cut from five days to under two hours; that is not a field-labor saving.", "source_ids": ["S24", "RS006"]}], "limitations": ["The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified."], "assessment": "The maintenance-hours estimate now has a traceable source. Whether it becomes a saving depends on adoption, reliability and costs.", "sources": [{"id": "RS006", "title": "Right fix at right time – AI delivers for Woodside", "url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "date": "2026-01-05", "publisher": "Woodside", "locator": "Angel pilot / North West Shelf launch, AWS collaboration and model processing-time paragraphs", "excerpt": "five days to under two hours", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "review_status": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}, {"id": "S24", "title": "Artificial Intelligence at Woodside", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "date": "Updated 2026-05-27; publication day unknown", "publisher": "Woodside", "locator": "Maintenance section; last-updated stamp at foot of page", "excerpt": "a potential 15% reduction of maintenance hours", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "review_status": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["RS006", "S24"], "workflow": ["RS006", "S24"], "technology_data": ["RS006", "S24"], "deployment_scope": ["RS006", "S24"], "economics": ["RS006", "S24"], "integration_human_role": ["RS006", "S24"], "practical_relevance": ["RS006", "S24"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-006", "source_ids": ["RS006", "S24"], "report_date": "2026-10-18", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-042"], "earlier_review": {"report_date": "2026-09-27", "assessment": "Analyst inference: Named workflow, partner, pilot and reported computational improvement; potential maintenance-hour savings are prospective and not independently verified.", "limitations": ["The reopened source does not substantiate the 15% annual maintenance-hours estimate present in the prior packet; that figure should be omitted pending a traceable source.", "Further global rollout was scheduled during 2026; completion was not confirmed in this source."], "note": "Superseded by targeted 18 October 2026 source check."}, "last_source_check": "2026-10-18"}, {"key": "repsol-multiagent", "existing_ids": ["AI-031"], "company": "Repsol", "title": "Multi-agent workflow platform for employee productivity", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Pilot deployed from 2025 for over four months; 2026 annual report snapshot updated to 34 agents / 3 workflows", "practical_relevance": "Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.", "business_problem": "Employees need to coordinate multi-step information work and apply internal knowledge across business processes.", "workflow": "Platform configures agents into workflows; pilot teams supervised agents on departmental processes; separate upstream Automated Well Data Ingestion automates daily drilling/geology reports.", "technology_data": "Repsol-built agent platform with Accenture; custom knowledge bases, professional interface and language models. Annual report lists 34 agents/3 workflows, while Oct 2025 snapshot reported 22 agents/3 teams.", "integration_human_role": "Over 50 people supervised pilot teams; employee review role is explicit.", "deployment_scope": "Pilot in Digitalization and Services department (>700 employees) plus reported downstream workflows and an upstream automated-report workflow.", "partners": ["Accenture"], "disclosed_results": [{"metric": "Agents in multi-agent workflows", "value": "34 across 3 workflows", "period": "2025 annual report", "attribution": "Company-reported", "source_ids": ["RS007"]}], "economics": "No quantified efficiency metric disclosed; Repsol claims numerous efficiencies and improvements.", "timeline": [{"date": "2025-10-15", "event": "Multi-agent pilot announced; 22 agents across three teams.", "source_ids": ["S19"]}, {"date": "2026-02", "event": "Annual report reports 34 agents in three workflows and Automated Well Data Ingestion.", "source_ids": ["RS007"]}], "limitations": ["Agent counts vary by disclosure date; not a contradiction. Annual report window within research range."], "assessment": "Analyst inference: Concrete supervised agent workflows with measured cohort/scale; operational productivity outcomes remain unquantified.", "sources": [{"id": "S19", "title": "Repsol launches intelligent multi-agent platform", "url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "publisher": "Repsol", "date": "2025-10-15", "type": "company announcement", "locator": "Pilot project / GenAI competence center", "review_status": "full text reviewed", "supports": ["Accenture collaboration produced an agent-construction platform and pilot in Digitalization and Services department.", "Three teams with 22 agents supported processes for over four months; over 50 people supervised the agent teams.", "Repsol reports 60 production GenAI use cases and 22 scaled as of Oct 2025; these are company-reported portfolio measures."], "accessed": "2026-09-27", "original_id": "majors-010"}, {"id": "RS007", "title": "Consolidated Management Report 2025 — Repsol", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "publisher": "Repsol", "date": "2026-02-19", "type": "annual report", "locator": "5.5 Technology and digitalization", "review_status": "excerpt reviewed", "supports": ["Repsol reports Automated Well Data Ingestion automates daily drilling and geology reports and integrates data into corporate systems to speed exploration decisions.", "Annual report later lists 34 agents in three workflows, distinct from the Oct 2025 pilot snapshot of 22 agents."], "accessed": "2026-09-27", "original_id": "majors-011"}], "field_sources": {"business_problem": ["S19"], "workflow": ["S19", "RS007"], "technology_data": ["S19", "RS007"], "integration_human_role": ["S19"], "deployment_scope": ["S19", "RS007"], "economics": ["S19"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": ["AI-032"], "cohort": "majors", "packet_id": "majors-007", "source_ids": ["S19", "RS007"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-031"]}, {"key": "omv-ai-hub", "existing_ids": ["AI-084"], "company": "OMV", "title": "AI Hub and enterprise assistants", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "AI Hub launched 2025; 2025 annual report reviewed 2026-04-02", "practical_relevance": "Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.", "business_problem": "Employees need secure tools for document analysis, information retrieval and navigation of internal procedures.", "workflow": "Central Hub provides chatbots, assistants and agents; GPT assistant retrieves/analyzes documents; Regulations Assistant guides users through internal policies.", "technology_data": "Built on Azure and integrated in OMV digital environment; details of model, retrieval architecture and guardrails not disclosed.", "integration_human_role": "Employees use outputs for daily work; policies assistant surfaces rules; review remains with staff.", "deployment_scope": "Group-wide hub; in-house GPT used monthly by >2,500 employees; 10,000+ attended GenAI training.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Monthly active users", "value": ">2,500 employees", "period": "2025", "attribution": "Company-reported", "source_ids": ["R037"]}, {"metric": "GenAI training reach", "value": ">10,000 employees", "period": "2025", "attribution": "Company-reported", "source_ids": ["R037"]}], "economics": "No directly attributable productivity or financial metric; adoption and training counts reported.", "timeline": [{"date": "2025", "event": "AI Hub launched; 2025 usage metrics reported in annual report.", "source_ids": ["R037"]}], "limitations": ["Employee-use count and project pipeline are company-reported."], "assessment": "Analyst inference: Good adoption evidence with identifiable workflows but business outcomes and model architecture not disclosed.", "sources": [{"id": "R037", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": null, "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27", "original_id": "majors-012"}], "field_sources": {"business_problem": ["R037"], "workflow": ["R037"], "technology_data": ["R037"], "integration_human_role": ["R037"], "deployment_scope": ["R037"], "economics": ["R037"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-008", "source_ids": ["R037"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-084"]}, {"key": "omv-norwegian-companion", "existing_ids": ["AI-132"], "company": "OMV", "title": "Norwegian AI Companion geological and engineering search", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported in OMV 2025 annual report", "practical_relevance": "Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.", "business_problem": "Engineers face large archives of geology and petroleum-engineering material on the Norwegian Continental Shelf.", "workflow": "Companion synthesizes answers to complex geology/geoscience/petroleum engineering questions from a large curated source corpus.", "technology_data": "Searches 2.5m pages of public and internal geology, geophysics and reports; precise model and retrieval methods not specified.", "integration_human_role": "Engineering teams consume concise answers and focus on technical work; validation workflow not described.", "deployment_scope": "Relevant to Norwegian Continental Shelf teams; user count and rollout stage not disclosed.", "partners": [], "disclosed_results": [{"metric": "Corpus", "value": ">2.5 million pages", "period": "2025", "attribution": "Company-reported", "source_ids": ["R037"]}], "economics": "No output quality or time-saved result stated.", "timeline": [{"date": "2026-04-02", "event": "2025 annual report describes NAIC.", "source_ids": ["R037"]}], "limitations": ["OMV describes a digital assistant, but details on what counts as actionable response and hallucination controls are absent."], "assessment": "Analyst inference: Useful domain search workflow with corpus scale and business users identified; efficacy evidence is absent.", "sources": [{"id": "R037", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": null, "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27", "original_id": "majors-012"}], "field_sources": {"business_problem": ["R037"], "workflow": ["R037"], "technology_data": ["R037"], "integration_human_role": ["R037"], "deployment_scope": ["R037"], "economics": ["R037"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-009", "source_ids": ["R037"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-132"]}, {"key": "omv-iDEA", "existing_ids": ["AI-134"], "company": "OMV", "title": "iDEA well-engineering knowledge system", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Introduced by 2025 annual-report period", "practical_relevance": "A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.", "business_problem": "Drilling design can repeat avoidable problems when prior global well experience is hard to find.", "workflow": "Knowledge system identifies global analogues and lessons learned during active drilling-program design to flag potential issues and costs.", "technology_data": "Connects current well design to OMV and OMV Petrom global experience; underlying AI/model details not disclosed.", "integration_human_role": "Drilling engineers use timely insights to inform program; final design accountability remains with engineers.", "deployment_scope": "OMV and OMV Petrom well-design workflow; exact adoption and well count not reported.", "partners": [], "disclosed_results": [], "economics": "Avoided incidents/cost savings not quantified.", "timeline": [{"date": "2026-04-02", "event": "2025 annual report describes introduction of iDEA.", "source_ids": ["R037"]}], "limitations": ["System is described as knowledge-based; avoid asserting generative AI or autonomous design."], "assessment": "Analyst inference: A grounded knowledge workflow with obvious operational application, but no quantified validation or architecture.", "sources": [{"id": "R037", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": null, "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27", "original_id": "majors-012"}], "field_sources": {"business_problem": ["R037"], "workflow": ["R037"], "technology_data": ["R037"], "integration_human_role": ["R037"], "deployment_scope": ["R037"], "economics": ["R037"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-010", "source_ids": ["R037"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-134"]}, {"key": "omv-pump-ml", "existing_ids": ["AI-135"], "company": "OMV", "title": "Predictive pump-failure analysis in Romania", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Machine-learning pilot reported in 2025 annual report", "practical_relevance": "Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.", "business_problem": "Critical well failures and process interventions can interrupt production; proactive planning can reduce downtime.", "workflow": "Model analyzes historical pump-operation and event data to predict potential pump failures up to 90 days ahead.", "technology_data": "Machine-learning failure prediction; specific model, test performance and deployment integration not disclosed.", "integration_human_role": "Supports proactive maintenance planning; human maintenance intervention implied.", "deployment_scope": "Romania pilot; critical well failures; no number of wells/pumps given.", "partners": [], "disclosed_results": [{"metric": "Forecast horizon", "value": "up to 90 days", "period": "Romania pilot", "attribution": "Company-reported", "source_ids": ["R037"]}], "economics": "Downtime avoided or prediction precision not reported.", "timeline": [{"date": "2026-04-02", "event": "OMV reports Romanian ML pilot and 90-day forecast horizon.", "source_ids": ["R037"]}], "limitations": ["Pilot only; predictive horizon is not accuracy."], "assessment": "Analyst inference: Named predictive-maintenance pilot with forecast horizon; lack of validation metrics prevents assessment of decision quality.", "sources": [{"id": "R037", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": null, "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27", "original_id": "majors-012"}], "field_sources": {"business_problem": ["R037"], "workflow": ["R037"], "technology_data": ["R037"], "integration_human_role": ["R037"], "deployment_scope": ["R037"], "economics": ["R037"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-011", "source_ids": ["R037"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-135"]}, {"key": "omv-field-agent", "existing_ids": ["AI-133"], "company": "OMV", "title": "OMV tests field-development decisions with analytical agents", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "CCS planning pilot; early internal results", "practical_relevance": "Do better plans still outperform after engineering review and actual field execution?", "business_problem": "Low-carbon field development needs better decisions under subsurface uncertainty and safety constraints.", "workflow": "Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "technology_data": "Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "integration_human_role": "Engineers challenge assumptions and validate recommendations before action.", "deployment_scope": "Norwegian CCS planning pilot; article published 14 April 2026. Test dates, asset name and project count unknown.", "partners": ["Stanford University", "TerraAI"], "disclosed_results": [], "economics": "OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "timeline": [{"date": "2026-04-02", "event": "OMV describes ongoing test partnership in annual report.", "source_ids": ["R037"]}, {"date": "2026-04-14", "review_on": "2026-10-18", "event": "OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI. OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "source_ids": ["WATCH18PM-OMV"]}], "limitations": ["No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132)."], "assessment": "Test whether the selected plan works after execution. Faster scenario analysis and early project-value claims are not collected cash.", "sources": [{"id": "R037", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": null, "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27", "original_id": "majors-012"}, {"id": "WATCH18PM-OMV", "title": "How AI is powering better decisions in the race to net zero", "url": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "date": "2026-04-14", "publisher": "OMV", "locator": "Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker", "excerpt": "not generative AI, but analytical, agent-based AI", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "review_status": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["R037", "WATCH18PM-OMV"], "workflow": ["R037", "WATCH18PM-OMV"], "technology_data": ["R037", "WATCH18PM-OMV"], "deployment_scope": ["R037", "WATCH18PM-OMV"], "economics": ["R037", "WATCH18PM-OMV"], "integration_human_role": ["R037", "WATCH18PM-OMV"], "practical_relevance": ["R037", "WATCH18PM-OMV"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-012", "source_ids": ["R037", "WATCH18PM-OMV"], "report_date": "2026-10-18", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-133"], "earlier_review": {"report_date": "2026-09-27", "assessment": "Analyst inference: Early-stage named technical collaboration with preliminary outcomes; claims remain company-reported and unquantified.", "limitations": ["Early results are qualitative; do not treat as deployed or proven value."], "note": "Superseded by targeted 18 October 2026 source check."}, "last_source_check": "2026-10-18"}, {"key": "eni-open-hpc-ai", "existing_ids": [], "company": "Eni", "title": "Open access to supercomputing for external AI innovation", "department": "Research & innovation", "secondary_departments": [], "category": "AI infrastructure & enablement", "stage": "HPC6/HPC7 access program announced July 2026", "practical_relevance": "Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.", "business_problem": "Startups, companies and research organizations may lack compute and specialized expertise needed to develop AI solutions.", "workflow": "Eni makes HPC capacity, expertise and tools available to external firms, startups and research centers developing AI and digital innovation.", "technology_data": "Green Data Center hosts HPC6/HPC7; combined performance described by Eni as above exascale-class. The infrastructure supports broad HPC/AI experimentation, not a specific oilfield AI workflow.", "integration_human_role": "External researchers/companies access Eni tools and expertise; project selection/terms are not in reviewed announcement.", "deployment_scope": "Innovation ecosystem access; specific program scale, users and compute allocation not disclosed.", "partners": [], "disclosed_results": [{"metric": "Combined compute", "value": "Exascale-class performance", "period": "July 2026", "attribution": "Company-reported", "source_ids": ["RS008"]}], "economics": "No commercial returns or AI project outcomes quantified.", "timeline": [{"date": "2026-07-28", "event": "Eni announces access program for HPC infrastructure.", "source_ids": ["RS008"]}], "limitations": ["Kept separate from data-center supply/infrastructure activity; disclosure concerns research compute made available to third parties."], "assessment": "Analyst inference: Strategic enablement infrastructure for AI research, not evidence of operational AI adoption.", "sources": [{"id": "RS008", "title": "Eni opens its supercomputing capabilities to the innovation ecosystem", "url": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "publisher": "Eni", "date": "2026-07-28", "type": "company announcement", "locator": "Research and Development", "review_status": "full text reviewed", "supports": ["Eni opened HPC infrastructure, expertise and tools to companies, startups and research centres to develop AI and digital innovation; HPC6 and HPC7 together exceed exascale-class performance."], "accessed": "2026-09-27", "original_id": "majors-013"}], "field_sources": {"business_problem": ["RS008"], "workflow": ["RS008"], "technology_data": ["RS008"], "integration_human_role": ["RS008"], "deployment_scope": ["RS008"], "economics": ["RS008"]}, "verification": "verified primary", "corrections_to_existing": "AI-027 is an aggregate portfolio record; this is a named 2026 infrastructure initiative and not evidence of a particular oil/gas workflow.", "related_existing_ids": ["AI-027"], "cohort": "majors", "packet_id": "majors-013", "source_ids": ["RS008"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-137"]}, {"key": "eni-ai-portfolio", "existing_ids": ["AI-027"], "company": "Eni", "title": "AI agents, digital plant and drilling automation portfolio", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "2025 outcomes and targets reported in Eni for 2025", "practical_relevance": "Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.", "business_problem": "Industrial plants need improved uptime/energy efficiency and drilling must reduce cycle time while maintaining safety.", "workflow": "Eni reports AI applications for digital plant optimization and drilling automation; AI agents are developing across portfolio.", "technology_data": "About 300 AI use cases, about 20 AI agents; specific models/data and production coverage absent.", "integration_human_role": "Human role, automation scope and workflow approval details not described in summary disclosure.", "deployment_scope": "Company-wide, with digital plant and drilling applications; specific sites not named.", "partners": [], "disclosed_results": [{"metric": "Use cases", "value": "~300 (+40% vs. 2024)", "period": "2025", "attribution": "Company-reported", "source_ids": ["RS009"]}, {"metric": "Digital plant target", "value": "2–3% production uplift and emissions reduction", "period": "target", "attribution": "Company-reported", "source_ids": ["RS009"]}, {"metric": "Drilling automation target", "value": "35% less drilling time", "period": "target", "attribution": "Company-reported", "source_ids": ["RS009"]}], "economics": "Targets: digital plant 2–3% production enhancement and 2–3% emissions reduction; drilling automation 35% reduction in drilling time. These are targets, not verified realized results.", "timeline": [{"date": "2026-04-22", "event": "Eni for 2025 reports use cases, agents and expected impacts.", "source_ids": ["RS009"]}], "limitations": ["Targets can be misread as actual results; metrics are company-reported targets."], "assessment": "Analyst inference: Portfolio-scale evidence and measurable targets exist, but the short disclosure does not map them to specific deployments or realized outcomes.", "sources": [{"id": "RS009", "title": "Eni for 2025 — Just Transition", "url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "publisher": "Eni", "date": "2026-04", "type": "annual report", "locator": "Data and artificial intelligence", "review_status": "excerpt reviewed", "supports": ["Eni reports around 300 AI use cases (+40% vs. 2024), around 20 AI agents; digital plant target ranges of 2–3% production enhancement and emissions reduction and drilling automation target of 35% less drilling time."], "accessed": "2026-09-27", "original_id": "majors-014"}], "field_sources": {"business_problem": ["RS009"], "workflow": ["RS009"], "technology_data": ["RS009"], "integration_human_role": ["RS009"], "deployment_scope": ["RS009"], "economics": ["RS009"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-014", "source_ids": ["RS009"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-027"]}, {"key": "aker-yggdrasil-agents", "existing_ids": [], "company": "Aker BP", "title": "AI agents for Yggdrasil remote operations", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Planned from day one of Yggdrasil operations in 2027; disclosed 2026", "practical_relevance": "A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.", "business_problem": "Remote/low-manned operations need rapid condition monitoring, maintenance planning and troubleshooting with safe operations.", "workflow": "AI agents planned for condition-based maintenance and planning at launch, with scaling to more operational areas as solutions mature; data to be shared with onshore control room and external experts.", "technology_data": "Cognite Data Fusion structures thousands of sensors and plant data; detailed agent design and safeguards not disclosed.", "integration_human_role": "Agents work alongside control-room technicians; people decide on operations and gradual scope increase is planned.", "deployment_scope": "Yggdrasil remote onshore operations centre planned from 2027.", "partners": ["Cognite"], "disclosed_results": [], "economics": "No measured uptime or cost benefit disclosed.", "timeline": [{"date": "2026-04-29", "event": "Aker BP described day-one AI agent plans for Yggdrasil.", "source_ids": ["RS010"]}, {"date": "2026-02-11", "event": "CEO said AI agents reduce operational troubleshooting time and improve uptime; company-wide capability claim.", "source_ids": ["RS011"]}], "limitations": ["Company claims operational agents already reduce troubleshooting time, but no example or quantified result was found in reviewed material."], "assessment": "Analyst inference: A strong future deployment design with human-in-the-loop and structured data foundation; it remains planned for 2027 and no outcome yet.", "sources": [{"id": "RS010", "title": "Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation", "url": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "publisher": "Aker BP", "date": null, "type": "company announcement", "locator": "Yggdrasil remote operations", "review_status": "full text reviewed", "supports": ["Yggdrasil operations from onshore centre planned from 2027; AI agents planned from day one for condition-based maintenance and planning, with scaling to other workflows as mature.", "Aker BP says industrial data platform Cognite Data Fusion structures thousands of sensor streams and data is to be shared with people and AI agents."], "accessed": "2026-09-27", "original_id": "majors-016"}, {"id": "RS011", "title": "Aker BP Q4 2025 transcript", "url": "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf", "publisher": "Aker BP", "date": "2026-02-11", "type": "earnings call", "locator": "PDF pp. 5–6", "review_status": "full text reviewed", "supports": ["CEO describes AI-enabled earlier exploration decisions and AI agents reducing operational troubleshooting time and improving uptime.", "The CEO describes high-quality structured real-time data connection across the company; 10 years of digital investment is contextual foundation.", "Microsoft speaker (partner attribution) says Copilot adoption nearly 100%; not independently audited."], "accessed": "2026-09-27", "original_id": "majors-015"}], "field_sources": {"business_problem": ["RS010"], "workflow": ["RS011", "RS010"], "technology_data": ["RS010"], "integration_human_role": ["RS010"], "deployment_scope": ["RS010"], "economics": ["RS011"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": ["AI-131"], "cohort": "majors", "packet_id": "majors-015", "source_ids": ["RS010", "RS011"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-138"]}, {"key": "santos-ai-program", "existing_ids": [], "company": "Santos", "title": "Production AI, virtual pressure measurement and drilling controls", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Company reports production-AI benefits in 2025; pressure-model savings are potential; deployment population undisclosed.", "practical_relevance": "Virtual pressure measurement could avoid spending on physical instrumentation. That is a different value mechanism from production optimization. The potential benefit still depends on where engineers accept the model as a replacement and what it costs to keep it reliable.", "business_problem": "Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.", "workflow": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "technology_data": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "integration_human_role": "Drilling controls are monitored from an operating center; pressure-model acceptance and oversight are not described.", "deployment_scope": "26 May 2026 investor briefing; the deck reports 2025 production-AI benefits and 2029 potential. The stimulation column compares 2026 productivity with a 2023 baseline. No well/rig population or model-specific measurement period is supplied.", "partners": [], "disclosed_results": [{"metric": "Column-specific achieved benefits and potential", "value": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "period": "26 May 2026 investor briefing; the deck reports 2025 production-AI benefits and 2029 potential. The stimulation column compares 2026 productivity with a 2023 baseline. No well/rig population or model-specific measurement period is supplied.", "attribution": "Santos; not independently validated; excluded from numerical counts pending scope reconciliation", "source_ids": ["RS012"]}], "economics": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "timeline": [{"date": "2026-05-26", "event": "Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.", "source_ids": ["RS012"]}], "limitations": ["The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology."], "assessment": "Virtual pressure measurement could avoid spending on physical instrumentation. That is a different value mechanism from production optimization. The potential benefit still depends on where engineers accept the model as a replacement and what it costs to keep it reliable.", "sources": [{"id": "RS012", "title": "Santos 2026 Investor Briefing Day", "url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "publisher": "Santos", "date": "2026-05-26", "type": "investor presentation", "locator": "Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross", "review_status": "Original 84-page issuer PDF retrieved and slide 72 rendered/reviewed 14 October 2026. Four column-specific metrics kept separate.", "supports": ["Santos says technology program applies digital and AI capability across integrated value chain, including cost reduction and production uplift across drilling, completions, projects and facilities; specific quantified AI-only impacts not isolated."], "accessed": "2026-10-14", "original_id": "majors-017"}], "field_sources": {"business_problem": ["RS012"], "workflow": ["RS012"], "technology_data": ["RS012"], "integration_human_role": ["RS012"], "deployment_scope": ["RS012"], "economics": ["RS012"]}, "verification": "verified primary", "corrections_to_existing": "AI-083 is secondary report on Xecta and should not be certified from this briefing excerpt; Xecta source verification remains outstanding.", "related_existing_ids": ["AI-083"], "cohort": "majors", "packet_id": "majors-016", "source_ids": ["RS012"], "report_date": "2026-10-14", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-139"]}, {"key": "total-mistral-frontier-models", "existing_ids": ["AI-077"], "company": "TotalEnergies", "title": "Frontier AI models for reservoir exploration and engineering", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Three-year program announced 2026-09-15", "practical_relevance": "Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.", "business_problem": "Subsurface experts need to interpret extensive seismic, geological and reservoir data and consider multiple development scenarios for prospects and existing projects.", "workflow": "Joint scientific lab will develop models to support exploration, reservoir characterization and development, generate multiple scenarios for exploration opportunities, and help optimize and extend existing projects.", "technology_data": "Agentic AI is specified; TotalEnergies says program will combine nearly 10 petaflops of data with accumulated subsurface expertise. Detailed training, inference architecture and benchmarks not disclosed.", "integration_human_role": "Models are intended to support subsurface experts in analysis and decision-making; expert decision authority is explicit.", "deployment_scope": "Three-year joint program; development and scope across subsurface activities, not yet deployed operationally.", "partners": ["Mistral AI"], "disclosed_results": [], "economics": "Investment >€100 million over three years; no return or production impact disclosed.", "timeline": [{"date": "2026-09-15", "event": "Three-year frontier-model partnership announced.", "source_ids": ["R021"]}], "limitations": ["Development program is newly announced; no models, deployments, validation results or returns disclosed."], "assessment": "Analyst inference: Named business uses and material investment show commitment, but performance results and in-production scope remain undisclosed.", "sources": [{"id": "R021", "title": "TotalEnergies Announces a Partnership with Mistral to Develop Frontier AI Models for Reservoir Exploration and Engineering", "url": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "publisher": "TotalEnergies", "date": "2026-09-15", "accessed": "2026-09-27", "type": "company announcement", "locator": "Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data", "review_status": "full text reviewed", "supports": ["Three-year joint program represents over €100 million investment to develop frontier AI models supporting exploration, characterization and development of oil and gas reservoirs.", "Program will use agentic AI and combine almost 10 petaflops of data with geoscience/reservoir expertise; develop multiple scenarios for exploration opportunities and optimizing existing projects.", "Scientific lab will combine TotalEnergies subsurface experts with Mistral AI science and technology."], "original_id": "majors-020"}], "field_sources": {"business_problem": ["R021"], "workflow": ["R021"], "technology_data": ["R021"], "integration_human_role": ["R021"], "deployment_scope": ["R021"], "economics": ["R021"]}, "verification": "verified primary", "corrections_to_existing": "AI-077 is a distinct frontier-model program and should remain separate from the 2025 Mistral AI innovation-lab use cases.", "related_existing_ids": ["AI-030"], "cohort": "majors", "packet_id": "majors-017", "source_ids": ["R021"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-077"]}, {"key": "shell-agentic-reliability", "existing_ids": ["AI-076"], "company": "Shell", "title": "Agentic root-cause analysis in global reliability program", "department": "Maintenance & reliability", "secondary_departments": ["Production & field operations"], "category": "Internal adoption", "stage": "New multi-year agreement announced 2026-06-04; implementation planned", "practical_relevance": "Extends predictive maintenance beyond anomaly detection toward analysis and remediation of reliability issues across asset operations.", "business_problem": "Shell needs to identify equipment anomalies and determine causes/remediation across a large distributed asset base.", "workflow": "C3 AI reliability platform monitors equipment for predictive maintenance; new agreement will add agent-based root-cause analysis and remediation.", "technology_data": "C3 AI Reliability and Agentic AI Platform on Microsoft Azure; partner says historical deployment monitors >13,000 equipment items.", "integration_human_role": "Decision rights and human review in agent RCA are not disclosed.", "deployment_scope": "Global operations; multi-year expansion. Existing predictive maintenance ran since 2018; new agentic scope announced for future rollout.", "partners": ["C3 AI", "Microsoft Azure"], "disclosed_results": [], "economics": "C3 AI executive claims hundreds of millions of dollars in economic value, but basis and Shell validation are not supplied.", "timeline": [{"date": "2026-06-04", "event": "New multi-year expansion agreement announced; agentic RCA planned.", "source_ids": ["RS013"]}], "limitations": ["Expansion announcement is prospective; no results for the added agent functionality.", "13,000 equipment coverage and hundreds of millions value are vendor-reported, not independent."], "assessment": "Analyst inference: this is a substantial global reliability program with clearly named predictive-maintenance footprint; agentic remediation scope is only newly contracted.", "sources": [{"id": "RS013", "title": "C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations", "url": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "publisher": "C3 AI", "date": "2026-06-04", "accessed": "2026-09-27", "type": "partner case study", "locator": "News Details, lines 285–297", "review_status": "full text reviewed", "supports": ["C3 AI says it has worked with Shell since 2018 on predictive maintenance monitoring over 13,000 equipment items.", "A new multi-year agreement announced June 2026 extends reliability AI and introduces agent-based root cause analysis and remediation; the agreement is prospective.", "Program uses C3 AI Reliability and C3 Agentic AI Platform on Microsoft Azure.", "Hundreds of millions of dollars value is a C3 AI executive claim, not independent or Shell-attributed validation."], "original_id": "majors-018"}], "field_sources": {"business_problem": ["RS013"], "workflow": ["RS013"], "technology_data": ["RS013"], "integration_human_role": ["RS013"], "deployment_scope": ["RS013"], "economics": ["RS013"]}, "verification": "verified partner", "corrections_to_existing": "Official Shell/C3 partner disclosure corroborates AI-076 with vendor evidence and reveals prior predictive maintenance footprint; classify the latest agentic scope as prospective.", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-018", "source_ids": ["RS013"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-076"]}, {"key": "shell-pearl-machine-vision", "existing_ids": [], "company": "Shell", "title": "Real-time machine vision and equipment health AI at Pearl GTL", "department": "HSE & emissions", "secondary_departments": ["Maintenance & reliability"], "category": "Internal adoption", "stage": "Deployed by September 2025", "practical_relevance": "Near-real-time identification of unsafe conditions and equipment issues supports earlier action and reduces manual inspection needs.", "business_problem": "Large plant work and critical machinery create safety risks and inspection burden; timely signals can allow intervention before incidents or equipment problems escalate.", "workflow": "Computer vision identifies unsafe behavior and hazardous conditions and alerts supervisors; a separate AI capability provides near-real-time equipment-health insights for field intervention.", "technology_data": "Machine vision and AI; video/imagery inputs or sensor types and model details are not disclosed.", "integration_human_role": "Supervisors receive alerts; operations teams act on equipment insights.", "deployment_scope": "Pearl GTL plant in Qatar; Shell says the site was recognized by WEF Global Lighthouse Network in Sep 2025.", "partners": [], "disclosed_results": [], "economics": "No AI-specific safety or maintenance outcomes quantified.", "timeline": [{"date": "2025-09", "event": "Pearl GTL recognized as WEF Global Lighthouse; Shell describes deployed safety and equipment-health AI.", "source_ids": ["RS014"]}], "limitations": ["Shell does not report model accuracy, alert volume, incidents avoided or maintenance savings."], "assessment": "Analyst inference: specific operating use of AI in safety and asset reliability is verified, although outcome evidence is limited.", "sources": [{"id": "RS014", "title": "Safety at Pearl: Every second counts, every role matters", "url": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "publisher": "Shell", "date": null, "accessed": "2026-09-27", "type": "company announcement", "locator": "Being smart about safety", "review_status": "full text reviewed", "supports": ["Shell says Pearl GTL uses machine vision to flag unsafe behaviors and hazardous conditions in real time and alert supervisors.", "Pearl uses AI for near-real-time equipment health insights to reduce manual inspections and support earlier intervention.", "The plant was recognized by the World Economic Forum Global Lighthouse Network in September 2025."], "original_id": "majors-019"}], "field_sources": {"business_problem": ["RS014"], "workflow": ["RS014"], "technology_data": ["RS014"], "integration_human_role": ["RS014"], "deployment_scope": ["RS014"], "economics": ["RS014"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-019", "source_ids": ["RS014"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-140"]}, {"key": "woodside-startup-advisor", "existing_ids": ["AI-043"], "company": "Woodside", "title": "Startup Advisor for LNG plant commissioning/startup", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Proven in live operations; date of first deployment not disclosed", "practical_relevance": "The Startup Advisor combines AI, automation and digital procedures to guide LNG plant start-up operators step by step and compare live performance with prior proven standards; Woodside reports it has been used in live operations and critical situations, without a dated first deployment or quantified safety/production result.", "business_problem": "LNG plant startups are complex, safety-sensitive and depend on experienced operator knowledge.", "workflow": "Guides operators through digital procedures and compares live startup conditions with established performance standards; it can review past startups and inform real-time decisions in critical situations.", "technology_data": "AI, automation and digital procedures; data inputs and model architecture are not disclosed.", "integration_human_role": "Control-room leads and panel operators use and endorse the tool; tool supports rather than replaces operational judgment.", "deployment_scope": "Woodside LNG operations; specific sites and deployment dates not disclosed.", "partners": [], "disclosed_results": [], "economics": "No quantified time, safety or production benefit disclosed.", "timeline": [{"date": "undated", "event": "Woodside describes live-operational use in official AI page; reviewed 2026-09-27.", "source_ids": ["S24"]}], "limitations": ["Official page does not date the use case, so deployment cannot be confirmed to have started in research window."], "assessment": "Analyst inference: practical decision-support workflow is concrete, but the deployment date is undated and no performance measure is disclosed.", "sources": [{"id": "S24", "title": "Woodside Artificial Intelligence — Startup Advisor", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "publisher": "Woodside", "date": null, "accessed": "2026-09-27", "type": "company announcement", "locator": "Startup Advisor", "review_status": "full text reviewed", "supports": ["Startup Advisor uses AI, automation and digital procedures to guide LNG plant operators through startup steps and compare real-time performance to a standard.", "Woodside says it has been proven in live operations and used during critical situations to support decisions; control-room leads and panel operators endorsed the tool."], "original_id": "majors-021"}], "field_sources": {"business_problem": ["S24"], "workflow": ["S24"], "technology_data": ["S24"], "integration_human_role": ["S24"], "deployment_scope": ["S24"], "economics": ["S24"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "majors", "packet_id": "majors-020", "source_ids": ["S24"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-043"]}, {"key": "aramco-fadhili-agr", "existing_ids": [], "company": "Aramco", "title": "Reinforcement-learning control for acid gas removal at Fadhili", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Permanent adoption at Fadhili after pilot; additional facilities assessed for suitability", "practical_relevance": "This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.", "business_problem": "Acid-gas removal performance affects amine and steam consumption, power, stability and operator workload.", "workflow": "The AI system monitors the existing AGR process and makes ongoing adjustments to improve sour-to-sweet gas conversion. The pilot was conducted Oct 2024-Apr 2025, followed by permanent adoption.", "technology_data": "Yokogawa-developed reinforcement-learning AI; data sources, controls architecture and operating envelope are not disclosed.", "integration_human_role": "Operators remain responsible for plant operations; Aramco reports fewer manual interventions. Human approval/safety interlocks are not described.", "deployment_scope": "Fadhili Gas Plant; further Aramco sites under assessment.", "partners": ["Yokogawa Middle East & Africa", "Yokogawa Digital"], "disclosed_results": [{"metric": "amine and steam use reduction", "value": "10%-15%", "period": "Fadhili pilot, Oct 2024-Apr 2025", "attribution": "Company-reported", "source_ids": ["RS015"]}, {"metric": "power use reduction", "value": "around 5%", "period": "Fadhili pilot, Oct 2024-Apr 2025", "attribution": "Company-reported", "source_ids": ["RS015"]}], "economics": "Company reports 10-15% reduction in amine and steam use, ~5% lower power use, improved stability and fewer manual interventions. Quantified dollar savings/emissions are not disclosed.", "timeline": [{"date": "2024-10 to 2025-04", "event": "Fadhili pilot period", "source_ids": ["RS015"]}, {"date": "After pilot; exact date undisclosed", "event": "Company says technology permanently adopted at Fadhili; other sites assessed", "source_ids": ["RS015"]}], "limitations": ["Metric attribution is company-reported and evaluation details are not disclosed.", "The company source page is undated; accessed 2026-09-27. The specific month permanent adoption occurred is not disclosed."], "assessment": "Analyst inference: This is a specific process-control deployment with reported operating metrics and permanent use at one gas plant. RL control in a continuous process could reduce reagent, steam and power demand; wider rollout and repeatability remain unproven.", "sources": [{"id": "RS015", "title": "Aramco AI and Big Data", "url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "publisher": "Aramco", "date": "Undated; accessed 2026-09-27", "type": "company announcement", "locator": "Fadhili Gas Plant section", "review_status": "full text reviewed", "supports": ["RL pilot for AGR optimization ran Oct 2024-Apr 2025; Aramco says permanent adoption followed.", "Company reports 10-15% lower amine and steam use, around 5% lower power use, greater stability and fewer manual operator interventions."], "accessed_date": "2026-09-27", "original_id": "aramco_fadhili"}], "field_sources": {"business_problem": ["RS015"], "workflow": ["RS015"], "technology_data": ["RS015"], "integration_human_role": ["RS015"], "deployment_scope": ["RS015"], "economics": ["RS015"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": ["AI-038"], "cohort": "nocs", "packet_id": "nocs-001", "source_ids": ["RS015"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-141"]}, {"key": "adnoc-energyai", "existing_ids": ["AI-040"], "company": "ADNOC", "title": "ENERGYai agentic subsurface workflows", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "90-day PoC completed Jan 2025; planned operational version/test deployment was forward-looking", "practical_relevance": "Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.", "business_problem": "Accelerate seismic interpretation and improve reservoir monitoring/production insights using large volumes of company subsurface and operational data.", "workflow": "Agents were evaluated for seismic survey analysis, reservoir-performance insight and anomaly detection; users could query results in natural language. Trial findings included improved data quality through error detection, standardization and dataset enrichment.", "technology_data": "70B-parameter LLM, 50+ years ADNOC knowledge, petabytes of internal data; trial used data from >15% of onshore/offshore wells. Detailed models, data controls and integrations not disclosed.", "integration_human_role": "ADNOC subject-matter experts participated; 100+ specialists and secure AI infrastructure supported the trial. Engineers interact with natural-language outputs.", "deployment_scope": "Proof-of-concept scope; next version was described as five agents with field tests and intended scale to thousands of wells, but completion evidence not included here.", "partners": ["AIQ", "G42", "Microsoft"], "disclosed_results": [{"metric": "accuracy improvement in major seismic interpretation aspects", "value": "70%", "period": "90-day PoC concluded Jan 2025", "attribution": "Company-reported", "source_ids": ["RS016"]}], "economics": "No financial return or operating impact disclosed; company reports 70% improvement in accuracy in major seismic interpretation aspects.", "timeline": [{"date": "2024-11-04", "event": "ENERGYai launched/development announced (outside primary source window but confirmed retrospectively in Jan 2025 release)", "source_ids": ["RS016"]}, {"date": "2025-01-16", "event": "90-day PoC completion announced; next operating version expected H1 2025, per plan", "source_ids": ["RS016"]}], "limitations": ["Jan 2025 source verifies a PoC and forward plan, not the later completion or broad deployment claims in existing records."], "assessment": "Analyst inference: Shows concrete agent workflows and a mixed subsurface/production trial. The accuracy claim is promising but denominator, benchmark and independent validation are not supplied.", "sources": [{"id": "RS016", "title": "ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution", "url": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "publisher": "ADNOC", "date": "2025-01-16", "type": "company announcement", "locator": "Trial findings and next-phase paragraphs", "review_status": "full text reviewed", "supports": ["90-day ENERGYai proof-of-concept used a 70-billion-parameter LLM with ADNOC knowledge and proprietary operational data.", "Trial supported seismic interpretation, reservoir monitoring and anomaly detection; ADNOC reports 70% better accuracy in major seismic interpretation aspects.", "Initial planned operational version had five subsurface agents, asset test deployment and intention to scale across thousands of wells."], "original_id": "adnoc_energyai"}], "field_sources": {"business_problem": ["RS016"], "workflow": ["RS016"], "technology_data": ["RS016"], "integration_human_role": ["RS016"], "deployment_scope": ["RS016"], "economics": ["RS016"]}, "verification": "verified primary", "corrections_to_existing": "Exact-match audit: ENERGYai maps only to AI-040. AI-071 is employee capability-building, not the same initiative.", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-002", "source_ids": ["RS016"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-040"]}, {"key": "adnoc-aipso", "existing_ids": ["AI-011"], "company": "ADNOC", "title": "AiPSO production system optimization", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Launched Nov 2025; initial deployment across eight fields; rollout plan through 2027", "practical_relevance": "A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.", "business_problem": "Speed diagnosis of well-performance issues and production-system optimization that otherwise takes engineers days.", "workflow": "Connects office and field work so engineers can identify issues and optimize wells in minutes; platform is intended to coordinate work across assets.", "technology_data": "AI-powered production system optimization platform; source does not disclose model architecture, sensors, workflow integrations or specific recommendations.", "integration_human_role": "Engineers use platform outputs for operational decisions; autonomous actuation is not claimed.", "deployment_scope": "Initial eight fields with planned onshore/offshore deployment across ADNOC by 2027.", "partners": ["SLB"], "disclosed_results": [{"metric": "complex tasks speed-up", "value": "up to 10x", "period": "Nov 2025 launch disclosure", "attribution": "Company-reported", "source_ids": ["S06"]}], "economics": "No monetary return disclosed; ADNOC claims complex work up to 10x faster.", "timeline": [{"date": "2025-11-03", "event": "Platform launched; initial deployment across eight fields", "source_ids": ["S06"]}, {"date": "2027", "event": "Target for all onshore and offshore fields", "source_ids": ["S06"]}], "limitations": ["Company-reported disclosure; no independent validation identified in reviewed sources."], "assessment": "Analyst inference: A named, asset-level deployment with a defined rollout horizon. The workflow claims are operationally plausible but source provides no baseline methodology or production uplift.", "sources": [{"id": "S06", "title": "ADNOC and SLB Launch Transformative AI-Powered Solution to Boost Upstream Productivity", "url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "publisher": "ADNOC", "date": "2025-11-03", "type": "company announcement", "locator": "AiPSO deployment paragraphs", "review_status": "full text reviewed", "supports": ["Production System Optimization platform launched with initial deployment across eight fields.", "Connects office and field workflows so engineers diagnose issues and optimize wells in minutes rather than days; planned deployment across all onshore/offshore fields by 2027.", "ADNOC says complex tasks can be completed up to ten times faster; this is a company claim, not independently validated."], "original_id": "adnoc_aipso"}], "field_sources": {"business_problem": ["S06"], "workflow": ["S06"], "technology_data": ["S06"], "integration_human_role": ["S06"], "deployment_scope": ["S06"], "economics": ["S06"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-003", "source_ids": ["S06"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-011"]}, {"key": "petronas-ermai", "existing_ids": [], "company": "PETRONAS", "title": "ERMAI subsurface opportunity identification", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "In-house AI solution; value uplift reported for 2025", "practical_relevance": "A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.", "business_problem": "Find bypassed hydrocarbons and accelerate subsurface analysis.", "workflow": "ERMAI analyzes subsurface data for behind-casing opportunities and real-time petrophysics.", "technology_data": "In-house AI; models/data integration unspecified.", "integration_human_role": "Decision rights/validation not described.", "deployment_scope": "PETRONAS upstream; asset count unspecified.", "partners": ["Internal PETRONAS development"], "disclosed_results": [{"metric": "subsurface analysis time", "value": "~6 days to 6 seconds", "period": "2025 report", "attribution": "Company-reported", "source_ids": ["RS017"]}], "economics": "PETRONAS reports analysis time falling from ~6 days to 6 seconds and 2025 recovery value uplift; amount not given.", "timeline": [{"date": "2025-12-31", "event": "2025 performance reported in Integrated Report 2025", "source_ids": ["RS017"]}], "limitations": ["Source is excerpt review of the annual report; details of the metric baseline and the value-uplift calculation are not disclosed."], "assessment": "Analyst inference: A sharp cycle-time claim and connection to reserve recovery make this commercially material. Benefit magnitude and model/asset coverage require additional substantiation; report does not quantify dollars.", "sources": [{"id": "RS017", "title": "PETRONAS Integrated Report 2025", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "publisher": "PETRONAS", "date": "2026-02-27", "type": "annual report", "locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)", "review_status": "excerpt reviewed", "supports": ["ERMAI subsurface analysis and company-reported analysis-time/recovery claims."], "original_id": "petronas_report"}], "field_sources": {"business_problem": ["RS017"], "workflow": ["RS017"], "technology_data": ["RS017"], "integration_human_role": ["RS017"], "deployment_scope": ["RS017"], "economics": ["RS017"]}, "verification": "verified primary", "corrections_to_existing": "Exact-match audit: ERMAI is a separate initiative; AI-050 (TIRAI) and AI-080 (myPROdata) are related, not exact matches.", "related_existing_ids": ["AI-050", "AI-080"], "cohort": "nocs", "packet_id": "nocs-004", "source_ids": ["RS017"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-142"]}, {"key": "petronas-tricipta", "existing_ids": [], "company": "PETRONAS", "title": "TriCiptaAI exploration AI products", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "JDA announced Jun 2025; two initial MVP products delivered in Nov 2025; later deployed per 2026 report", "practical_relevance": "Joint development turns the partnership into named products and reported deployment. Material execution proof exists, though value and comparative quality metrics remain undisclosed.", "business_problem": "Reduce uncertainty in exploration screening.", "workflow": "AI.SEEK and Global Exploration Basin join geoscience data for basin screening.", "technology_data": "AI geospatial tools; models and integrations unspecified.", "integration_human_role": "Exploration teams interpret results; autonomy not described.", "deployment_scope": "PETRONAS Carigali; basin scope undisclosed.", "partners": ["Beicip-Franlab", "AFED Digital"], "disclosed_results": [], "economics": "No quantified return or accuracy disclosed.", "timeline": [{"date": "2025-06-18", "event": "TriCipta AI joint development agreement announced", "source_ids": ["RS018"]}, {"date": "2025-11", "event": "AI.SEEK and GEB first MVPs delivered, per 2025 Integrated Report", "source_ids": ["RS017"]}], "limitations": ["AI.SEEK and GEB are identified as first MVPs in 2025 and later reported as deployed; rollout scope is not specified."], "assessment": "Analyst inference: Joint development turns the partnership into named products and reported deployment. Material execution proof exists, though value and comparative quality metrics remain undisclosed.", "sources": [{"id": "RS018", "title": "PETRONAS Carigali Unveils Pioneering AI Venture - TriCipta AI", "url": "https://www.petronas.com/bm/node/1688", "publisher": "PETRONAS", "date": "2025-06-18", "type": "company announcement", "locator": "Media release", "review_status": "full text reviewed", "supports": ["PETRONAS Carigali, Beicip-Franlab and AFED Digital entered a joint development agreement through TriCipta AI.", "Work is aimed at exploration, development and production challenges, including discovery, uncertainty reduction and recovery."], "original_id": "petronas_tricipta"}, {"id": "RS017", "title": "PETRONAS Integrated Report 2025", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "publisher": "PETRONAS", "date": "2026-02-27", "type": "annual report", "locator": "Upstream performance / Enhanced Resource Monetisation Artificial Intelligence (ERMAI)", "review_status": "excerpt reviewed", "supports": ["PETRONAS describes TriCiptaAI MVPs AI.SEEK and Global Exploration Basin."], "original_id": "petronas_report"}], "field_sources": {"business_problem": ["RS017"], "workflow": ["RS017"], "technology_data": ["RS017"], "integration_human_role": ["RS017"], "deployment_scope": ["RS017"], "economics": ["RS017"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-005", "source_ids": ["RS018", "RS017"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-143"]}, {"key": "petronas-tirai", "existing_ids": ["AI-050"], "company": "PETRONAS", "title": "TIRAI GenAI assistant for myPROdata subsurface data", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Commercial AI product", "stage": "Live from 18 Oct 2025 for MBR2025 block pages", "practical_relevance": "A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.", "business_problem": "Reduce friction in exploring Malaysia petroleum acreage data and finding relevant technical documents.", "workflow": "Chatbot assists users on SmartViewer, SmartData Catalog, MBR and MBR+ pages and links to ElasticDocs reports, maps, images, seismic data and well markers.", "technology_data": "Generative AI chatbot connected to PETRONAS myPROdata; underlying model, retrieval/indexing and safeguards not disclosed.", "integration_human_role": "Supports subscribers/investors in data exploration; human interpretation remains necessary for subsurface evaluation.", "deployment_scope": "myPROdata pages for MBR2025 blocks.", "partners": ["PETRONAS myPROdata"], "disclosed_results": [], "economics": "Commercial value and user outcomes not disclosed.", "timeline": [{"date": "2025-10-18", "event": "TIRAI went live on selected MBR2025 pages", "source_ids": ["S29"]}], "limitations": ["No subscriber adoption or impact metrics disclosed."], "assessment": "Analyst inference: A named AI feature within a real upstream data-access product supports acreage evaluation; evidence covers launch and function, not impact.", "sources": [{"id": "S29", "title": "TIRAI - Your New PETRONAS myPROdata Intelligence Assistance is Live!", "url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "publisher": "PETRONAS myPROdata", "date": "2025-11-03", "type": "company announcement", "locator": "TIRAI release", "review_status": "full text reviewed", "supports": ["TIRAI GenAI chatbot went live on 18 October 2025 on SmartViewer, SmartData Catalog, MBR and MBR+ pages for MBR2025 blocks.", "Assistant links users to ElasticDocs resources including reports, maps, images, seismic data and well markers for subsurface data exploration."], "original_id": "petronas_myprodata"}], "field_sources": {"business_problem": ["S29"], "workflow": ["S29"], "technology_data": ["S29"], "integration_human_role": ["S29"], "deployment_scope": ["S29"], "economics": ["S29"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-006", "source_ids": ["S29"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-050"]}, {"key": "cnpc-kunlun", "existing_ids": ["AI-081"], "company": "PetroChina", "title": "Kunlun large model with value-chain applications", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "70B model and 43 petroleum scenarios in 2024; 2026 update reports 152 deployed scenarios", "practical_relevance": "The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.", "business_problem": "Apply a petroleum-domain language/model platform to analysis and tasks across exploration, development, refining and planning.", "workflow": "2026 example: crude procurement question answered with an evaluation, processing plan and risk notes in five minutes versus several hours or days of manual analysis. CNPC describes autonomous planning/task decomposition, tool execution, calculation, prediction and explainability capabilities.", "technology_data": "Company-developed platform, grown from 70B parameters in 2024 to 300B by May 2025, per official reports; joint development with China Mobile, Huawei and iFLYTEK. Uses sector-specific data, AI middleware and computing environment.", "integration_human_role": "Domain teams apply outputs; example comes from a refinery planning manager. Governance, model boundaries and human review are not detailed in reviewed disclosures.", "deployment_scope": "CNPC says 152 scenarios across the value chain are in scaled application by May 2026; scenario-level deployment counts not given.", "partners": ["China Mobile", "Huawei", "iFLYTEK"], "disclosed_results": [{"metric": "crude purchase evaluation cycle", "value": "5 minutes vs hours or days", "period": "May 2026 example", "attribution": "Company-reported", "source_ids": ["RS020"]}, {"metric": "deployed scenarios", "value": "152", "period": "May 2026", "attribution": "Company-reported", "source_ids": ["RS020"]}], "economics": "No consolidated financial impact disclosed. One refinery manager says crude assessment shortened from hours/days to five minutes; not independently validated.", "timeline": [{"date": "2024-08", "event": "Kunlun launched (outside window; baseline detail in 2024 annual report)", "source_ids": ["RS019"]}, {"date": "2025-05", "event": "300B parameter milestone per CNPC announcement; report source documents current scale-up", "source_ids": ["RS020"]}, {"date": "2026-05-28", "event": "152 application scenarios and higher-order AI capabilities announced", "source_ids": ["RS020"]}], "limitations": ["2024 launch predates window start; included only to establish baseline, while 2025-26 capability and scale updates fall inside the window. 2026 source reviewed in full."], "assessment": "Analyst inference: The strongest evidence is the named broad platform and numerous launched use cases, with concrete procurement and technical scenarios. Scenario count is company-reported; quality and value vary by use case.", "sources": [{"id": "RS019", "title": "2024 Annual Report", "url": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "publisher": "CNPC", "date": "2025-08-01", "type": "annual report", "locator": "Launch of 70-billion-parameter Kunlun Large Model", "review_status": "excerpt reviewed", "supports": ["CNPC reports launching a 70-billion-parameter Kunlun model in 2024 with 43 professional/general scenarios for petroleum.", "Report names a joint development group of CNPC, China Mobile, Huawei and iFLYTEK; platform included eight trained large models, 18 application scenarios, specialized datasets, AI middleware and computing environment."], "original_id": "cnpc_annual"}, {"id": "RS020", "title": "中国石油昆仑大模型迈入主动智能新阶段", "url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml", "publisher": "CNPC News Center", "date": "2026-05-29", "type": "company announcement", "locator": "Release overview and application examples", "review_status": "full text reviewed", "supports": ["CNPC says Kunlun reached 152 deployed scenarios across the energy and chemical value chain in May 2026.", "Specific example: crude purchase evaluation with processing plans and risk advice in five minutes versus hours or days of manual analysis, as reported by a refinery manager."], "original_id": "cnpc_2026"}], "field_sources": {"business_problem": ["RS019", "RS020"], "workflow": ["RS019", "RS020"], "technology_data": ["RS019", "RS020"], "integration_human_role": ["RS019", "RS020"], "deployment_scope": ["RS019", "RS020"], "economics": ["RS019", "RS020"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-007", "source_ids": ["RS019", "RS020"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-081"]}, {"key": "sinopec-jianghan", "existing_ids": [], "company": "Sinopec", "title": "Jianghan Oilfield AI for shale gas, drilling, safety and production", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Multiple use cases described as in use in 2025", "practical_relevance": "This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.", "business_problem": "Improve shale-gas development and field work by refreshing models, warning about drilling hazards, tuning fracturing, monitoring safety and alerting on liquid loading.", "workflow": "AI use cases include a shale-gas digital twin with dynamically updated geologic models; drilling risk list generation based on planned well path and complex-event proximity; fracture parameter optimization; video detection of unsafe acts/PPE; drone/video/human inspection; liquid-loading warning; and Jianghan Zhiyan LLM for routine work.", "technology_data": "AI models, video analytics, digital twin and LLM using geological, drilling, fracture, production and safety data. Vendor/model and integration architecture not stated.", "integration_human_role": "Warnings and recommendations support field staff/engineers. Automated surveillance and alerting are described; full autonomous control is not.", "deployment_scope": "Jianghan Oilfield and Fuling shale-gas operating area, including >800 well-instance liquid-loading warning applications.", "partners": ["Sinopec Jianghan Oilfield"], "disclosed_results": [{"metric": "fracture simulation accuracy", "value": ">80%", "period": "2025 disclosure", "attribution": "Company-reported", "source_ids": ["RS021"]}, {"metric": "safety inspection efficiency", "value": ">50% increase", "period": "2025 disclosure", "attribution": "Company-reported", "source_ids": ["RS021"]}, {"metric": "liquid-loading alert accuracy", "value": ">95%", "period": "2025 disclosure", "attribution": "Company-reported", "source_ids": ["RS021"]}], "economics": "Company-reported: fracturing simulation accuracy >80%; safety supervision efficiency +50%, missed-violation rate <10%; liquid-loading warnings accuracy >95%, response efficiency >90%. Causal attribution/baselines not fully described.", "timeline": [{"date": "2025-05-28", "event": "Jianghan deployment portfolio and use cases published", "source_ids": ["RS021"]}], "limitations": ["Chinese subsidiary/field source; group-wide deployment is not established. Performance measures are company-reported."], "assessment": "Analyst inference: This is the broadest detailed field-use disclosure in the cohort and exposes real workflows rather than an AI ambition. Metrics span distinct systems and should not be pooled into a single AI benefit claim.", "sources": [{"id": "RS021", "title": "〖中国石化报〗多场景推进人工智能技术应用", "url": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "publisher": "Sinopec / Jianghan Oilfield", "date": "2025-05-28", "type": "company announcement", "locator": "Jianghan Oilfield AI deployment section", "review_status": "full text reviewed", "supports": ["Jianghan Oilfield describes AI for shale-gas digital-twin reservoir updates, drilling risk alerts, fracture optimization, video safety monitoring, field inspection and work assistance.", "Reported examples include fracture simulation accuracy above 80%, video safety inspection efficiency up over 50% and missed-violation detection below 10%.", "A liquid-loading warning use case had more than 800 well instances, parsed 150,000 records/day, and company-reported warning accuracy above 95%; treatment efficiency above 90%."], "original_id": "sinopec"}], "field_sources": {"business_problem": ["RS021"], "workflow": ["RS021"], "technology_data": ["RS021"], "integration_human_role": ["RS021"], "deployment_scope": ["RS021"], "economics": ["RS021"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-008", "source_ids": ["RS021"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-144"]}, {"key": "cnooc-ai-plus", "existing_ids": [], "company": "CNOOC", "title": "AI+ smart field, intelligent drilling and typhoon response", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Priority scenarios deployed by mid-2025; smart drilling demonstration results reported", "practical_relevance": "Evidence of operationally connected deployments is relevant, but an important drilling metric belongs to an integrated smart-drilling program and should not be represented as an AI-only result.", "business_problem": "Improve offshore production/injection optimization, drilling/completion performance and weather emergency response across complex offshore assets.", "workflow": "CNOOC scaled intelligent injection-production technology, advanced intelligent drilling/completion, and combined satellite remote sensing, unmanned equipment and AI for typhoon response.", "technology_data": "AI technology named, but algorithms/data and whether systems share common architecture are not disclosed.", "integration_human_role": "Operational personnel coordinate responses; AI aids monitoring and decision support. Autonomous actuation not established.", "deployment_scope": "Includes Shenhai-1 smart gas field and Qinhuangdao 32-6 construction benchmarks; deployment counts not stated.", "partners": ["CNOOC subsidiaries and technical teams"], "disclosed_results": [{"metric": "demonstration project drilling speed", "value": "+26%", "period": "2025 H1", "attribution": "Company-reported; AI-specific causality not established", "source_ids": ["RS022"]}], "economics": "26% drilling-speed improvement is disclosed for demonstration projects but not attributed exclusively to AI. No AI-specific financial returns.", "timeline": [{"date": "2025-08-27", "event": "CNOOC mid-year results describe AI+ scenes, injection-production, smart drilling and weather response", "source_ids": ["RS022"]}], "limitations": ["Existing AI-082 ~20% failure reduction claim is from secondary China Daily coverage and is not re-verified by the cited CNOOC primary source; retain as unverified pending a company/technical source."], "assessment": "Analyst inference: Evidence of operationally connected deployments is relevant, but an important drilling metric belongs to an integrated smart-drilling program and should not be represented as an AI-only result.", "sources": [{"id": "RS022", "title": "2025 Interim Results", "url": "https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html", "publisher": "CNOOC", "date": "2025-08-27", "type": "company announcement", "locator": "AI+ key scenarios and intelligent drilling paragraphs", "review_status": "full text reviewed", "supports": ["CNOOC reported implementing priority AI+ scenarios, scaling intelligent injection-production technology, and advancing intelligent drilling/completion construction.", "Company reported demonstration project drilling speed improvement of 26%; the release does not allocate this result exclusively to AI.", "Deep Sea No.1 smart gas field named as a national excellent-level smart factory; satellite remote sensing, unmanned equipment and AI algorithms were combined for typhoon response."], "original_id": "cnooc2025"}], "field_sources": {"business_problem": ["RS022"], "workflow": ["RS022"], "technology_data": ["RS022"], "integration_human_role": ["RS022"], "deployment_scope": ["RS022"], "economics": ["RS022"]}, "verification": "verified primary", "corrections_to_existing": "Exact-match audit: CNOOC smartfield/drilling/typhoon umbrella is distinct from AI-068 terminal Symmetry process optimization and AI-082 predictive-maintenance metric; retain them only as related records.", "related_existing_ids": ["AI-068", "AI-082"], "cohort": "nocs", "packet_id": "nocs-009", "source_ids": ["RS022"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-145"]}, {"key": "petrobras-smart-tocha", "existing_ids": [], "company": "Petrobras", "title": "Smart Tocha AI flare combustion monitoring and adjustment", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Technology developed; company page says availability across all new units by 2026", "practical_relevance": "A specific closed-loop operational application, with a named research partner and rollout target. Actual adoption count and measured results remain unknown.", "business_problem": "Monitor flaring and optimize steam flow to flare safely and efficiently.", "workflow": "AI monitors flare burning and adjusts steam flow automatically.", "technology_data": "AI monitoring and flow adjustment; data sources, model, safeguards and control system interface not disclosed.", "integration_human_role": "Automatic adjustment claimed; operator oversight and fail-safe arrangements are not described.", "deployment_scope": "Planned availability on new units by 2026; operational count not supplied.", "partners": ["PUC-Rio"], "disclosed_results": [], "economics": "No quantified emission, gas or cost result in the cited page.", "timeline": [{"date": "2026", "event": "Company-stated planned availability across new units by 2026; actual completion not verified", "source_ids": ["RS023"]}], "limitations": ["Page also describes other technologies; only the AI-linked flare control is represented here. Source is a live corporate page with no publication date; crawl/access date 2026-09-27."], "assessment": "Analyst inference: A specific closed-loop operational application, with a named research partner and rollout target. Actual adoption count and measured results remain unknown.", "sources": [{"id": "RS023", "title": "Inovação e tecnologia: nossa fonte de energia", "url": "https://petrobras.com.br/inovacao-e-tecnologia", "publisher": "Petrobras", "date": "Undated; accessed 2026-09-27", "type": "company announcement", "locator": "Smart Tocha entry", "review_status": "full text reviewed", "supports": ["Smart Tocha, developed with PUC-Rio, uses AI to monitor flare combustion and automatically adjust steam flow.", "Petrobras says the technology will be available in all new units by 2026; this wording is prospective and page does not establish full rollout or quantified outcome."], "accessed_date": "2026-09-27", "original_id": "petrobras_smarttocha"}], "field_sources": {"business_problem": ["RS023"], "workflow": ["RS023"], "technology_data": ["RS023"], "integration_human_role": ["RS023"], "deployment_scope": ["RS023"], "economics": ["RS023"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-010", "source_ids": ["RS023"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-146"]}, {"key": "ecopetrol-ai-ecosystem", "existing_ids": [], "company": "Ecopetrol", "title": "AI and advanced analytics ecosystem; intelligent upstream planning", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Ecosystem implemented in 2025; maturity assessment identified constraints", "practical_relevance": "The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.", "business_problem": "Create shared AI/analytics capabilities and enable business initiatives; an upstream planner addresses repeated planning work.", "workflow": "Ecopetrol reports an AI and advanced analytics ecosystem and an intelligent upstream planner that released staff hours. Other named application details were not disclosed in this report.", "technology_data": "Shared capabilities, resources and infrastructure; report cites limited data access and autonomy as maturity gaps.", "integration_human_role": "Business teams and AI community; human decision processes implied, details not disclosed.", "deployment_scope": "Group ecosystem; intelligent planner in upstream, asset count not specified.", "partners": ["Internal Ecopetrol teams"], "disclosed_results": [{"metric": "labor hours released by intelligent planner", "value": "4,609 hours", "period": "2025", "attribution": "Company-reported", "source_ids": ["RS024"]}], "economics": "Planner reportedly released 4,609 labor hours. Company reporting does not establish this as a monetary return or validate the causal contribution of AI.", "timeline": [{"date": "2025-12-31", "event": "Ecosystem and planner outcomes included in 2025 report", "source_ids": ["RS024"]}], "limitations": ["AI maturity and access issues remain; annual-report excerpt reviewed, so claims limited to the surfaced passages."], "assessment": "Analyst inference: The disclosure shows both an enabling capability and a narrow productivity result, while candidly naming maturity limits. Avoid attributing Ecopetrol-wide innovation benefits to AI.", "sources": [{"id": "RS024", "title": "Informe de gestión y resultados 2025", "url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "publisher": "Ecopetrol", "date": "2026-03-31", "type": "annual report", "locator": "Technology and innovation section; AI ecosystem and upstream planner", "review_status": "excerpt reviewed", "supports": ["Ecopetrol reports establishing an AI and advanced analytics ecosystem and identifies data access and autonomy limits in its maturity assessment.", "Upstream intelligent planner reportedly released 4,609 labor hours; report does not attribute all AI ecosystem benefits to that planner or quantify causal AI-specific savings."], "original_id": "ecopetrol_2025"}], "field_sources": {"business_problem": ["RS024"], "workflow": ["RS024"], "technology_data": ["RS024"], "integration_human_role": ["RS024"], "deployment_scope": ["RS024"], "economics": ["RS024"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-011", "source_ids": ["RS024"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-147"]}, {"key": "ecopetrol-cade", "existing_ids": [], "company": "Ecopetrol", "title": "CADE energy dispatch decision platform", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Opened May 2026; AI described as a development capability", "practical_relevance": "Operationally useful energy dispatch is a concrete workflow, but it should be coded as AI only at the development/roadmap level until the AI component is shown in production.", "business_problem": "Optimize energy procurement and dispatch decisions across the group.", "workflow": "A single platform selects energy source, time and quantity using self-generation, renewable generation and national grid purchases; analytics and AI are planned to support optimization.", "technology_data": "Decision platform plus analytics/AI development; source does not confirm AI currently controls dispatch or disclose model/data architecture.", "integration_human_role": "Source describes an energy dispatch decision platform; energy managers are the likely users (analyst inference). It does not disclose human approval steps or claim AI currently controls dispatch.", "deployment_scope": "Ecopetrol group operations in Colombia; current generation portfolio and expansion described, but AI deployment footprint not specified.", "partners": ["ISA"], "disclosed_results": [], "economics": "AI-specific economic or energy savings not quantified; source forecasts broader future generation coverage.", "timeline": [{"date": "2026-05-04", "event": "CADE opened; company says AI/analytics will be developed for energy optimization", "source_ids": ["RS025"]}], "limitations": ["AI described prospectively; platform combines deterministic energy management and advanced analytics, so AI deployment is not confirmed."], "assessment": "Analyst inference: Operationally useful energy dispatch is a concrete workflow, but it should be coded as AI only at the development/roadmap level until the AI component is shown in production.", "sources": [{"id": "RS025", "title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/", "publisher": "Ecopetrol", "date": "2026-05-04", "type": "company announcement", "locator": "CADE paragraphs", "review_status": "full text reviewed", "supports": ["CADE is an Ecopetrol-ISA energy dispatch decision platform combining self-generation, renewables and grid purchases to choose source, hour and quantity.", "Company says it will develop technology, advanced analytics and AI for energy optimization; source does not establish AI as the system currently making autonomous dispatch decisions."], "original_id": "ecopetrol_cade"}], "field_sources": {"business_problem": ["RS025"], "workflow": ["RS025"], "technology_data": ["RS025"], "integration_human_role": ["RS025"], "deployment_scope": ["RS025"], "economics": ["RS025"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-012", "source_ids": ["RS025"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-148"]}, {"key": "ypf-three-workflows", "existing_ids": [], "company": "YPF", "title": "Generative AI for contract cataloging, customer service and refinery inspection", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Use cases carried out in 2024 and reported in sustainability report", "practical_relevance": "Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.", "business_problem": "Increase productivity in contract cataloging, make strategic customer information easier to access, and improve inspector decisions/training.", "workflow": "GenAI/NLP automates contract classification; a chatbot organizes strategic information for commercial decisions; a refinery inspection assistant draws on reports, plant shutdown records and historical asset data to support material choice, practices and technical learning.", "technology_data": "Generative AI and NLP; the inspection assistant connects technical document/history sources. Exact models, integrations and user counts not disclosed.", "integration_human_role": "Provides assistance to cataloguers, commercial teams and refinery inspectors; YPF positions AI as human-empowering. No autonomous decisions reported.", "deployment_scope": "Corporate cataloging and customer support; refinery inspection deployment scope not stated.", "partners": ["Internal YPF development"], "disclosed_results": [], "economics": "No quantified results; company says contract cataloging productivity/information quality improve and inspection support can reduce risk, improve material durability and accelerate learning.", "timeline": [{"date": "2024", "event": "YPF reports the three AI projects as use cases conducted in 2024", "source_ids": ["RS026"]}], "limitations": ["2024 initiatives fall inside window but original launch dates or scale are not specified. Annual-report text accessed as excerpt."], "assessment": "Analyst inference: Three practical workflows span shared-service and operational teams. Source documents use cases but provides no performance or rollout scale, so evidence is stronger on existence than outcome.", "sources": [{"id": "RS026", "title": "Reporte de Sustentabilidad 2024", "url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "publisher": "YPF", "date": "2025; publication day not confirmed", "type": "annual report", "locator": "Innovation and Technology / Use of artificial intelligence", "review_status": "excerpt reviewed", "supports": ["YPF lists 2024 use cases: generative AI/NLP contract cataloging; customer-service chatbot; and refinery inspection virtual assistant.", "The inspection assistant uses inspection reports, shutdown records and historical asset data to support material selection, good practices and technical learning; YPF says it can reduce operational risk.", "YPF reports governance standards for LLM security and generative AI and general principles for privacy, cybersecurity, bias and labeling."], "original_id": "ypf_report"}], "field_sources": {"business_problem": ["RS026"], "workflow": ["RS026"], "technology_data": ["RS026"], "integration_human_role": ["RS026"], "deployment_scope": ["RS026"], "economics": ["RS026"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-013", "source_ids": ["RS026"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-149"]}, {"key": "pertamina-pintar", "existing_ids": [], "company": "Pertamina", "title": "PINTAR GenAI HSE question-answering for Rokan workers", "department": "Workforce & learning", "secondary_departments": [], "category": "Internal adoption", "stage": "Described in Pertamina Hulu Energi 2024 annual report", "practical_relevance": "A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.", "business_problem": "Help workers find HSE rules/procedures and improve practical safety awareness at work.", "workflow": "Conversational GenAI answers questions about HSE regulations and procedures; intended to support workplace safety learning.", "technology_data": "ChatGPT-powered GenAI; underlying retrieval corpus, access controls, validation and interface not disclosed.", "integration_human_role": "Worker self-service Q&A; no human-review pathway or escalation behavior described.", "deployment_scope": "Pertamina Hulu Rokan Zone; user count, production deployment scope and usage volume not disclosed.", "partners": ["Pertamina Hulu Rokan"], "disclosed_results": [], "economics": "No outcome metric disclosed.", "timeline": [{"date": "2024", "event": "PINTAR reported in Pertamina Hulu Energi annual report, published 2025", "source_ids": ["RS027"]}], "limitations": ["Primary annual-report excerpt reviewed; product scope and AI design details are sparse."], "assessment": "Analyst inference: A narrowly defined HSE knowledge workflow is identifiable, but operational effect, user adoption and controls remain unknown.", "sources": [{"id": "RS027", "title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "publisher": "Pertamina Hulu Rokan", "date": "2025-09-01", "type": "company announcement", "locator": "PINTAR description", "review_status": "excerpt reviewed", "supports": ["PINTAR is a generative AI conversational platform for answering workplace health, safety and environment regulations and procedures and supporting worker safety learning.", "Source is a 2024 upstream-subholding annual-report document excerpt; user scale, deployment metrics and outcome measures are not disclosed."], "original_id": "pertamina_pintar"}], "field_sources": {"business_problem": ["RS027"], "workflow": ["RS027"], "technology_data": ["RS027"], "integration_human_role": ["RS027"], "deployment_scope": ["RS027"], "economics": ["RS027"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-014", "source_ids": ["RS027"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-150"]}, {"key": "pertamina-geothermal-gbionic", "existing_ids": [], "company": "Pertamina", "title": "PGE G-Bionic machine learning and digital twins", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Digital program described Oct 2025; value figures are estimates", "practical_relevance": "Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.", "business_problem": "Improve geothermal drilling economics, plant efficiency, well productivity and lost-production performance.", "workflow": "PGE links machine learning and digital twins to well optimization, plant efficiency and reduced lost production opportunities. Specific workflows or asset-level examples not detailed in this release.", "technology_data": "Machine learning and digital twins; source does not detail data, models, controls or specific production integrations.", "integration_human_role": "Engineering and operating teams presumably apply recommendations; exact human role not stated.", "deployment_scope": "PGE geothermal operations; site coverage not disclosed.", "partners": ["Pertamina Geothermal Energy internal teams"], "disclosed_results": [{"metric": "estimated production increase", "value": "up to 6%", "period": "program estimate in Oct 2025", "attribution": "Company-reported estimate; machine learning/digital twin program", "source_ids": ["RS028"]}, {"metric": "estimated annual net income equivalent value creation", "value": "5%-7%", "period": "program estimate in Oct 2025", "attribution": "Company-reported estimate; not AI-only", "source_ids": ["RS028"]}], "economics": "PGE estimates up to 6% drilling cost reduction, production increase up to 6%, and value equivalent to 5-7% annual net income increase. These are program estimates, not realized/independently audited AI-only returns.", "timeline": [{"date": "2025-10-27", "event": "G-Bionic program benefits presented at PGE Digital Day 2025", "source_ids": ["RS028"]}], "limitations": ["Figures are estimates for a combined digital transformation program and should not be attributed solely to AI.", "Pertamina Geothermal Energy is a geothermal subsidiary; this is an adjacent-energy example rather than oil and gas operations."], "assessment": "Analyst inference: Material possible economics and a clear business area justify tracking, while attribution remains at a bundled digital program level rather than AI-only.", "sources": [{"id": "RS028", "title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "publisher": "Pertamina Geothermal Energy", "date": "2025-10-27", "type": "company announcement", "locator": "G-Bionic value case", "review_status": "full text reviewed", "supports": ["PGE describes a digital transformation program using machine learning and digital twins for well optimization, plant efficiency and lost-production opportunity reduction.", "PGE estimates potential production increases up to 6%, drilling cost reductions up to 6%, and value equivalent to 5-7% net income annually; these are program estimates, not audited realized AI benefits."], "original_id": "pertamina_geothermal"}], "field_sources": {"business_problem": ["RS028"], "workflow": ["RS028"], "technology_data": ["RS028"], "integration_human_role": ["RS028"], "deployment_scope": ["RS028"], "economics": ["RS028"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-015", "source_ids": ["RS028"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-151"]}, {"key": "ongc-gaslift-tender", "existing_ids": [], "company": "ONGC", "title": "AI/ML-enabled offshore gas-lift optimization procurement specification", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Tender specification issued Nov 2025; procurement is not evidence of operational deployment", "practical_relevance": "The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.", "business_problem": "Automate gas-lift monitoring and optimize injection while detecting equipment degradation to reduce downtime.", "workflow": "Tender calls for real-time well-data analysis, gas-lift injection optimization and predictive maintenance alerts from valve/pressure behavior, integrated with existing wellhead control panels and SCADA.", "technology_data": "AI/ML specified; no selected vendor, algorithm, data history or acceptance results disclosed.", "integration_human_role": "Operators/engineers receive local HMI and remote SCADA visibility; autonomous lift behavior is included as demonstration requirement, but actual commissioned control is unverified.", "deployment_scope": "Offshore wells under the tender scope; installed site count not verified.", "partners": ["Prospective bidder(s) not identified"], "disclosed_results": [], "economics": "No outcome or cost data; tender is a planned/procurement specification.", "timeline": [{"date": "2025-11-03", "event": "Tender requirements include AI-driven optimization and predictive maintenance", "source_ids": ["RS029"]}], "limitations": ["Tender publication date metadata surfaced as 3 Nov 2025; full procurement result and commissioning were not located in this pass."], "assessment": "Analyst inference: The tender demonstrates a well-defined intended AI workflow and industrial integration, but it should not be reported as an operating ONGC AI deployment until award and commissioning are confirmed.", "sources": [{"id": "RS029", "title": "ONGC tender for automated gas-lift system", "url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "publisher": "ONGC", "date": "2025-11-03", "type": "company announcement", "locator": "AI-driven optimization and predictive maintenance requirements", "review_status": "excerpt reviewed", "supports": ["Tender technical scope requires AI/ML algorithms to analyze well data and optimize gas-lift injection in real time, with predictive maintenance based on valve response and pressure trends.", "SCADA and existing wellhead control panel integration are specified. Procurement requirements show intended scope, not completed deployment."], "original_id": "ongc_tender"}], "field_sources": {"business_problem": ["RS029"], "workflow": ["RS029"], "technology_data": ["RS029"], "integration_human_role": ["RS029"], "deployment_scope": ["RS029"], "economics": ["RS029"]}, "verification": "verified primary", "corrections_to_existing": "", "related_existing_ids": [], "cohort": "nocs", "packet_id": "nocs-016", "source_ids": ["RS029"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-152"]}, {"key": "adnoc-ai-workforce", "existing_ids": ["AI-071"], "related_existing_ids": [], "company": "ADNOC", "title": "Enterprise AI workforce training and agentic AI capability building", "department": "Workforce & learning", "secondary_departments": [], "category": "Internal adoption", "stage": "By 2025, >40,000 employees completed AI training; 2025 report also records a separate agentic AI learning program", "practical_relevance": "The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.", "business_problem": "Develop employee AI literacy and practical capability as ADNOC expands enterprise AI use.", "workflow": "ADNOC reports enterprise-wide generative AI rollout with Microsoft Copilot in Nov 2023; the 2025 sustainability report separately identifies Introduction to Agentic AI training, engaging over 18,000 participants and resulting in over 10,000 AI agents created across operations.", "technology_data": "Microsoft Copilot rollout, internal AI learning programs and agent-creation activity; detailed curriculum, agent architecture and usage controls not disclosed.", "integration_human_role": "Training and enablement are human capability measures; the sources do not attribute productivity gains specifically to the training program.", "deployment_scope": "ADNOC-wide workforce figures; 2025 report gives learner and agent counts but no breakdown by business unit or operational risk tier.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Employees completing AI training", "value": ">40,000", "period": "As reported Nov. 2025", "attribution": "Company-reported aggregate", "source_ids": ["RS030"]}, {"metric": "Monthly productivity hours gained", "value": ">70,000 hours/month", "period": "As reported Nov. 2025 since 2023 enterprise Copilot rollout", "attribution": "Company-reported; not isolated to training", "source_ids": ["RS030"]}, {"metric": "Agentic AI training participants / agents created", "value": ">18,000 participants / >10,000 agents", "period": "2025", "attribution": "Company-reported", "source_ids": ["RS031"]}], "economics": "No training cost, controlled productivity baseline or causal attribution to training disclosed.", "timeline": [{"date": "2023-11", "event": "ADNOC says it rolled out enterprise generative AI with Microsoft Copilot; pre-window context.", "source_ids": ["RS030"]}, {"date": "2025-11-02", "event": "ADNOC reports training and productivity aggregates alongside a new co-development agreement with Microsoft.", "source_ids": ["RS030"]}, {"date": "2025", "event": "Sustainability report documents agentic AI learning program activity.", "source_ids": ["RS031"]}], "limitations": ["40,000 trained employees, >90% utilization and 70,000 hours/month are corporate aggregate disclosures, not training-attributable outcomes.", "The reported >10,000 created agents are not evidence that all were deployed, used, or agentic to the same extent."], "assessment": "Analyst inference: The initiative is practical organizational enablement, but published counts measure reach and self-reported aggregate productivity rather than causal training return.", "sources": [{"id": "RS030", "title": "ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI", "url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "publisher": "ADNOC", "date": "2025-11-02", "type": "company announcement", "locator": "Copilot history, training and productivity aggregate", "review_status": "full text reviewed", "supports": ["ADNOC reports enterprise Copilot rollout in 2023, >40,000 employees completing AI training, utilization above 90%, and >70,000 monthly productivity hours; no training causal attribution stated."], "original_id": "adnoc_msft_ai"}, {"id": "RS031", "title": "ADNOC Sustainability Report 2025", "url": "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "publisher": "ADNOC", "date": "2026", "type": "annual report", "locator": "AI and digital capabilities / workforce training", "review_status": "full text reviewed", "supports": ["In 2025, ADNOC says its Introduction to Agentic AI program engaged >18,000 participants and resulted in >10,000 agents created across operations."], "original_id": "adnoc_sustainability"}], "field_sources": {"business_problem": ["RS030", "RS031"], "workflow": ["RS030", "RS031"], "technology_data": ["RS030", "RS031"], "integration_human_role": ["RS030", "RS031"], "deployment_scope": ["RS030", "RS031"], "economics": ["RS030", "RS031"]}, "verification": "verified primary", "corrections_to_existing": "Verified primary source for the workforce training/enablement record; it is separate from ENERGYai (AI-040).", "cohort": "nocs", "packet_id": "nocs-017", "source_ids": ["RS030", "RS031"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-071"]}, {"key": "petronas-myprodata-agentic", "existing_ids": ["AI-080"], "related_existing_ids": ["AI-050"], "company": "PETRONAS", "title": "Agentic AI enhancement to Malaysia’s myPROdata upstream investment platform", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Commercial AI product", "stage": "Agreement signed Sept. 3, 2026; platform enhancement/rollout planned", "practical_relevance": "The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.", "business_problem": "Help investors and industry participants navigate Malaysia’s upstream opportunity data and perform faster E&P screening/evaluation.", "workflow": "Agentic AI is intended to connect and analyze geological, geophysical, engineering and other datasets, supporting prospective-area discovery, resource assessment, development evaluation and commerciality analysis.", "technology_data": "Existing web myPROdata platform with subsurface/surface data; Iraya Energies to integrate the agentic-AI capability and establish technical, commercial and security foundations.", "integration_human_role": "Investors remain decision makers; public release says agents analyze and connect data but does not define human approval controls or autonomous investment decisions.", "deployment_scope": "Malaysia upstream data platform offered to investors and industry; agreement signed, broad rollout not yet evidenced.", "partners": ["Iraya Energies"], "disclosed_results": [], "economics": "RM50-60 billion/year upstream investment and reducing discovery-to-first-hydrocarbon timeline from 100 to 50 months are PETRONAS/MPM national ambitions, not project-specific AI savings or outcomes.", "timeline": [{"date": "2026-09-03", "event": "MPM/PETRONAS signed an agreement with Iraya Energies to add agentic AI to myPROdata.", "source_ids": ["RS032"]}], "limitations": ["Agreement and intended capabilities are disclosed; production deployment, usage and achieved decision speed are not yet reported.", "Investment and cycle-time goals are national upstream targets, not attribution to the myPROdata AI enhancement."], "assessment": "Analyst inference: The disclosure identifies a tangible upstream investment workflow and named integrator, but the platform enhancement is at agreement stage and no results have been measured publicly.", "sources": [{"id": "RS032", "title": "PETRONAS Advances Malaysia’s Upstream Data Platform with Agentic AI to Accelerate E&P Investment", "url": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep", "publisher": "PETRONAS", "date": "2026-09-03", "type": "company announcement", "locator": "Agreement, platform workflow and MPM targets", "review_status": "full text reviewed", "supports": ["PETRONAS/MPM signed an agreement with Iraya Energies to add agentic AI to myPROdata for Malaysia exploration and production data evaluation.", "Iraya is lead integrator; the release states investment and cycle-time targets as broader national ambitions, not achieved product outcomes."], "original_id": "petronas_myprodata_ai"}], "field_sources": {"business_problem": ["RS032"], "workflow": ["RS032"], "technology_data": ["RS032"], "integration_human_role": ["RS032"], "deployment_scope": ["RS032"], "economics": ["RS032"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "nocs", "packet_id": "nocs-018", "source_ids": ["RS032"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-080"]}, {"key": "koc-g-agent-drilling", "existing_ids": ["AI-085"], "related_existing_ids": [], "company": "Kuwait Oil Company", "title": "G Agent AI drilling-rig resource management and scheduling", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "KOC’s Oct–Dec 2025 digest describes G Agent and an agentic rig-scheduling project as operational initiatives at its AI Innovation Center", "practical_relevance": "This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.", "business_problem": "Improve rig resource allocation, drilling planning, system coordination and real-time operational decision-making.", "workflow": "G Agent supports real-time data analysis and rig resource management; KOC separately describes an agentic AI rig-scheduling project that it says improved productivity, planning and operational quality.", "technology_data": "AI platform arising from KOC partnership with Microsoft and Halliburton; Microsoft Azure and licenses are part of broader government collaboration; Ghaia AI supports technology-program/training management at the Center.", "integration_human_role": "KOC says AI agents support decision-making; human role and authorization boundary for rig-scheduling outputs are not specified.", "deployment_scope": "KOC AI Innovation Center and drilling operations; no rig count, field/site or deployment breadth quantified in the reviewed KOC bulletin.", "partners": ["Microsoft", "Halliburton", "Ghaia AI"], "disclosed_results": [], "economics": "No quantified operational result, contract value or AI-specific cost/savings disclosed.", "timeline": [{"date": "2025-10 to 2025-12", "event": "KOC bulletin describes AI Innovation Center, G Agent and agentic rig scheduling.", "source_ids": ["RS033"]}], "limitations": ["The KOC company publication confirms the named platform and workflow, but gives no measured results or rollout scope.", "The bulletin describes G Agent and agentic scheduling; exact technical architecture and whether they are one system are not clear."], "assessment": "Analyst inference: This is more concrete than a planned AI center: KOC names drilling resource management and AI scheduling. Deployment breadth and performance remain unquantified.", "sources": [{"id": "RS033", "title": "The Kuwait Digest 15 (October–December 2025)", "url": "https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "publisher": "Kuwait Oil Company", "date": "2025-Q4", "type": "company periodical", "locator": "PDF pp. 14–16, AI Innovation Center / G Agent", "review_status": "full text reviewed", "supports": ["KOC says G Agent uses AI for drilling-rig resource management, systems integration and real-time data analysis; it also describes an agentic drilling-rig scheduling project.", "KOC names Microsoft and Halliburton partnership and says Ghaia AI manages technology programs/training at the Center."], "original_id": "koc_digest"}], "field_sources": {"business_problem": ["RS033"], "workflow": ["RS033"], "technology_data": ["RS033"], "integration_human_role": ["RS033"], "deployment_scope": ["RS033"], "economics": ["RS033"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "nocs", "packet_id": "nocs-019", "source_ids": ["RS033"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-085"]}, {"key": "adnoc-microsoft-ai-agents", "existing_ids": ["AI-108"], "related_existing_ids": [], "company": "ADNOC", "title": "ADNOC–Microsoft co-development and deployment of operational AI agents", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Strategy only", "stage": "Strategic agreement announced Nov. 2, 2025; co-development/deployment scope announced, no named agent deployment in the release", "practical_relevance": "The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.", "business_problem": "Extend ADNOC’s existing enterprise AI program into operationally focused agents and autonomous workflows across the value chain.", "workflow": "ADNOC and Microsoft plan to co-develop and deploy AI agents; Microsoft will also provide advanced AI tools and upskilling. The same agreement includes separate Masdar/XRG energy-supply work for Microsoft data-center growth.", "technology_data": "Microsoft AI tools and ADNOC operational data/workflows; detailed agent architecture, integrations and model governance not disclosed.", "integration_human_role": "Autonomous operations are an ambition; human approvals, safety boundaries and production status for the agents are unspecified.", "deployment_scope": "ADNOC value chain-wide agreement; data-center energy work involves Masdar and XRG and is not itself AI adoption.", "partners": ["Microsoft", "Masdar", "XRG"], "disclosed_results": [], "economics": "No deal value, implementation budget, deployed-agent count or AI-specific results disclosed.", "timeline": [{"date": "2025-11-02", "event": "ADNOC and Microsoft announced AI-agent collaboration; Masdar/XRG separately joined the energy-for-AI component.", "source_ids": ["RS030"]}], "limitations": ["Strategic agreement and plans, not evidence of agents already deployed.", "Separate energy supply for Microsoft data centers is infrastructure supporting AI demand, not an ADNOC operational AI workflow.", "Do not merge this partnership with ENERGYai (AI-040) or employee-training record (AI-071)."], "assessment": "Analyst inference: The release clarifies a broad co-development route but offers insufficient named workflow detail to count as a deployed operating system.", "sources": [{"id": "RS030", "title": "ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI", "url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "publisher": "ADNOC", "date": "2025-11-02", "type": "company announcement", "locator": "Operational AI agents and separate energy-for-AI component", "review_status": "full text reviewed", "supports": ["ADNOC and Microsoft announced plans to co-develop and deploy AI agents across ADNOC’s value chain; no particular agent or deployment was named.", "The wider release separately includes Masdar/XRG energy and infrastructure solutions for Microsoft AI/data-center growth."], "original_id": "adnoc_msft_ai"}], "field_sources": {"business_problem": ["RS030"], "workflow": ["RS030"], "technology_data": ["RS030"], "integration_human_role": ["RS030"], "deployment_scope": ["RS030"], "economics": ["RS030"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "nocs", "packet_id": "nocs-020", "source_ids": ["RS030"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-108"]}, {"key": "devon-closed-loop-lift-ai", "existing_ids": ["AI-006"], "company": "Devon Energy", "title": "Closed-loop AI adjusts gas-lift settings at operating wells", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "850+ Delaware wells live as of May 2026; ESP and rod-pump pilots", "practical_relevance": "Closed-loop operating control at a disclosed multi-well scale; management-reported pilot results remain separate from full-rollout economics.", "business_problem": "Well conditions require recurring lift adjustment.", "workflow": "Physics-informed AI calculates gas-lift injection settings and controls wells in a closed loop. Management reported more than 850 Delaware wells operating, with ESP and rod-pump models still in pilot.", "technology_data": "Model, input data and validation are undisclosed.", "integration_human_role": "Control limits and override procedures were not disclosed.", "deployment_scope": "850+ Delaware wells; 1,500 portfolio wells targeted.", "partners": [], "disclosed_results": [{"metric": "Gas-lift pilot production uplift", "value": "approximately 2–3%", "period": "2025 pilot; May 2026 disclosure", "attribution": "Management-reported", "source_ids": ["FU001"]}, {"metric": "Operating gas-lift scope", "value": "850+ Delaware wells", "period": "May 2026", "attribution": "Management-reported", "source_ids": ["FU001"]}], "economics": "The full-rollout uplift and AI-only cash benefit were not quantified.", "timeline": [{"date": "2026-05-06", "event": "Issuer call distinguishes pilot uplift, live gas-lift scope and additional lift pilots.", "source_ids": ["FU001"]}, {"date": "2026-08", "event": "CEO discussed 1,000-well scale and future rollout.", "source_ids": ["RS034"]}], "limitations": ["Pilot uplift is not the result for all 850+ wells. Broader rollout performance was not quantified; 1,500 wells was a target. Exclude the company-wide optimization program from AI-only benefits. Later secondary-hosted disclosures retain their source limits."], "assessment": "Scaled control is established as a management claim; pilot economics cannot be extrapolated mechanically.", "sources": [{"id": "RS034", "title": "Devon Energy Q2 2026 Earnings Call Transcript", "url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "publisher": "Yahoo Finance", "date": "2026-08", "type": "secondary", "locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted", "review_status": "excerpt reviewed", "supports": ["CEO reports 1,000 wells live; >2,000 more were potential rollout."], "accessed": "2026-09-27", "original_id": "na-003"}, {"id": "S02", "title": "Devon Q4 2025 Earnings Q&A Webcast Commentary Transcript", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "publisher": "Devon Energy", "date": "2026-02", "type": "earnings call", "locator": "Existing register cites second-half trials and 2026 rollout; PDF not independently opened in this pass", "review_status": "inaccessible", "supports": ["Existing profile claims second-half 2025 lift trials and Q1–Q2 2026 broader rollout; those details remain unverified in this review."], "accessed": "2026-09-27", "original_id": "na-024"}, {"id": "FU001", "title": "Devon Q1 2026 earnings call — Smart Gas Lift", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "publisher": "Devon Energy", "date": "2026-05-06", "type": "Issuer-hosted transcript", "locator": "Prepared remarks p. 4; Q&A pp. 12–14", "review_status": "full text reviewed", "supports": ["2025 gas-lift pilot: approximately 2–3% production uplift; 850+ Delaware wells live by May 2026 (management-reported)."], "accessed": "2026-09-28"}], "field_sources": {"business_problem": ["RS034"], "workflow": ["FU001"], "technology_data": ["RS034"], "integration_human_role": ["RS034"], "deployment_scope": ["FU001"], "economics": ["FU001"]}, "verification": "verified primary", "corrections_to_existing": "Follow-up on 28 September 2026 reviewed the issuer-hosted Q1 2026 transcript, resolving the earlier secondary-only provenance limitation for the May gas-lift scope and pilot claims. Later secondary updates and the earlier Q4 source retain their separate access limitations. AI-007, AI-086 and AI-087 describe distinct workflows.", "related_existing_ids": ["AI-007", "AI-086", "AI-087"], "cohort": "north_america", "packet_id": "north_america-001", "source_ids": ["RS034", "S02", "FU001"], "report_date": "2026-09-28", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-006"]}, {"key": "devon-chatdvn-enablement", "existing_ids": ["AI-086"], "company": "Devon Energy", "title": "ChatDVN and AI-101 workforce enablement", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "In use; expansion planned", "practical_relevance": "Named internal AI tools and employee adoption are more concrete than generic AI strategy, but productivity estimates have no disclosed measurement method.", "business_problem": "Employees need fast access to company data and assistance with repetitive analysis across technical and support roles.", "workflow": "Internal natural-language data access and assistant use supports everyday engineering, geoscience and software tasks; AI-101 trains leaders and employees on use. Devon planned computer vision/video analytics for anomaly detection.", "technology_data": "ChatDVN is described as combining AI tools, ML and LLMs; Data at Devon enables natural-language information retrieval. Proprietary video analytics compares visual feeds to operating data.", "integration_human_role": "AOGR interview states humans oversee and approve recommendations; company frames AI as augmenting employees.", "deployment_scope": "Companywide training; about 1,000 employees completed AI-101 by June 2025. Plans for video-to-data use in 2026 were prospective.", "partners": ["Internal platform; no external partner named."], "disclosed_results": [{"metric": "Reported productivity improvement among engineers, software developers and geoscientists", "value": "15%-30%", "period": "July 2025 interview", "attribution": "Company-reported", "source_ids": ["R039"]}, {"metric": "AI-101 completions", "value": "about 1,000 employees by June 2025", "period": "June 2025", "attribution": "Company-reported", "source_ids": ["R039"]}], "economics": "No spend, hours saved, quality measures or independent productivity evaluation disclosed.", "timeline": [{"date": "2025-06", "event": "About 1,000 employees had completed AI-101.", "source_ids": ["R039"]}, {"date": "2025-07", "event": "AOGR interview described ChatDVN and video analytics roadmap.", "source_ids": ["R039"]}], "limitations": ["Trade-publication interview rather than issuer filing; claims are attributed to Devon.", "Video anomaly detection was a plan, not a confirmed deployment.", "Historical interview packet: does not establish issuer-verified AI-101 or 15–30% productivity. See watch21extra-ai-086 for the separate primary daily-use evidence."], "assessment": "Named internal AI tools and employee adoption are more concrete than generic AI strategy, but productivity estimates have no disclosed measurement method.", "sources": [{"id": "R039", "title": "Devon Energy Stays Ahead Of Technology Curve", "url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "publisher": "The American Oil & Gas Reporter", "date": "2025-07", "type": "partner case study", "locator": "Interview with CTO Trey Lowe", "review_status": "full text reviewed", "supports": ["ChatDVN internal AI platform; leaders trained via AI-101; engineers, software developers and geoscientists report 15%-30% productivity improvement; humans review/approve AI recommendations; video analytics plans for 2026."], "accessed": "2026-09-27", "original_id": "na-001"}], "field_sources": {"business_problem": ["R039"], "workflow": ["R039"], "technology_data": ["R039"], "integration_human_role": ["R039"], "deployment_scope": ["R039"], "economics": ["R039"]}, "verification": "verified partner / company-attributed interview", "corrections_to_existing": "Not independently reverified against Devon filing; original profile’s productivity and scale claims are attributed claims, not independently established results.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-002", "source_ids": ["R039"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-086"]}, {"key": "eog-field-ai", "existing_ids": ["AI-088"], "company": "EOG Resources", "title": "Generative AI for geology, drilling and safety", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported in 2025–2026 calls; scope and results require issuer transcript confirmation", "practical_relevance": "Reported GenAI use across named functions; primary corroboration outstanding.", "business_problem": "Organize geology information and improve operations.", "workflow": "Transcript reports GenAI for geological data and AI for drilling and safety.", "technology_data": "GenAI is named; models and validation are undisclosed.", "integration_human_role": "Human review is not described in the excerpt.", "deployment_scope": "Geology, drilling and safety; scale undisclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-only outcome verified.", "timeline": [{"date": "2025-11", "event": "Third-party transcript reported generative AI and operational AI applications.", "source_ids": ["R045"]}, {"date": "2026-02", "event": "Third-party transcript reported ML production optimizers and remote Utica operations.", "source_ids": ["R046"]}], "limitations": ["Details are from third-party transcripts; reviewed issuer materials did not confirm these exact claims.", "Unverified does not mean disproved."], "assessment": "Plausible operational lead, still secondary-only pending issuer confirmation.", "sources": [{"id": "R045", "title": "EOG Q3 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2025-11-27", "type": "secondary", "locator": "CEO/management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript reproduces EOG comments on generative AI, geology, drilling, safety, predictive maintenance and in-house applications."], "accessed": "2026-09-27", "original_id": "na-014"}, {"id": "R046", "title": "EOG Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "publisher": "The Motley Fool", "date": "2026-02-25", "type": "secondary", "locator": "Management remarks on production optimizers", "review_status": "excerpt reviewed", "supports": ["Third-party transcript reproduces EOG claims on ML production optimizers and remote Utica operations."], "accessed": "2026-09-27", "original_id": "na-015"}], "field_sources": {"business_problem": ["R045", "R046"], "workflow": ["R045", "R046"], "technology_data": ["R045", "R046"], "integration_human_role": ["R045", "R046"], "deployment_scope": ["R045", "R046"], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-088 covers generative AI across geology, drilling and safety. AI-089 production optimizers/control-room workflow is distinct and is profiled separately.", "related_existing_ids": ["AI-089"], "cohort": "north_america", "packet_id": "north_america-003", "source_ids": ["R045", "R046"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-088"]}, {"key": "conocophillips-enterprise-ai", "existing_ids": ["AI-090"], "company": "ConocoPhillips", "title": "Enterprise AI strategy to embed AI in core processes", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Strategy only", "stage": "Enterprise framework described; named workflows not disclosed", "practical_relevance": "Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.", "business_problem": "Manual effort and inconsistent decision cycles across geographically diverse operations.", "workflow": "CEO said the company is embedding AI in core business processes to reduce manual work, speed decisions and improve consistency.", "technology_data": "No named models, data, platform or partner disclosed.", "integration_human_role": "Enterprise governance, security and responsible-use framework; human roles and specific controls not detailed.", "deployment_scope": "Companywide ambition; no concrete deployment footprint reported.", "partners": [], "disclosed_results": [], "economics": "No spend, savings, productivity or operating metrics disclosed.", "timeline": [{"date": "2026", "event": "CEO described enterprise AI framework at annual meeting.", "source_ids": ["R049"]}], "limitations": ["Disclosure is high-level; no practical named workflow or results identified."], "assessment": "Governance plus stated process objectives indicate intent, but this disclosure belongs in strategy-only rather than a demonstrated workflow register.", "sources": [{"id": "R049", "title": "ConocoPhillips Annual Meeting 2026 Transcript", "url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "publisher": "ConocoPhillips", "date": "2026", "type": "company announcement", "locator": "CEO answer on AI", "review_status": "excerpt reviewed", "supports": ["CEO described focus on embedding AI into core processes to reduce manual work, accelerate decisions and improve consistency under enterprise governance/security/responsible-use framework; no named application or outcome disclosed."], "accessed": "2026-09-27", "publication_date_status": "year only / exact publication date not established in reviewed page", "original_id": "na-006"}], "field_sources": {"business_problem": ["R049"], "workflow": ["R049"], "technology_data": ["R049"], "integration_human_role": ["R049"], "deployment_scope": ["R049"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "AI-090 is correctly characterized as an enterprise-level statement; it does not establish deployed applications or benefits.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-004", "source_ids": ["R049"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-090"]}, {"key": "oxy-enterprise-use-cases", "existing_ids": [], "company": "Occidental", "title": "AI Center of Excellence and cross-functional operating use cases", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "Center of Excellence established; individual workflow maturity not disclosed", "practical_relevance": "A coordination structure and candidate use domains are named; results are not.", "business_problem": "Oxy identified supply-chain, integrity and reservoir-characterization opportunities.", "workflow": "A Center of Excellence coordinates initiatives; specific deployed workflows are not named.", "technology_data": "Models, data and architecture are undisclosed.", "integration_human_role": "Central coordination is reported; review controls are not.", "deployment_scope": "Gulf of America and company functions; scale unquantified.", "partners": [], "disclosed_results": [], "economics": "No AI-specific returns disclosed.", "timeline": [{"date": "2025-02-19", "event": "CEO named AI Center of Excellence and Gulf of America use areas.", "source_ids": ["RS035"]}], "limitations": ["Named application areas are not enough to establish live deployment or results.", "AI is grouped with workforce efficiency but no job or process impact is quantified."], "assessment": "Evidence supports organizational direction, not proven deployment or value.", "sources": [{"id": "RS035", "title": "Q4 2024 Occidental Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf", "publisher": "Occidental", "date": "2025-02-19", "type": "earnings call", "locator": "CEO remarks: AI initiatives and AI Center of Excellence", "review_status": "excerpt reviewed", "supports": ["Oxy named AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization; established an AI Center of Excellence to align cross-company initiatives."], "accessed": "2026-09-27", "original_id": "na-004"}], "field_sources": {"business_problem": ["RS035"], "workflow": ["RS035"], "technology_data": ["RS035"], "integration_human_role": ["RS035"], "deployment_scope": ["RS035"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "AI-044’s subsurface advancement claim is broad; the 2025 earnings call corroborates named AI use areas, not an AI-enhanced subsurface development workflow. AI-091’s workforce-efficiency point is company-reported but general. Neither existing broad profile exactly matches the AI Center of Excellence / Gulf of America disclosure; they are related, not duplicates.", "related_existing_ids": ["AI-044"], "cohort": "north_america", "packet_id": "north_america-005", "source_ids": ["RS035"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-153"]}, {"key": "suncor-ai-mine-operations", "existing_ids": ["AI-099"], "company": "Suncor Energy", "title": "Autonomous haulage fleet expansion at Base Plant (AI not established)", "department": "Production & field operations", "secondary_departments": [], "category": "Automation / AI not established", "stage": "Autonomous haulage operating; undated page snapshot; 140 trucks is a prospective target", "practical_relevance": "Suncor reports an operating autonomous haulage fleet with control-room oversight. The source does not establish AI use; almost 120 is a snapshot and 140 is a prospective target.", "business_problem": "Move oil-sands material safely and consistently across Base Plant mine routes.", "workflow": "Driverless haul trucks move material using GPS, wireless/LTE communications and perceptive technologies. AHS software provides central control-room operators with real-time location and movement visibility so they can manage traffic flow and adapt to driving conditions.", "technology_data": "Suncor names GPS, wireless/LTE, perceptive technologies, AHS software and collision/obstacle detection. It does not identify an AI model, optimization algorithm or input-data architecture.", "integration_human_role": "Central control-room operators monitor real-time truck movement and manage flow; field deployment teams, operators and dispatchers contributed. Human oversight is explicit.", "deployment_scope": "Suncor says all ore at Base Plant is moved by AHS trucks and reports almost 120 trucks across two integrated mines in an undated page snapshot. A target of 140 equipped trucks is prospective; its year is not stated on the page.", "partners": ["Komatsu (haul trucks; software vendor not specified in the reviewed page)"], "disclosed_results": [{"metric": "Autonomous haul trucks", "value": "Almost 120 across two integrated mines in the undated page snapshot; 140 planned", "period": "Undated page snapshot; target year not stated", "attribution": "Company-reported", "source_ids": ["RS036"]}], "economics": "Suncor describes safety, reliability, efficiency and maintenance benefits of the AHS fleet but gives no quantified economic result in this page. These are automation benefits, not AI-attributed returns.", "timeline": [{"date": "2018", "event": "Suncor says Base Plant operations began using autonomous haul trucks.", "source_ids": ["RS036"]}, {"date": "undated", "event": "The page snapshot reports almost 120 trucks across two integrated mines and a prospective target of 140 by an unspecified year-end.", "source_ids": ["RS036"]}], "limitations": ["The page does not mention AI; classify as autonomous automation unless another technical source establishes an AI component.", "The source page is undated; the almost-120 snapshot and 140 target cannot be assigned to a publication year.", "The 140 count is planned, not deployed; the target year is not stated in the reviewed page.", "No quantified economics or workforce/job-impact result is disclosed in this source.", "Suncor's dynamic dispatch workflow is separate and is not included here."], "assessment": "Analyst inference: the disclosure supports a large-scale autonomous mine-transport system with a human control-room response loop. It does not support inclusion in an AI-adoption count without evidence of an AI component.", "sources": [{"id": "RS036", "title": "The Future of Canadian Mining is Autonomous — and Suncor is Leading the Way", "url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "publisher": "Suncor", "date": null, "type": "company feature", "locator": "Base Plant fleet, command centre and end-of-page target statement", "review_status": "full text reviewed", "supports": ["Suncor describes driverless AHS trucks using GPS, wireless/LTE communications and perceptive technologies; its page reports an almost-120 fleet snapshot across two mines, control-room oversight, and a prospective 140-truck target.", "The reviewed page contains no AI/artificial-intelligence claim and supplies no publication date; the target year therefore remains unspecified."], "accessed": "2026-09-27", "publication_date_status": "undated page; accessed 2026-09-27", "original_id": "na-007"}], "field_sources": {"business_problem": ["RS036"], "workflow": ["RS036"], "technology_data": ["RS036"], "integration_human_role": ["RS036"], "deployment_scope": ["RS036"], "economics": ["RS036"]}, "verification": "verified primary", "corrections_to_existing": "Remove the unsupported AI-optimization sentence and avoid inferring a page publication year or dated 140-truck target.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-006", "source_ids": ["RS036"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-099"]}, {"key": "cnrl-governed-enterprise-ai", "existing_ids": ["AI-098"], "company": "Canadian Natural Resources", "title": "Cost-justified AI with production-critical use gated", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Strategy only", "stage": "Policy in place; individual deployments not described", "practical_relevance": "This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.", "business_problem": "Use AI for employee knowledge, data insights and efficiency while controlling cyber, IP and operational risk.", "workflow": "AIF describes incremental, cost-justified adoption and bars reliance on AI in time-sensitive or production-critical processes until controls meet standards.", "technology_data": "No specific AI tools or datasets named.", "integration_human_role": "Employees must participate in creation/review/decisions and remain accountable; AI risk enters ERM and board committee reporting.", "deployment_scope": "Enterprise-wide policy; no application count or operational footprint disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-specific spend or return disclosed.", "timeline": [{"date": "2025-03-25", "event": "AIF described responsible AI program and risk policy.", "source_ids": ["RS037"]}, {"date": "2026", "event": "Company governance page summarized knowledge, data insight and cost-efficiency uses.", "source_ids": ["RS038"]}], "limitations": ["Policy statement does not verify any named AI product or production deployment."], "assessment": "This is a meaningful governance disclosure with a clear restriction on critical process autonomy; it should not be read as evidence of production AI.", "sources": [{"id": "RS037", "title": "Canadian Natural 2025 Annual Information Form", "url": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "publisher": "Canadian Natural Resources", "date": "2026-03-25", "type": "annual report", "locator": "Artificial Intelligence section, p.21", "review_status": "excerpt reviewed", "supports": ["AI is used to expand employee knowledge access, deliver data-driven insight and improve efficiency via incremental, cost-justified deployment; company will not rely on it for time-sensitive/production-critical processes until controls meet standards; employees remain involved/accountable; board risk oversight described."], "accessed": "2026-09-27", "original_id": "na-009"}, {"id": "RS038", "title": "Governance and Risk Management", "url": "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/", "publisher": "Canadian Natural Resources", "date": null, "type": "company announcement", "locator": "Use of AI in operations", "review_status": "excerpt reviewed", "supports": ["Corroborates program-level AI scope; refers readers to AIF for policy specifics."], "accessed": "2026-09-27", "publication_date_status": "undated page; accessed 2026-09-27", "original_id": "na-013"}], "field_sources": {"business_problem": ["RS037", "RS038"], "workflow": ["RS037", "RS038"], "technology_data": ["RS037", "RS038"], "integration_human_role": ["RS037", "RS038"], "deployment_scope": ["RS037", "RS038"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "AI-098 is supported by primary AIF; wording “no AI in production-critical processes” is conditional pending control validation.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-007", "source_ids": ["RS037", "RS038"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-098"]}, {"key": "matador-ai-vendor-fast-follower", "existing_ids": ["AI-096"], "company": "Matador Resources", "title": "AI analytics in production, completions and drilling workflows", "department": "Production & field operations", "secondary_departments": ["Drilling & completions", "Drilling & completions", "IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "Management described active integration of AI-driven analytics across operations in Q1 2026; operating footprint and duration by workflow not quantified.", "practical_relevance": "Management describes real-time monitoring and decision support spanning control-room production data, hydraulic fracturing operations, and lateral placement. The transcript provides practical workflow detail while leaving model architecture, vendor, validation and AI-attributable operating results undisclosed.", "business_problem": "Matador seeks to make operational signals actionable quickly to reduce downtime and identify inefficiencies, while improving completion execution and well placement.", "workflow": "Management says the control room and field staff monitor more than 40 million data points per day in real time to make production signals actionable. For completions, the company monitors hydraulic-fracturing water and sand pressures and volumes plus logistics. For drilling, management says AI assists targeting the preferred portion of the lateral target. The CEO described committee oversight and a measured adoption pace.", "technology_data": "The call names AI-driven analytics and a daily stream of more than 40 million data points, but does not specify model types, vendors, source systems, data governance or validation methods.", "integration_human_role": "Matador describes a cross-functional committee overseeing AI use. Control-room personnel monitor data with field staff; AI is described as assisting targeting and analytics, not controlling drilling or completions autonomously.", "deployment_scope": "Management described use across production, completions and drilling, including real-time control-room/field monitoring and lateral targeting. Counts of wells, rigs, crews or assets covered were not disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-attributed savings, production uplift, cost reduction or return disclosed. The call discusses cycle-time and cost levers but does not attribute those results exclusively to AI.", "timeline": [{"date": "2026-05-07", "event": "Q1 2026 earnings call described AI-driven analytics in production monitoring, completion monitoring and drilling lateral targeting, with a committee-led measured approach.", "source_ids": ["R058"]}], "limitations": ["Earnings call transcript is third-party-hosted, although the statements are attributed to Matador management.", "More than 40 million daily data points is an operating scale statement, not an AI outcome.", "No named vendor, model, asset counts, independent validation or AI-specific economics disclosed.", "Natural-gas transport/sales to AI data centers is infrastructure and remains distinct from internal AI adoption."], "assessment": "Analyst inference: this is more than strategy-only because management describes ongoing, workflow-specific AI use with human operators in the loop. Its practical value is plausible in surfacing anomalies, monitoring completions and improving target placement; deployment maturity and measurable causal impact remain unproven.", "sources": [{"id": "R058", "title": "Matador Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "publisher": "The Motley Fool (transcript of Matador management statements)", "date": "2026-05-07", "type": "earnings call transcript (secondary transcription)", "locator": "Q&A, management response on AI adoption and operations", "review_status": "full text reviewed", "supports": ["Matador executives say a committee is overseeing measured AI adoption and that AI-driven analytics are being integrated across operations.", "The CEO and operations executive describe control-room and field review of >40 million data points per day, real-time monitoring of frac pressures/volumes/logistics, and AI-assisted lateral targeting.", "Management does not provide an AI-specific performance or economic result."], "accessed": "2026-09-27", "original_id": "R058"}], "field_sources": {"business_problem": ["R058"], "workflow": ["R058"], "technology_data": ["R058"], "integration_human_role": ["R058"], "deployment_scope": ["R058"], "economics": ["R058"]}, "verification": "Management statements reviewed in secondary-hosted earnings transcript", "corrections_to_existing": "The prior strategy-only packet said no workflow was named. Q1 2026 management remarks identify production control-room monitoring, completions monitoring, and lateral targeting workflows; no AI-attributable results are established.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-008", "source_ids": ["R058"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-096"]}, {"key": "eqt-procurement-ai", "existing_ids": ["AI-095"], "company": "EQT", "title": "AI-assisted service rebids and procurement targeting", "department": "Supply chain & logistics", "secondary_departments": [], "category": "Internal adoption", "stage": "Tools discussed for prospective rebid targeting; deployment extent undisclosed", "practical_relevance": "Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return.", "business_problem": "Identify service contract rebids and procurement savings opportunities.", "workflow": "Management identified AI tools that could help prioritize service rebids. Broader purchasing activity was discussed separately.", "technology_data": "No tools, supplier-spend data or process design disclosed.", "integration_human_role": "No review/approval workflow disclosed.", "deployment_scope": "No deployment scope or implementation date specified.", "partners": [], "disclosed_results": [{"metric": "Expected procurement cost reduction", "value": "low-single-digit", "period": "management expectation at Q4 2025 call", "attribution": "Management expectation in issuer-hosted transcript; not AI-only", "source_ids": ["EQTD01"]}], "economics": "Prospective procurement opportunity; the call does not specify a unit or spend denominator, achieved savings or an AI-only effect.", "timeline": [{"date": "2026-02-18", "event": "Management discussed AI-supported rebid targeting and procurement focus.", "source_ids": ["R056"]}], "limitations": ["Low-single-digit procurement cost reduction is an opportunity estimate, not a realized saving or an isolated AI effect. Tool, vendor, scope and baseline remain undisclosed."], "assessment": "The use case is specific, but the public statement provides no spend denominator, named system, implementation detail or achieved outcome.", "sources": [{"id": "EQTD01", "title": "EQT Q4 2025 earnings call — AI-assisted procurement targeting", "url": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "publisher": "EQT / FactSet CallStreet", "date": "2026-02-18", "type": "Issuer-hosted corrected earnings-call transcript", "locator": "Printed pp. 20–21; Toby Rice, service rebids and procurement", "review_status": "full text reviewed", "supports": ["Management identifies AI-assisted targeting of service rebids; the procurement opportunity is prospective and not an AI-only return."], "accessed": "2026-09-29"}, {"id": "R056", "title": "EQT Q4 2025 Earnings Call", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2026-02-18", "type": "secondary", "locator": "management remarks on service rebids/procurement", "review_status": "excerpt reviewed", "supports": ["Management said AI tools would help target service rebids and procurement, expecting low-single-digit procurement cost reductions; not isolated to AI and no method, baseline or achieved value disclosed."], "accessed": "2026-09-27", "original_id": "na-012"}], "field_sources": {"business_problem": ["EQTD01"], "workflow": ["EQTD01"], "technology_data": ["EQTD01"], "integration_human_role": ["EQTD01"], "deployment_scope": ["EQTD01"], "economics": ["EQTD01"]}, "verification": "verified primary", "corrections_to_existing": "Issuer-hosted corrected transcript now reviewed. The prior secondary-only access limitation is resolved; implementation and economic attribution remain unknown.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-009", "source_ids": ["EQTD01", "R056"], "report_date": "2026-09-29", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-095"]}, {"key": "eqt-homer-city-ai-load", "existing_ids": ["AI-063"], "company": "EQT", "title": "Homer City gas supply for AI data-centre power demand", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Agreement in principle; proposed power project", "practical_relevance": "Useful exposure to AI electricity demand, but it should not be counted as EQT adopting AI in its own operations.", "business_problem": "Homer City redevelopment requires long-term gas supply to fuel large-scale power generation for AI/HPC loads.", "workflow": "EQT agreed in principle to supply gas to a planned 4.4 GW generation campus; this is commodity supply and generation infrastructure supporting AI demand, not an internal AI workflow.", "technology_data": "No AI system at EQT; project partners include Knighthead, GE Vernova and Kiewit in project roles.", "integration_human_role": "Supply agreement terms and delivery responsibilities require definitive documents.", "deployment_scope": "Homer City, Pennsylvania; target first power 2027 per existing record; current review found issuer/official project document link but did not re-open full terms.", "partners": ["Knighthead; GE Vernova; Kiewit."], "disclosed_results": [], "economics": "Gas volumes/pricing in the agreement in principle were not disclosed in source reviewed; infrastructure economics are distinct from AI adoption.", "timeline": [{"date": "2025-07", "event": "EQT gas-supply agreement in principle announced for Homer City AI/HPC project.", "source_ids": ["S42"]}], "limitations": ["Agreement in principle is not proof of final contract, construction completion or first power.", "This belongs in AI-demand infrastructure."], "assessment": "Useful exposure to AI electricity demand, but it should not be counted as EQT adopting AI in its own operations.", "sources": [{"id": "S42", "title": "Homer City Redevelopment announces agreement in principle for EQT Corporation to supply natural gas", "url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "publisher": "Homer City Redevelopment", "date": "2025-07", "type": "company announcement", "locator": "Agreement in principle", "review_status": "excerpt reviewed", "supports": ["Project announcement describes EQT agreement in principle for gas supply to redevelopment targeting AI/HPC campus; not an executed definitive supply contract."], "accessed": "2026-09-27", "original_id": "na-016"}], "field_sources": {"business_problem": ["S42"], "workflow": ["S42"], "technology_data": ["S42"], "deployment_scope": ["S42"], "integration_human_role": [], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-063 category is correct; primary project source for this existing claim was not available in current search excerpts, so leave source review gap and do not certify the full gas-supply details.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-010", "source_ids": ["S42"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-063"]}, {"key": "cenovus-productivity-pilots", "existing_ids": ["AI-061"], "company": "Cenovus Energy", "title": "Governed productivity and decision-support pilots", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Strategy only", "stage": "Pilot program disclosed; named pilots unavailable", "practical_relevance": "A pilot claim may be valid but should not be expanded into use cases or certified from a bare SEC URL without the supporting text.", "business_problem": "Prior-register claim (not verified in this review): Cenovus identified productivity and decision support as starting points for AI adoption.", "workflow": "Prior-register claim (not verified in this review): the 2025 MD&A was said to describe governed AI pilots; cited text was not retrieved and no workflow is established here.", "technology_data": "No model, data, vendor or platform verified.", "integration_human_role": "Governance/human role details not verified in available source excerpt.", "deployment_scope": "Pilot-stage claim; no users, sites or timing verified.", "partners": [], "disclosed_results": [], "economics": "No AI-specific results or costs verified.", "timeline": [{"date": "2025", "event": "Existing record cites AI productivity/decision-support pilots in 2025 MD&A; claim not re-opened during this review.", "source_ids": ["S40"]}], "limitations": ["No qualifying source text was reopened; keep the profile unverified until MD&A is checked.", "Cenovus public financial and annual report indexes were reviewed, but indexing alone does not prove review of all underlying reports."], "assessment": "A pilot claim may be valid but should not be expanded into use cases or certified from a bare SEC URL without the supporting text.", "sources": [{"id": "S40", "title": "Cenovus 2025 Management’s Discussion and Analysis", "url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "publisher": "Cenovus Energy / SEC", "date": "2026", "type": "annual report", "locator": "Existing register cites MD&A AI disclosure; not reopened", "review_status": "inaccessible", "supports": ["Existing register associates this filing with productivity and decision-support pilots; claim remains unverified in this cohort pass."], "accessed": "2026-09-27", "publication_date_status": "year only / exact publication date not established in reviewed page", "original_id": "na-017"}], "field_sources": {"business_problem": [], "workflow": [], "technology_data": [], "integration_human_role": [], "deployment_scope": [], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-061 is pending source re-verification; no evidence surfaced to contradict it.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-011", "source_ids": ["S40"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-061"]}, {"key": "ovintiv-operational-ai", "existing_ids": ["AI-092"], "company": "Ovintiv", "title": "In-house AI for artificial lift, drilling and completions", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported use cases and scale; primary transcript confirmation outstanding", "practical_relevance": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "business_problem": "Improve production uptime and lift settings, reduce drilling times/costs and tune completions.", "workflow": "CEO described an early AI portfolio in production operations (uptime and artificial-lift optimization), drilling and completions. Component details for DRIVE Center rig optimization and real-time frac design are maintained in separate records.", "technology_data": "In-house algorithms paired with a proprietary data set are reported; architecture, data lineage and model validation not disclosed.", "integration_human_role": "Operations and drilling optimization centers use company algorithms; transcript describes rapid acquired-asset integration. Human review/process not fully described.", "deployment_scope": "Broad functions named; no well or rig counts confirmed in issuer-hosted source review.", "partners": [], "disclosed_results": [], "economics": "No AI-attributable cost or production uplift verified; management discussed combined cost/well results and synergy context without allocation.", "timeline": [{"date": "2025-11", "event": "Q3 transcript described early AI in operations, drilling and completions.", "source_ids": ["R053"]}, {"date": "2026-02", "event": "Q4 transcript described acquired rig integration through DRIVE Center.", "source_ids": ["R054"]}, {"date": "2026-05", "event": "Q1 transcript described real-time tuning of 70-odd frac design inputs.", "source_ids": ["R055"]}], "limitations": ["Available transcript sources are third-party hosted; exact claims need issuer-hosted transcript/audio corroboration.", "Q1 2026 mentions combined operational performance; attribution to AI alone is not established."], "assessment": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "sources": [{"id": "R053", "title": "Ovintiv Q3 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2025-11-06", "type": "secondary", "locator": "Management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "accessed": "2026-09-27", "original_id": "na-018"}, {"id": "R054", "title": "Ovintiv Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2026-02-24", "type": "secondary", "locator": "Management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "accessed": "2026-09-27", "original_id": "na-019"}, {"id": "R055", "title": "Ovintiv Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2026-05-12", "type": "secondary", "locator": "Management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "accessed": "2026-09-27", "original_id": "na-020"}], "field_sources": {"business_problem": ["R053"], "workflow": ["R053"], "technology_data": ["R053"], "integration_human_role": ["R053"], "deployment_scope": ["R053"], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-092 is the broader operational AI portfolio. AI-093 DRIVE Center and AI-094 frac-design tuning are separate workflows linked below.", "related_existing_ids": ["AI-093", "AI-094"], "cohort": "north_america", "packet_id": "north_america-012", "source_ids": ["R053", "R054", "R055"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-092"]}, {"key": "chord-ai-rod-pumps", "existing_ids": ["AI-097"], "company": "Chord Energy", "title": "AI-assisted rod-pump optimization on mature Bakken wells", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported 2026 adjustment; issuer-source confirmation outstanding", "practical_relevance": "The reported use case is plausible and operationally specific, but causal attribution and exact deployment scope are unresolved.", "business_problem": "Improve production from older, declining wells.", "workflow": "Management reportedly adjusted AI rod-pump controls alongside deeper pump settings and added workover rigs; it described a positive base-production effect for the target well group.", "technology_data": "Model/vendor and sensor data not disclosed.", "integration_human_role": "Human role, monitoring and control limits not disclosed.", "deployment_scope": "Targeted legacy wells; counts and production baseline not disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-only decline-rate or production metric verified.", "timeline": [{"date": "2026-05-06", "event": "Third-party Q1 call transcript reports adjustment and effect on base decline.", "source_ids": ["R059"]}], "limitations": ["Secondary transcript only; “halt in decline” cannot be attributed solely to AI given simultaneous workover and pump-depth interventions."], "assessment": "The reported use case is plausible and operationally specific, but causal attribution and exact deployment scope are unresolved.", "sources": [{"id": "R059", "title": "Chord Energy Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "publisher": "The Motley Fool", "date": "2026-05-06", "type": "secondary", "locator": "Management remarks on rod pumps", "review_status": "excerpt reviewed", "supports": ["Transcript attributes improved legacy-well base production to AI pump tuning along with deeper pump placement and workovers; no AI-only effect is isolated."], "accessed": "2026-09-27", "original_id": "na-021"}], "field_sources": {"business_problem": ["R059"], "workflow": ["R059"], "technology_data": ["R059"], "deployment_scope": ["R059"], "integration_human_role": [], "economics": []}, "verification": "secondary only", "corrections_to_existing": "AI-097 rests on an unverified secondary transcript; combined interventions confound the implied AI impact.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-013", "source_ids": ["R059"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-097"]}, {"key": "diamondback-lift-ai", "existing_ids": ["AI-045"], "company": "Diamondback Energy", "title": "Artificial-lift optimization and downtime reduction", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Early/unspecified; secondary source only", "practical_relevance": "A useful lead for follow-up, but present evidence does not support certification of a deployed initiative.", "business_problem": "Prior-register claim (not verified by issuer source): lift optimization was linked to base-production and downtime improvement.", "workflow": "Prior-register claim (not verified): a third-party Q2 2026 transcript was said to mention early AI/automation for daily lift optimization; no primary confirmation was located.", "technology_data": "No model, sensor, vendor or workflow integration details available from reviewed primary sources.", "integration_human_role": "Not disclosed.", "deployment_scope": "Not disclosed.", "partners": [], "disclosed_results": [], "economics": "Any 1% decline figure in transcript context is illustrative rather than an achieved AI outcome.", "timeline": [{"date": "2026-08", "event": "Third-party Q2 2026 transcript is cited by existing record; issuer materials were not reviewed for AI passage.", "source_ids": ["RS040"]}], "limitations": ["No primary-source confirmation located in this cohort review.", "Existing transcript uses illustrative decline-rate discussion; not a realized AI result."], "assessment": "A useful lead for follow-up, but present evidence does not support certification of a deployed initiative.", "sources": [{"id": "RS040", "title": "Diamondback Q2 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "publisher": "The Motley Fool", "date": "2026-08-10", "type": "secondary", "locator": "Existing register citation; not fully re-opened in primary materials", "review_status": "excerpt reviewed", "supports": ["Existing record links discussion of lift optimization and downtime to transcript; use case and outcomes remain unverified by issuer source."], "accessed": "2026-09-27", "original_id": "na-022"}], "field_sources": {"business_problem": [], "workflow": [], "technology_data": [], "integration_human_role": [], "deployment_scope": [], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-045 is only secondary-source supported; do not report scenario math as company performance.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-014", "source_ids": ["RS040"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-045"]}, {"key": "coterra-ml-capital-planning", "existing_ids": ["AI-087"], "company": "Coterra Energy", "title": "ML well prediction informs capital planning", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported by Devon after merger close; pre-merger scale and integration state not fully disclosed", "practical_relevance": "Named capital-planning application; implementation and value remain unclear.", "business_problem": "Improve well forecasts used in capital choices.", "workflow": "CEO said ML well predictions inform capital planning.", "technology_data": "ML named; inputs, validation and architecture undisclosed.", "integration_human_role": "Capital-planning use is reported; review controls are unknown.", "deployment_scope": "Legacy Coterra assets; well coverage is undisclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-specific financial outcome disclosed.", "timeline": [{"date": "2026-08", "event": "CEO said models were being integrated after the merger.", "source_ids": ["RS034"]}], "limitations": ["Third-party-hosted transcript; no realized value disclosed."], "assessment": "Distinct planning use from Devon lift controls; integration is ongoing and secondary-reported.", "sources": [{"id": "RS034", "title": "Devon Energy Q2 2026 Earnings Call Transcript", "url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "publisher": "Yahoo Finance", "date": "2026-08", "type": "secondary", "locator": "CEO remarks on Slide 9 and Q&A; source transcript not issuer-hosted", "review_status": "excerpt reviewed", "supports": ["CEO said Coterra ML models support capital planning."], "accessed": "2026-09-27", "original_id": "na-003"}], "field_sources": {"business_problem": ["RS034"], "workflow": ["RS034"], "technology_data": ["RS034"], "integration_human_role": ["RS034"], "deployment_scope": ["RS034"], "economics": []}, "verification": "secondary only", "corrections_to_existing": "AI-087 exactly matches the Coterra well-prediction/capital-planning workflow. This post-close disclosure is secondary-hosted and needs issuer transcript re-verification.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-015", "source_ids": ["RS034"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-087"]}, {"key": "eog-production-optimizers", "existing_ids": ["AI-089"], "company": "EOG Resources", "title": "ML production optimizers and remote Utica operations", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported in 2025–2026 calls; scope and results require issuer transcript confirmation", "practical_relevance": "Potential production/remote-operations use; issuer verification is outstanding.", "business_problem": "Improve production uptime through optimization and remote oversight.", "workflow": "Transcript reports ML production optimizers and remote management of acquired Utica wells.", "technology_data": "ML named; data and model design undisclosed.", "integration_human_role": "A 24-hour control room is reported; control limits are unclear.", "deployment_scope": "Utica and base production; scale undisclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-only result verified.", "timeline": [{"date": "2026-02-25", "event": "Third-party transcript described ML production optimizers and Utica remote operations.", "source_ids": ["R046"]}], "limitations": ["Claim is from a third-party transcript and was not confirmed in reviewed issuer materials.", "No scale or financial attribution disclosed."], "assessment": "Keep as an unverified lead; scale and realized results remain unknown.", "sources": [{"id": "R046", "title": "EOG Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "publisher": "The Motley Fool", "date": "2026-02-25", "type": "secondary", "locator": "Management remarks on production optimizers", "review_status": "excerpt reviewed", "supports": ["Third-party transcript reproduces EOG claims on ML production optimizers and remote Utica operations."], "accessed": "2026-09-27", "original_id": "na-015"}], "field_sources": {"business_problem": ["R046"], "workflow": ["R046"], "technology_data": ["R046"], "integration_human_role": ["R046"], "deployment_scope": ["R046"], "economics": []}, "verification": "secondary only", "corrections_to_existing": "AI-089 exactly matches this production-optimizer/remote-operations profile; issuer-source verification remains outstanding.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-016", "source_ids": ["R046"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-089"]}, {"key": "ovintiv-drive-center", "existing_ids": ["AI-093"], "company": "Ovintiv", "title": "DRIVE Center applies AI to acquired rigs", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported use cases and scale; primary transcript confirmation outstanding", "practical_relevance": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "business_problem": "Improve drilling/frac performance and operational cost through AI optimization.", "workflow": "Ovintiv’s Q4 2025 third-party transcript says recently acquired drilling rigs were linked to the DRIVE Center, an AI optimization tool, within weeks of closing.", "technology_data": "In-house algorithms paired with a proprietary/private data set; no architecture or named external partners disclosed.", "integration_human_role": "Operations and drilling optimization centers use company algorithms; transcript describes rapid acquired-asset integration. Human review/process not fully described.", "deployment_scope": "Claims span acquired wells/rigs, production operations centers and frac jobs; no well/rig totals verified.", "partners": [], "disclosed_results": [], "economics": "No AI-attributable cost or production uplift verified; management discussed combined cost/well results and synergy context without allocation.", "timeline": [{"date": "2026-02-24", "event": "Ovintiv’s Q4 2025 third-party transcript says recently acquired drilling rigs were linked to the DRIVE Center, an AI optimization tool, within weeks of closing.", "source_ids": ["R054"]}], "limitations": ["Available transcript sources are third-party hosted; exact claims need issuer-hosted transcript/audio corroboration.", "Q1 2026 mentions combined operational performance; attribution to AI alone is not established."], "assessment": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "sources": [{"id": "R054", "title": "Ovintiv Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2026-02-24", "type": "secondary", "locator": "Management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "accessed": "2026-09-27", "original_id": "na-019"}], "field_sources": {"business_problem": ["R054"], "workflow": ["R054"], "technology_data": ["R054"], "integration_human_role": ["R054"], "deployment_scope": ["R054"], "economics": []}, "verification": "secondary only", "corrections_to_existing": "AI-093 exactly matches this workflow. Evidence is a third-party transcript excerpt; primary issuer confirmation remains outstanding.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-017", "source_ids": ["R054"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-093"]}, {"key": "ovintiv-real-time-frac-ai", "existing_ids": ["AI-094"], "company": "Ovintiv", "title": "AI tunes frac-design factors in real time", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Reported use cases and scale; primary transcript confirmation outstanding", "practical_relevance": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "business_problem": "Improve drilling/frac performance and operational cost through AI optimization.", "workflow": "Ovintiv’s Q1 2026 third-party transcript describes in-house algorithms using a proprietary data set to tune roughly 70 frac-design inputs in real time.", "technology_data": "In-house algorithms paired with a proprietary/private data set; no architecture or named external partners disclosed.", "integration_human_role": "Operations and drilling optimization centers use company algorithms; transcript describes rapid acquired-asset integration. Human review/process not fully described.", "deployment_scope": "Claims span acquired wells/rigs, production operations centers and frac jobs; no well/rig totals verified.", "partners": [], "disclosed_results": [], "economics": "No AI-attributable cost or production uplift verified; management discussed combined cost/well results and synergy context without allocation.", "timeline": [{"date": "2026-05-12", "event": "Ovintiv’s Q1 2026 third-party transcript describes in-house algorithms using a proprietary data set to tune roughly 70 frac-design inputs in real time.", "source_ids": ["R055"]}], "limitations": ["Available transcript sources are third-party hosted; exact claims need issuer-hosted transcript/audio corroboration.", "Q1 2026 mentions combined operational performance; attribution to AI alone is not established."], "assessment": "The workflows sound practical and technically specific, but much of the register detail comes from secondary transcripts; primary-source confirmation and measured AI returns remain gaps.", "sources": [{"id": "R055", "title": "Ovintiv Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2026-05-12", "type": "secondary", "locator": "Management remarks", "review_status": "excerpt reviewed", "supports": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "accessed": "2026-09-27", "original_id": "na-020"}], "field_sources": {"business_problem": ["R055"], "workflow": ["R055"], "technology_data": ["R055"], "integration_human_role": ["R055"], "deployment_scope": ["R055"], "economics": []}, "verification": "secondary only", "corrections_to_existing": "AI-094 exactly matches this workflow. Evidence is a third-party transcript excerpt; primary issuer confirmation remains outstanding.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-018", "source_ids": ["R055"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-094"]}, {"key": "devon-dc-monitoring-agents", "existing_ids": ["AI-007"], "company": "Devon Energy", "title": "AI agents for drilling and completions monitoring", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Scaling; issuer reported 1,000 autonomous wells as of Q2 2026", "practical_relevance": "This is the cohort’s most concrete AI disclosure: a named control workflow, current reported scale, and integration into core well operations. Evidence of deployment is company-reported; incremental production, costs and economics remain unisolated.", "business_problem": "Prior-register claim (not reverified): D&C monitoring models may identify unproductive time and repeat effective practices.", "workflow": "Prior-register claim (not reverified): the Q4 2025 presentation was said to describe D&C monitoring models and AI agents; the cited presentation could not be reopened.", "technology_data": "Not disclosed in verified source text.", "integration_human_role": "Not disclosed in verified source text.", "deployment_scope": "Not disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-specific economics verified; broader business optimization savings cannot be attributed to this workflow.", "timeline": [{"date": "2026-02", "event": "Existing register points to Q4 2025 earnings presentation; not reverified.", "source_ids": ["S03"]}], "limitations": ["The cited investor presentation was not readable/reopened during this pass. Treat as unverified, not disproved."], "assessment": "The workflow would be practical if confirmed, but available evidence here is only the existing register citation.", "sources": [{"id": "S03", "title": "Devon Q4 2025 Earnings Presentation", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf", "publisher": "Devon Energy", "date": "2026-02", "type": "investor presentation", "locator": "Existing profile cites D&C monitoring / AI agents; exact slide not reopened", "review_status": "inaccessible", "supports": ["Existing profile says models and agents identify unproductive time and repeat better drilling/completion practices; claim is not reverified in this pass."], "accessed": "2026-09-27", "original_id": "na-023"}], "field_sources": {"business_problem": [], "workflow": [], "technology_data": [], "integration_human_role": [], "deployment_scope": [], "economics": []}, "verification": "unable to verify", "corrections_to_existing": "AI-007 is a distinct D&C workflow and should not be merged into lift optimization. It remains unverified pending review of Devon’s presentation.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-019", "source_ids": ["S03"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-007"]}, {"key": "oxy-subsurface-ai-claim", "existing_ids": ["AI-044"], "company": "Occidental", "title": "AI-enhanced subsurface development", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "Management-reported current AI use in selected Gulf and Permian/onshore work; complex Gulf subsurface project in development in 2025 with execution expected in 1–2 years.", "practical_relevance": "Management named current AI use in operational functions and onshore subsurface work; a more complex Gulf subsurface project remained prospective. No AI-specific results were quantified.", "business_problem": "Apply AI to subsurface characterization/development and operational processes across complex assets.", "workflow": "Q4 2024 remarks named AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization. Q2 2025 remarks said AI already helped subsurface work in the Permian and other onshore areas; separately, a complex Gulf subsurface project was under development with execution expected in one to two years.", "technology_data": "Model, data inputs, software and validation are not disclosed in the reviewed transcript remarks.", "integration_human_role": "Management described an AI center of excellence coordinating initiatives; workflow-level human review and controls are not specified.", "deployment_scope": "Company remarks identify Gulf of America operational areas and Permian/other onshore subsurface work, but provide no asset counts, users or deployment scale. The complex Gulf subsurface project was prospective.", "partners": [], "disclosed_results": [], "economics": "No AI-specific economics or outcome metric disclosed. Do not attribute broader production or reliability results to AI.", "timeline": [{"date": "2025-02-19", "event": "Q4 2024 call described AI use in Gulf of America supply chain, asset integrity and reservoir characterization, and an AI center of excellence; Algeria seismic was described as future-use data.", "source_ids": ["R050"]}, {"date": "2025-08-07", "event": "Q2 2025 call described AI assistance in Permian/other onshore subsurface work and a separate complex Gulf project in development, with execution expected in 1–2 years.", "source_ids": ["R051"]}, {"date": "2026-05-06", "event": "Q1 2026 call described AI as taking a larger role across disciplines; no additional named subsurface workflow.", "source_ids": ["S25"]}], "limitations": ["Source record is based on third-party earnings-call transcripts of management remarks.", "Gulf and onshore use cases are management-reported; technical implementation and scale are not disclosed.", "The complex Gulf subsurface project was forward-looking in Aug. 2025 and must not be described as deployed.", "No AI-attributed result or economics disclosed."], "assessment": "Management remarks support several named uses and existing onshore assistance, alongside a distinct planned Gulf subsurface project. They do not quantify implementation scale or results.", "sources": [{"id": "R050", "title": "Occidental Petroleum Q4 2024 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "publisher": "The Motley Fool (transcription of Occidental management remarks)", "date": "2025-02-19", "type": "earnings-call transcript", "locator": "CEO remarks on Gulf of America and AI center of excellence", "review_status": "full text reviewed", "supports": ["Management named current AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization and an AI center of excellence.", "Algeria seismic acquisition was described as a dataset for future AI use; no outcome or implementation detail was disclosed."], "accessed": "2026-09-27", "original_id": "R050"}, {"id": "R051", "title": "Occidental (OXY) Q2 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "publisher": "The Motley Fool (transcription of Occidental management remarks)", "date": "2025-08-07", "type": "earnings-call transcript", "locator": "CEO remarks on subsurface AI", "review_status": "full text reviewed", "supports": ["CEO said AI was already helping subsurface work in Permian and other onshore areas.", "Complex Gulf subsurface AI project remained under development; execution was expected in the following one to two years."], "accessed": "2026-09-27", "original_id": "R051"}, {"id": "S25", "title": "Occidental Q1 2026 Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "publisher": "Occidental", "date": "2026-05-06", "type": "issuer-hosted earnings-call transcript", "locator": "CEO remarks on AI across disciplines", "review_status": "excerpt reviewed", "supports": ["Management said AI was taking a larger role across disciplines; this is broad context and does not add a named subsurface workflow."], "accessed": "2026-09-27", "original_id": "na-025"}], "field_sources": {"business_problem": ["R050", "R051"], "workflow": ["R050", "R051"], "technology_data": [], "integration_human_role": ["R050"], "deployment_scope": ["R050", "R051"], "economics": ["R050", "R051"]}, "verification": "management-reported transcript evidence reviewed; not independently validated", "corrections_to_existing": "Q4 2024 and Q2 2025 calls support current named uses and onshore assistance; the complex Gulf subsurface project was still prospective. This corrects the earlier Q1-2026-only review framing.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-020", "source_ids": ["R050", "R051", "S25"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-044"]}, {"key": "oxy-workforce-ai", "existing_ids": ["AI-091"], "company": "Occidental", "title": "AI in workforce efficiency and process review", "department": "Workforce & learning", "secondary_departments": [], "category": "Internal adoption", "stage": "Center of Excellence established; individual workflow maturity not disclosed", "practical_relevance": "A stated workforce-efficiency objective, with no named workflow or measured result.", "business_problem": "Management linked AI with workforce efficiency and process review.", "workflow": "COO said AI was expanding across disciplines; no process was specified.", "technology_data": "Model, data and platform are undisclosed.", "integration_human_role": "Review/control process is undisclosed.", "deployment_scope": "No user or asset scope disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-specific savings disclosed.", "timeline": [{"date": "2026-05", "event": "On Q1 2026 call, COO described AI as taking a larger role across disciplines and connected it to workforce efficiency; no named process or result was disclosed.", "source_ids": ["S25"]}], "limitations": ["The claim does not establish a named deployment or measurable effect."], "assessment": "A strategic direction, not a verified workflow or labor saving.", "sources": [{"id": "S25", "title": "Q1 2026 Occidental Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "publisher": "Occidental", "date": "2026-05-06", "type": "earnings call", "locator": "COO answer on workforce efficiency", "review_status": "excerpt reviewed", "supports": ["COO said AI is taking a larger role across company disciplines as part of workforce efficiency; no workflow, deployed scale or AI-only savings quantified."], "accessed": "2026-09-27", "original_id": "na-005"}], "field_sources": {"business_problem": ["S25"], "workflow": ["S25"], "technology_data": [], "integration_human_role": [], "deployment_scope": [], "economics": []}, "verification": "verified primary, strategy-level", "corrections_to_existing": "AI-091 exactly matches the workforce-efficiency point. No named process, adoption scale or AI-specific productivity outcome was disclosed.", "related_existing_ids": [], "cohort": "north_america", "packet_id": "north_america-021", "source_ids": ["S25"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-091"]}, {"key": "suncor-dynamic-dispatch-ai", "existing_ids": [], "company": "Suncor Energy", "title": "Dynamic AI dispatch assigns mine haul trucks", "department": "IT, data & enterprise transformation", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "Operating system described on an undated company page; initial deployment date not disclosed.", "practical_relevance": "A named AI system automates routine truck assignments while keeping dispatchers active to improve route and ore-blend decisions.", "business_problem": "Mine dispatchers must continuously route haul trucks among shovels, dumps and refueling while responding to slowdowns.", "workflow": "Suncor says dynamic dispatch assigns trucks to dump stations, sends them to refueling and reroutes around slowdowns; dispatchers then review mine maps and optimize routes and blends.", "technology_data": "Suncor identifies an AI dispatch system; model, optimization method and input data are not disclosed.", "integration_human_role": "Human dispatchers remain on shift, monitor routes and use the system’s assignments to improve mine routing and blends.", "deployment_scope": "Suncor describes implementation at Syncrude’s Mildred Lake site; truck count, start date and broader site coverage are not disclosed.", "partners": [], "disclosed_results": [{"metric": "Routine dispatch functions", "value": "AI assigns dumps/refueling and reroutes around slowdowns; human dispatchers retained", "period": "undated company page", "attribution": "Company-reported", "source_ids": ["RS041"]}], "economics": "No dispatch-specific production, cost or revenue outcome is quantified.", "timeline": [{"date": "undated", "event": "Company page describes dynamic dispatch as recently implemented; publication and initial deployment dates are not stated.", "source_ids": ["RS041"]}], "limitations": ["Source page is undated; “recently implemented” has no dated reference point.", "No named vendor, system architecture, AI method, user count or quantitative outcome disclosed.", "This workflow is distinct from the autonomous haulage fleet in AI-099."], "assessment": "This is a practical AI decision-support/assignment workflow, not evidence that haulage fleet autonomy and dispatch are one system. Human dispatchers remain part of operations.", "sources": [{"id": "RS041", "title": "Dispatching innovation: Suncor's new dispatch system improves mine operations", "url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "publisher": "Suncor", "date": null, "type": "company announcement", "locator": "AI dispatch system; human dispatcher workflow", "review_status": "full text reviewed", "supports": ["Dynamic dispatch automates dump/refueling assignments and slowdown rerouting; human dispatchers remain active."], "accessed": "2026-09-27", "publication_date_status": "undated page; accessed 2026-09-27", "original_id": "na-008"}], "field_sources": {"business_problem": ["RS041"], "workflow": ["RS041"], "technology_data": ["RS041"], "integration_human_role": ["RS041"], "deployment_scope": ["RS041"], "economics": ["RS041"]}, "verification": "verified primary", "corrections_to_existing": "New distinct workflow identified in Suncor source; related to, but not the same initiative as, AI-099.", "related_existing_ids": ["AI-099"], "cohort": "north_america", "packet_id": "north_america-022", "source_ids": ["RS041"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-154"]}, {"key": "slb-tela-catalog", "existing_ids": ["AI-008"], "related_existing_ids": ["AI-009"], "company": "SLB", "title": "Tela agentic-AI product catalog and workflow capability", "department": "Drilling & completions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Tela launched Nov 2025; current catalog advertises 33 ready workflows", "practical_relevance": "Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.", "business_problem": "Provide upstream teams with domain-aware assistance inside engineering and operations tools for technical interpretation and repetitive workflow execution.", "workflow": "Catalog covers data discovery, reservoir/simulation queries, document ingestion for well planning, equipment recommendations, production forecasting/surveillance and other domains. SLB says some embedded agents recommend actions while more mature autonomous workflows can act within physics/domain guardrails.", "technology_data": "Lumi data/AI platform; LLMs plus domain foundation models; deployed cloud or on-prem. Catalog currently covers in-app Tela, Canvas and Innovation Factori workflows.", "integration_human_role": "Product documentation describes human oversight and supports recommendations; autonomy is configurable/guardrailed. No claim every agent is autonomous.", "deployment_scope": "Commercial catalog offers 33 workflows; customer installation per workflow not established.", "partners": ["SLB internal product and Innovation Factori teams"], "disclosed_results": [], "economics": "No disclosed sales, usage, or quantified customer return for the catalog itself.", "timeline": [{"date": "2025-11-03", "event": "Tela launch announced", "source_ids": ["S45"]}, {"date": "2026-09-27", "event": "Current product page lists 33 deployment-ready workflows; access date, not publication date", "source_ids": ["RS042"]}], "limitations": ["Product page is undated and mutable; accessed 2026-09-27. The current 33-workflow count may postdate original Nov 2025 launch."], "assessment": "Analyst inference: Substantial named technical functionality makes Tela a concrete product offer. The 33 count is a deployment-ready catalog count, not 33 deployments or validated outcomes.", "sources": [{"id": "RS042", "title": "Tela agentic-AI assistant catalog", "url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "publisher": "SLB", "date": "Undated; accessed 2026-09-27", "type": "company product documentation", "locator": "Tela feature and 33 workflow brochure sections", "review_status": "full text reviewed", "supports": ["SLB lists 33 deployment-ready workflows across subsurface, drilling, production, facilities and sustainability.", "Tela uses LLMs and domain foundation models on Lumi and is offered within SLB applications/cloud or on premises.", "The catalog is product capability and availability; individual workflows are not all evidence of customer deployment."], "accessed_date": "2026-09-27", "original_id": "slb_tela"}, {"id": "S45", "title": "SLB Unveils Groundbreaking New Agentic AI Technology for the Energy Industry", "url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "publisher": "SLB", "date": "2025-11-03", "type": "company announcement", "locator": "Tela architecture and deployment availability", "review_status": "full text reviewed", "supports": ["Tela announced as agentic AI assistant built on Lumi and using LLMs and domain foundation models.", "SLB described a five-step loop and says assistants can interpret logs, predict drilling issues and optimize equipment; autonomous actions depend on workflow and are not customer deployment evidence."], "original_id": "slb_tela_launch"}], "field_sources": {"business_problem": ["RS042", "S45"], "workflow": ["RS042", "S45"], "technology_data": ["RS042", "S45"], "integration_human_role": ["RS042", "S45"], "deployment_scope": ["RS042", "S45"], "economics": ["RS042", "S45"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-001", "source_ids": ["RS042", "S45"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-008"]}, {"key": "slb-marketplace", "existing_ids": ["AI-047"], "related_existing_ids": [], "company": "SLB", "title": "Digital Marketplace for energy AI agents and applications", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched June 2026", "practical_relevance": "A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.", "business_problem": "Help energy companies discover, evaluate and deploy specialized digital/AI applications in existing digital environments.", "workflow": "Governed store lists SLB and partner agents, models, skills, connectors and apps spanning Tela, Lumi and Delfi. Offers are vetted for security, interoperability and compatibility.", "technology_data": "Marketplace links into SLB Lumi/Delfi platforms and Tela AI assistant; detailed technical interface architecture not disclosed.", "integration_human_role": "Users evaluate/access tools; publication and certification controls provide pre-listing governance.", "deployment_scope": "Launch inventory ~200 products, including >30 partners; listing does not confirm downstream deployment.", "partners": ["Over 30 partners and ISVs"], "disclosed_results": [], "economics": "Marketplace monetization model not quantified; SLB describes sales and consumption-scale ecosystem.", "timeline": [{"date": "2026-06-15", "event": "Marketplace launched", "source_ids": ["RS043"]}], "limitations": ["Company announcement reports marketplace inventory, not paid adoption or customer operating results."], "assessment": "Analyst inference: A distribution/product ecosystem can shorten procurement and integration paths. Value depends on customer adoption and interoperability after listing; product availability is not deployment.", "sources": [{"id": "RS043", "title": "SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy", "url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "publisher": "SLB", "date": "2026-06-15", "type": "company announcement", "locator": "Marketplace offerings and governance", "review_status": "full text reviewed", "supports": ["Marketplace launched as channel for SLB and third parties to offer agents, domain models, tools, connectors and applications within existing environments.", "Approximately 200 products from SLB and 30+ partners; offers are certified for security, interoperability and compatibility before listing.", "Availability of products does not establish end-customer installation or operational results."], "original_id": "slb_market"}], "field_sources": {"business_problem": ["RS043"], "workflow": ["RS043"], "technology_data": ["RS043"], "integration_human_role": ["RS043"], "deployment_scope": ["RS043"], "economics": ["RS043"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-002", "source_ids": ["RS043"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-047"]}, {"key": "slb-petoro", "existing_ids": [], "related_existing_ids": [], "company": "Petoro", "title": "Petoro AI well-data extraction and field-development screening", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Commercial AI product", "stage": "Pilot/prototype workflow and customer validation reported June 2026", "practical_relevance": "A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.", "business_problem": "Review a large, fragmented set of well records and sidetrack candidates to improve field-development choices in mature Norwegian fields.", "workflow": "Well Schematic Extractor/Well History agents extract casing, completion and formation facts from reports/logs; deterministic checks plus visual QC help staff verify output. FDPlan and agent workflows aid opportunity selection and drilling portfolio optimization.", "technology_data": "Lumi AI workspace, LLMs, domain context, FDPlan, Delfi and OSDU-aligned data foundation.", "integration_human_role": "Petoro analysts validate extraction via visual QC and use insights for plans; user decisions remain with engineers/planners.", "deployment_scope": "Petoro NCS assets; field/schematic coverage is not quantified.", "partners": ["SLB"], "disclosed_results": [{"metric": "well schematic quality-control throughput", "value": "2 to 6-7 schematics/day", "period": "preliminary user testing reported Jun 2026", "attribution": "Vendor-reported", "source_ids": ["RS044"]}, {"metric": "productivity improvement", "value": "3x", "period": "2026 case study", "attribution": "Vendor-reported", "source_ids": ["RS044"]}], "economics": "SLB reports 3x productivity; preliminary QC throughput went from 2 schematics/day to 6-7/day. Vendor case-study figures; no economics disclosed.", "timeline": [{"date": "2025-10-17", "event": "SLB Q3 earnings says Petoro awarded generative AI field-planning solution", "source_ids": ["RS045"]}, {"date": "2026-06-05", "event": "Detailed Petoro well-data extraction and preliminary productivity published", "source_ids": ["RS044"]}], "limitations": ["Vendor-authored case study; Petoro 2025 report separately corroborates reservoir-complexity tool in Delfi, but not the specific GenAI schematic agent or productivity numbers."], "assessment": "Analyst inference: A named customer and two live data workflows support practical value beyond product claims. Reported productivity needs Petoro corroboration and broader operational rollout evidence.", "sources": [{"id": "RS044", "title": "Petoro and SLB: Pioneering AI-driven well planning on the Norwegian continental shelf", "url": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "publisher": "SLB", "date": "2026-06-05", "type": "partner case study", "locator": "Objectives, workflow, impact and deployment sections", "review_status": "full text reviewed", "supports": ["Petoro and SLB built Lumi/Innovation Factori workflows to extract data from well schematics/reports and rank field-development/sidetrack opportunities.", "Well Schematic Extractor and Well History use LLMs to extract casing, completion and formation information, with deterministic checks and user visual QC.", "SLB reports threefold productivity improvement; preliminary QC rose from two schematics/day to six or seven. This is vendor case-study reporting."], "original_id": "petoro"}, {"id": "RS045", "title": "SLB Announces Third-Quarter 2025 Results", "url": "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609", "publisher": "SLB", "date": "2025-10-17", "type": "earnings release", "locator": "Digital highlights / Petoro award", "review_status": "excerpt reviewed", "supports": ["Q3 earnings: Petoro awarded a generative-AI solution to optimize drilling resources using FDPlan and Lumi."], "original_id": "slb_q3"}], "field_sources": {"business_problem": ["RS044", "RS045"], "workflow": ["RS044", "RS045"], "technology_data": ["RS044", "RS045"], "integration_human_role": ["RS044"], "deployment_scope": ["RS044", "RS045"], "economics": ["RS044"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-003", "source_ids": ["RS044", "RS045"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-155"]}, {"key": "slb-bp-azerbaijan", "existing_ids": ["AI-012"], "related_existing_ids": [], "company": "bp", "title": "Remote offshore drilling operations for bp Azerbaijan rigs", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployed on multiple rigs; result reported in Jan 2026 earnings", "practical_relevance": "Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.", "business_problem": "Reduce offshore staffing and oversight burden.", "workflow": "Performance Live remotely supports offshore platforms with DrillOps/Neuro; Neuro automates trajectory delivery and downlinking.", "technology_data": "DrillOps models formation/drilling parameters; Neuro controls trajectory. Detailed model architecture undisclosed.", "integration_human_role": "Remote center specialists and cross-trained offshore drillers work together; override details unspecified.", "deployment_scope": "Several Western Asia offshore platforms; operator not named in SLB case study.", "partners": ["SLB"], "disclosed_results": [{"metric": "offshore well-placement crew size", "value": "reduced 66%", "period": "Q4 2025 results", "attribution": "Vendor-reported", "source_ids": ["RS047"]}, {"metric": "personnel-on-board days saved", "value": ">1,500", "period": "Q4 2025 results", "attribution": "Vendor-reported", "source_ids": ["RS047"]}], "economics": "SLB reports crew −66% and >1,500 offshore personnel-days saved; financial savings not stated.", "timeline": [{"date": "2026-01-16", "event": "SLB disclosed Azerbaijan remote drilling results in Q4 earnings release", "source_ids": ["RS046"]}], "limitations": ["Vendor earnings disclosure; no operator corroboration reviewed."], "assessment": "Analyst inference: Named customer and quantified offshore workforce result indicate an operationally material deployment, though metric is supplier-reported and baseline scope is unclear.", "sources": [{"id": "RS046", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "publisher": "SLB", "date": "2026-01-23", "type": "earnings release", "locator": "Technology highlights, pp. 18-19", "review_status": "full text reviewed", "supports": ["Q4 release also records the bp partnership update."], "original_id": "slb_q4"}, {"id": "RS047", "title": "SLB digital solutions save 1,500 personnel-on-board days and boost efficiency", "url": "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "publisher": "SLB", "date": "Undated; accessed 2026-09-27", "type": "company case study", "locator": "Summary, objective and results; excerpt reviewed", "review_status": "excerpt reviewed", "supports": ["Performance Live with DrillOps and Neuro reduced offshore well-placement crew size 66% and saved >1,500 POB days across several Western Asia platforms.", "Case attributes benefits to remote operations and autonomous directional drilling; exact operator not identified in the case page."], "accessed_date": "2026-09-27", "original_id": "bp_azerbaijan_case"}], "field_sources": {"business_problem": ["RS047"], "workflow": ["RS047"], "technology_data": ["RS047"], "integration_human_role": ["RS047"], "deployment_scope": ["RS047"], "economics": ["RS047"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-004", "source_ids": ["RS046", "RS047"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-012"]}, {"key": "slb-oq-neuro", "existing_ids": ["AI-013"], "related_existing_ids": [], "company": "OQ Exploration & Production", "title": "Neuro autonomous directional drilling on OQ wells", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Drilled four reported wells; results published Jan 2026", "practical_relevance": "This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.", "business_problem": "Improve drilling speed in Oman.", "workflow": "Neuro/PowerDrive results on three wells; PowerDrive One on another.", "technology_data": "Neuro, RSS; model details absent.", "integration_human_role": "Supervision details absent.", "deployment_scope": "Four wells.", "partners": ["SLB"], "disclosed_results": [{"metric": "ROP increase", "value": "up to 50% on three wells", "period": "reported Jan 2026", "attribution": "Vendor-reported", "source_ids": ["RS046"]}, {"metric": "ROP increase", "value": "40% on one well", "period": "reported Jan 2026", "attribution": "Vendor-reported", "source_ids": ["RS046"]}], "economics": "SLB: ROP up to +50% on three, +40% on one; less tripping/casing, no cost.", "timeline": [{"date": "2025", "event": "OQ Neuro and PowerDrive well campaigns conducted; exact well dates not disclosed", "source_ids": ["RS046"]}, {"date": "2026-01-16", "event": "Results disclosed by SLB", "source_ids": ["RS046"]}], "limitations": ["Primary vendor earnings evidence; operator-side corroboration and well-by-well denominators are not disclosed."], "assessment": "Analyst inference: This is a named well-level deployment with concrete performance figures. Gains are vendor-reported and well-specific, so should not be generalized to OQ’s full program.", "sources": [{"id": "RS046", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "publisher": "SLB", "date": "2026-01-23", "type": "earnings release", "locator": "Technology highlights, pp. 18-19", "review_status": "full text reviewed", "supports": ["Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile."], "original_id": "slb_q4"}], "field_sources": {"business_problem": ["RS046"], "workflow": ["RS046"], "technology_data": ["RS046"], "integration_human_role": ["RS046"], "deployment_scope": ["RS046"], "economics": ["RS046"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-005", "source_ids": ["RS046"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-013"]}, {"key": "slb-koc-seismic", "existing_ids": ["AI-014"], "related_existing_ids": [], "company": "Kuwait Oil Company", "title": "AI-assisted seismic processing contract for Kuwait Oil Company", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Five-year award reported Jan 2026", "practical_relevance": "The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.", "business_problem": "Process KOC seismic data.", "workflow": "Five-year contract includes AI reprocessing.", "technology_data": "AI workflow; methods absent.", "integration_human_role": "Not described.", "deployment_scope": "KOC; coverage absent.", "partners": ["SLB"], "disclosed_results": [], "economics": "Value/outcomes undisclosed.", "timeline": [{"date": "2025-12-31", "event": "Five-year seismic AI processing/reprocessing contract awarded (reported in Q4 2025 results)", "source_ids": ["RS046"]}], "limitations": ["Vendor-reported contract; no KOC source or specific AI performance metric reviewed."], "assessment": "Analyst inference: The award is a commitment to apply AI workflows in seismic interpretation, but the press disclosure does not describe delivered output or deployment milestones.", "sources": [{"id": "RS046", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "publisher": "SLB", "date": "2026-01-23", "type": "earnings release", "locator": "Technology highlights, pp. 18-19", "review_status": "full text reviewed", "supports": ["Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile."], "original_id": "slb_q4"}], "field_sources": {"business_problem": ["RS046"], "workflow": ["RS046"], "technology_data": ["RS046"], "integration_human_role": ["RS046"], "deployment_scope": ["RS046"], "economics": ["RS046"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-006", "source_ids": ["RS046"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-014"]}, {"key": "slb-koc-jurassic", "existing_ids": ["AI-066"], "related_existing_ids": [], "company": "Kuwait Oil Company", "title": "Four AI collaborations for KOC Jurassic gas field", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Contract awards reported July 2026; implementation plans, not verified completed rollout", "practical_relevance": "The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.", "business_problem": "Improve Jurassic-gas decisions and KOC AI skills.", "workflow": "Four awards: AI rig scheduling, GenAI production analytics, log QC/reconstruction, 3-month academy.", "technology_data": "Models and interfaces undisclosed.", "integration_human_role": "Academy trains KOC experts.", "deployment_scope": "Jurassic gas; breadth absent.", "partners": ["SLB"], "disclosed_results": [], "economics": "Value/outcomes absent.", "timeline": [{"date": "2026-07-24", "event": "Four AI collaborations disclosed in Q2 earnings", "source_ids": ["S27"]}], "limitations": ["SLB partner disclosure only. Distinguish awarded projects from completed applications."], "assessment": "Analyst inference: The earnings release identifies four named workstreams and training, improving materially on a generic collaboration label. Evidence shows awards and intended work, not delivered operational outcomes.", "sources": [{"id": "S27", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "earnings release", "locator": "Quarterly highlights, contract awards and technology highlights", "review_status": "full text reviewed", "supports": ["Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy)."], "original_id": "slb_q2"}], "field_sources": {"business_problem": ["S27"], "workflow": ["S27"], "technology_data": ["S27"], "integration_human_role": ["S27"], "deployment_scope": ["S27"], "economics": ["S27"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-007", "source_ids": ["S27"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-066"]}, {"key": "slb-ongc-drillops", "existing_ids": [], "related_existing_ids": [], "company": "ONGC", "title": "DrillOps intelligent well delivery across ONGC rig fleet", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Three-year enterprise contract; rollout scope up to 120 rigs announced July 2026", "practical_relevance": "This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.", "business_problem": "Coordinate large drilling fleet.", "workflow": "DrillOps monitoring, predictive analytics, central review.", "technology_data": "Models/sensors absent.", "integration_human_role": "Central review; no autonomous control claim.", "deployment_scope": "Three years; up to 120 rigs.", "partners": ["SLB"], "disclosed_results": [], "economics": "Value/results absent.", "timeline": [{"date": "2026-07-24", "event": "Three-year contract announced; up to 120 rigs in scope", "source_ids": ["S27"]}], "limitations": ["Vendor-announced contract; no ONGC announcement or measured outcomes found. AI contribution is not explicit in the deployment description."], "assessment": "Analyst inference: This is a named, broad-scale digital deployment with useful scope detail. Avoid describing it as AI-specific since the source emphasizes real-time monitoring/predictive analytics without naming AI in this paragraph.", "sources": [{"id": "S27", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "earnings release", "locator": "Quarterly highlights, contract awards and technology highlights", "review_status": "full text reviewed", "supports": ["Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs."], "original_id": "slb_q2"}], "field_sources": {"business_problem": ["S27"], "workflow": ["S27"], "technology_data": ["S27"], "integration_human_role": ["S27"], "deployment_scope": ["S27"], "economics": ["S27"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-008", "source_ids": ["S27"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-156"]}, {"key": "slb-pttep-pilot", "existing_ids": ["AI-067"], "related_existing_ids": [], "company": "PTTEP", "title": "PTTEP offshore production data and AI advisory pilot", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Pilot award reported July 2026; not yet evidenced as completed", "practical_relevance": "A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.", "business_problem": "Improve offshore production assurance.", "workflow": "Pilot: Cognite Data Fusion, OptiFlow, Tela advisory agents.", "technology_data": "Models unspecified.", "integration_human_role": "Advisory; approvals absent.", "deployment_scope": "One offshore asset.", "partners": ["SLB", "Cognite"], "disclosed_results": [], "economics": "Value/results absent.", "timeline": [{"date": "2026-07-24", "event": "Pilot contract disclosed by SLB", "source_ids": ["S27"]}], "limitations": ["Partner announcement only; no PTTEP confirmation or pilot results reviewed."], "assessment": "Analyst inference: A concrete asset pilot links data foundation and AI advisory to production workflows; evidence is only contract/pilot stage and gives no outcomes.", "sources": [{"id": "S27", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "earnings release", "locator": "Quarterly highlights, contract awards and technology highlights", "review_status": "full text reviewed", "supports": ["Q2 disclosure: PTTEP one-asset offshore pilot using Cognite, OptiFlow and Tela agents."], "original_id": "slb_q2"}], "field_sources": {"business_problem": ["S27"], "workflow": ["S27"], "technology_data": ["S27"], "integration_human_role": ["S27"], "deployment_scope": ["S27"], "economics": ["S27"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-009", "source_ids": ["S27"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-067"]}, {"key": "slb-cnooc-terminal", "existing_ids": ["AI-068"], "related_existing_ids": [], "company": "CNOOC", "title": "Closed-loop process optimization at CNOOC Bohai offshore terminal", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Integrated solution reported in July 2026; live closed-loop operation described", "practical_relevance": "This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.", "business_problem": "Optimize terminal process economics.", "workflow": "Symmetry twins link plant data to APC/DCS closed-loop optimization.", "technology_data": "AI, simulation, APC/DCS.", "integration_human_role": "Oversight unspecified.", "deployment_scope": "One CNOOC subsidiary terminal.", "partners": ["SLB", "China France Bohai Geoservices"], "disclosed_results": [], "economics": "No quantified benefit.", "timeline": [{"date": "2026-07-24", "event": "Deployment described in SLB Q2 results", "source_ids": ["S27"]}], "limitations": ["SLB vendor disclosure; no operating metric or CNOOC corroboration reviewed."], "assessment": "Analyst inference: This is among the most technically concrete AI-service cases: data, digital twin, APC/DCS and closed-loop control are named. Actual value and control safeguards are not disclosed.", "sources": [{"id": "S27", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "earnings release", "locator": "Quarterly highlights, contract awards and technology highlights", "review_status": "full text reviewed", "supports": ["Q2 disclosure: CNOOC subsidiary terminal connects Symmetry twins to APC/DCS."], "original_id": "slb_q2"}], "field_sources": {"business_problem": ["S27"], "workflow": ["S27"], "technology_data": ["S27"], "integration_human_role": ["S27"], "deployment_scope": ["S27"], "economics": ["S27"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-010", "source_ids": ["S27"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-068"]}, {"key": "halliburton-coterra-autofrac", "existing_ids": ["AI-109"], "related_existing_ids": [], "company": "Halliburton", "title": "Octiv Auto Frac autonomous stage execution with Coterra", "department": "Drilling & completions", "secondary_departments": [], "category": "Automation / AI not established", "stage": "Initial launch Jan 2025; rollout across remaining Halliburton-executed Permian programs", "practical_relevance": "A named customer rolled a real automated completion workflow beyond an initial pilot. This should be categorized as automation, not AI, because the source does not establish an AI algorithm.", "business_problem": "Execute customer-designed hydraulic-fracturing stages more consistently and reduce manual variability during pumping.", "workflow": "Coterra configures design and job controls; Octiv Auto Frac automatically executes stage delivery via ZEUS, using Sensori for monitoring. Initial rollout expanded to remaining Halliburton-run Coterra Permian completion programs.", "technology_data": "Automated control system, equipment telemetry and fracture monitoring; autonomy is process automation, not a claim of AI.", "integration_human_role": "Operator retains design configuration/control. System executes stage-by-stage after customer setup; human oversight specifics not disclosed.", "deployment_scope": "Coterra’s Halliburton-executed Permian completions.", "partners": ["Coterra Energy"], "disclosed_results": [{"metric": "stage efficiency increase", "value": "17%", "period": "initial deployment, announced Jan 2025", "attribution": "Vendor-reported", "source_ids": ["RS048"]}], "economics": "Halliburton reports 17% higher stage efficiency in initial deployment. No absolute throughput/cost definition or independent validation.", "timeline": [{"date": "2025-01-06", "event": "Coterra launch and rollout to remaining Halliburton-executed Permian programs announced", "source_ids": ["RS048"]}], "limitations": ["Stage-efficiency baseline/definition not given; all metrics are Halliburton-reported.", "The cited sources identify autonomous/intelligent automation but do not disclose an AI/ML method; treat as automation, not AI."], "assessment": "Analyst inference: This is an operationally relevant autonomous frac workflow with a reported stage-efficiency result. The reviewed sources do not establish AI/ML, so it is classified as automation.", "sources": [{"id": "RS048", "title": "Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program", "url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "publisher": "Halliburton", "date": "2025-01-06", "type": "company announcement", "locator": "Deployment and results paragraphs", "review_status": "full text reviewed", "supports": ["Octiv Auto Frac automates stage delivery against customer-configured designs, using ZEUS and Sensori.", "Coterra was first operator to automate and control its frac design/execution; Halliburton reports 17% higher stage efficiency in initial rollout and rollout to remaining Halliburton-executed Permian completion programs."], "original_id": "RS048"}], "field_sources": {"business_problem": ["RS048"], "workflow": ["RS048"], "technology_data": ["RS048"], "integration_human_role": ["RS048"], "deployment_scope": ["RS048"], "economics": ["RS048"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-011", "source_ids": ["RS048"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-109"]}, {"key": "halliburton-chevron-closed-loop", "existing_ids": ["AI-110"], "related_existing_ids": [], "company": "Halliburton", "title": "Closed-loop hydraulic fracturing with Chevron", "department": "Drilling & completions", "secondary_departments": [], "category": "Automation / AI not established", "stage": "Closed-loop autonomous operation achieved in Colorado by Q1 2025; platform described Apr 2025", "practical_relevance": "A significant move from preset automation to feedback-driven closed-loop completions. Since the partner calls the capability intelligent fracturing but does not establish AI, do not classify as AI.", "business_problem": "Adapt frac delivery to local subsurface feedback while treatment is underway.", "workflow": "ZEUS IQ couples automated OCTIV stage execution with Sensori fracture feedback; Chevron’s subsurface knowledge is embedded in an algorithm to adapt operations in real time.", "technology_data": "Wellsite-to-cloud loop, subsurface measurements and automated stage controls; AI not expressly claimed in source, so classify as digital feedback automation.", "integration_human_role": "No-human-intervention operation is reported; human design/oversight details and safety stop behavior are not described.", "deployment_scope": "Chevron Colorado completions; reported operation count not disclosed.", "partners": ["Chevron"], "disclosed_results": [], "economics": "No quantified productivity or recovery metric disclosed.", "timeline": [{"date": "2025-04-22", "event": "Halliburton says first closed-loop autonomous frac operation achieved during Q1 2025", "source_ids": ["RS049"]}, {"date": "2025-04-12", "event": "Chevron/Halliburton workflow described publicly", "source_ids": ["RS049"]}], "limitations": ["Deployment evidence comes from Halliburton/partner release; no measured performance result. Automation, not verified AI.", "The reviewed Halliburton source describes feedback-driven automation and a domain algorithm, but does not identify an AI/ML method."], "assessment": "Analyst inference: Feedback-driven completions are a meaningful control capability; on reviewed evidence this is closed-loop automation rather than established AI.", "sources": [{"id": "RS049", "title": "Chevron and Halliburton Enable Intelligent Hydraulic Fracturing", "url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "publisher": "Halliburton", "date": "2025-04-12", "type": "company announcement", "locator": "Closed-loop intelligent fracturing description", "review_status": "full text reviewed", "supports": ["ZEUS IQ combines OCTIV Auto Frac execution, Sensori fracture monitoring and Chevron subsurface knowledge embedded in an algorithm.", "Closed-loop system uses real-time feedback to adapt delivery to local formation conditions."], "original_id": "RS049"}], "field_sources": {"business_problem": ["RS049"], "workflow": ["RS049"], "technology_data": ["RS049"], "integration_human_role": ["RS049"], "deployment_scope": ["RS049"], "economics": ["RS049"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-012", "source_ids": ["RS049"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-110"]}, {"key": "halliburton-equinor-onbottom", "existing_ids": [], "related_existing_ids": [], "company": "Halliburton", "title": "Automated on-bottom directional drilling for Equinor", "department": "Drilling & completions", "secondary_departments": [], "category": "Automation / AI not established", "stage": "One North Sea Equinor well delivered Feb 2025", "practical_relevance": "Named live-well deployment and three-system integration make this a concrete autonomous-drilling case. Commercial scaling evidence is not in the source.", "business_problem": "Coordinate downhole, hydraulics and surface rig controls to optimize well placement and drilling parameters in real time.", "workflow": "Integrated LOGIX remote automation, Sekal Drilltronics and rig controls perform autonomous directional drilling with automated hydraulics and dynamic surface equipment control.", "technology_data": "Closed-loop models for subsurface, wellbore fluid and pressure systems integrated with rig automation.", "integration_human_role": "A one-push initiation is described; safe limits cited. Human override/monitoring architecture not detailed.", "deployment_scope": "One Equinor well on Norwegian Continental Shelf; future scale not disclosed.", "partners": ["Equinor", "Sekal"], "disclosed_results": [], "economics": "No ROP/time/cost measure disclosed.", "timeline": [{"date": "2025-02-26", "event": "Halliburton and Sekal reported delivery of an Equinor well with integrated automated on-bottom drilling.", "source_ids": ["RS050"]}], "limitations": ["Company-reported or vendor case study; no independent validation identified in reviewed sources.", "Source describes automated/autonomous drilling; AI/ML method not disclosed."], "assessment": "Analyst inference: The one-well integrated deployment is a material drilling-automation case; it should not be counted as AI without explicit AI evidence.", "sources": [{"id": "RS050", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "publisher": "Halliburton", "date": "2025-02-26", "type": "company announcement", "locator": "Deployment description", "review_status": "full text reviewed", "supports": ["A well for Equinor on the Norwegian Continental Shelf used LOGIX, Drilltronics and rig control for autonomous directional drilling, automated wellbore hydraulics and surface equipment control.", "Integrated models optimize drilling parameters and well placement in real time; a single action initiates the workflow."], "original_id": "RS050"}], "field_sources": {"business_problem": ["RS050"], "workflow": ["RS050"], "technology_data": ["RS050"], "integration_human_role": ["RS050"], "deployment_scope": ["RS050"], "economics": ["RS050"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-013", "source_ids": ["RS050"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-157"]}, {"key": "baker-repsol-leucipa", "existing_ids": ["AI-112"], "related_existing_ids": [], "company": "Baker Hughes", "title": "Leucipa GenAI assistant and lift optimizer for Repsol", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Agreement and early pilots announced June 2025", "practical_relevance": "A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.", "business_problem": "Improve access to production insights and reduce manual surveillance burden on plunger-lift well engineers.", "workflow": "Leucipa virtual assistant offers conversational data access and real-time/predictive insight. Repsol was piloting Lift Optimizer for plunger-lift wells to help monitor, control and maintain wells.", "technology_data": "GenAI chat capability and production/lift optimization analytics within automated field production solution; models and source systems not described.", "integration_human_role": "Supports production engineers; optimizer pilot not stated to autonomously control wells.", "deployment_scope": "Repsol plunger-lift wells; number/field footprint not disclosed.", "partners": ["Repsol"], "disclosed_results": [], "economics": "No measured outcomes, return or deployment scale disclosed.", "timeline": [{"date": "2025-06-10", "event": "Leucipa GenAI and Lift Optimizer announced; Repsol early-stage pilot", "source_ids": ["R081"]}], "limitations": ["Company announcement describes early-stage deployment/pilot as of June 2025."], "assessment": "Analyst inference: A named customer pilot links a GenAI assistant and specific artificial-lift optimization to an operational task. Evidence does not yet establish commercial results.", "sources": [{"id": "R081", "title": "Baker Hughes, Repsol Launching New AI-Powered Functionality in Leucipa", "url": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "publisher": "Baker Hughes", "date": "2025-06-10", "type": "company announcement", "locator": "Leucipa assistant and lift optimizer", "review_status": "full text reviewed", "supports": ["Agreement adds GenAI virtual assistant for conversational access to data and real-time insights and predictive analytics.", "Repsol was piloting Leucipa Lift Optimizer for plunger-lift wells to reduce time engineers spend monitoring, controlling and maintaining individual wells."], "original_id": "baker_repsol"}], "field_sources": {"business_problem": ["R081"], "workflow": ["R081"], "technology_data": ["R081"], "integration_human_role": ["R081"], "deployment_scope": ["R081"], "economics": ["R081"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-014", "source_ids": ["R081"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-112"]}, {"key": "weatherford-intelligence", "existing_ids": ["AI-113"], "related_existing_ids": [], "company": "Weatherford", "title": "Industrial Intelligence portfolio for AI and edge operations", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Portfolio launched Oct 2025", "practical_relevance": "The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.", "business_problem": "Connect heterogeneous industrial data and deliver analytics/control and AI tools within customers’ existing operations systems.", "workflow": "Portfolio includes AI/physics-based MetaFlow virtual flow metering, Engineer’s Briefcase document/transcription/image/language tools, and StrataCore production model, with unified data and edge/software layers.", "technology_data": "AI and physics models, edge systems, unified Synsera data model and software Launchpad; interoperable with customer systems/hardware.", "integration_human_role": "Domain engineers, field/surveillance engineers and executives are named user groups; specific autonomy/human controls vary by tool.", "deployment_scope": "Portfolio breadth advertised; specific customer deployments or performance evidence not disclosed in launch release.", "partners": ["Weatherford software partners and customers"], "disclosed_results": [], "economics": "No consolidated revenue or customer productivity metric disclosed.", "timeline": [{"date": "2025-10-01", "event": "Industrial Intelligence portfolio launched", "source_ids": ["RS051"]}, {"date": "2025", "event": "Internal agents embedded in company workflows, per 2025 sustainability report", "source_ids": ["RS052"]}], "limitations": ["Vendor portfolio description should not be mistaken for rollout. Separate internal AI agents reported in sustainability report from customer product capability."], "assessment": "Analyst inference: The launch is a product portfolio, not a deployment. Several named components have practical tasks, but evidence for adoption and measurable results is limited.", "sources": [{"id": "RS051", "title": "Weatherford Launches Industrial Intelligence at FWRD 2025", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "publisher": "Weatherford", "date": "2025-10-01", "type": "company announcement", "locator": "Product portfolio descriptions", "review_status": "full text reviewed", "supports": ["Portfolio includes MetaFlow AI/physics-based virtual flow meter, Engineer’s Briefcase AI document transcription/recognition/translation tools and StrataCore production foundational model.", "It also includes unified data model, edge and SaaS software; portfolio launch does not itself verify customer deployment."], "original_id": "weatherford_2025"}, {"id": "RS052", "title": "Weatherford 2025 Sustainability Report: Artificial Intelligence", "url": "https://www.weatherford.com/documents/sustainability/sustainability-report/", "publisher": "Weatherford", "date": "2026-04-01", "type": "sustainability report", "locator": "Artificial Intelligence section", "review_status": "excerpt reviewed", "supports": ["Weatherford reports an AI Impact Hub and agents embedded in internal workflows in 2025.", "Named agents review material test reports, cross-reference parts/inventory, classify spend and simplify bidding; training and human involvement stated."], "original_id": "weatherford_internal"}], "field_sources": {"business_problem": ["RS051", "RS052"], "workflow": ["RS051", "RS052"], "technology_data": ["RS051", "RS052"], "integration_human_role": ["RS051", "RS052"], "deployment_scope": ["RS051", "RS052"], "economics": ["RS051", "RS052"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-015", "source_ids": ["RS051", "RS052"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-113"]}, {"key": "nov-dba-egypt", "existing_ids": [], "related_existing_ids": [], "company": "NOV", "title": "Drilling Beliefs & Analytics anomaly warning in Egypt", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployed across operator’s Egypt rig fleet; specific first case Jan 2025", "practical_relevance": "Specific data signals, alert, field response and avoided cost make this a strong vendor case study; operator identity and independent validation are absent.", "business_problem": "Detect drilling dysfunction and extend bit performance before a costly trip or replacement.", "workflow": "Bayesian models and decision trees evaluate weight on bit, torque, mechanical specific energy and depth of cut at edge; Bit Balling Belief alert prompted crew anti-balling pill sweeps. RTOC engineers and crews jointly monitor conditions.", "technology_data": "Max edge platform ingests rig data; wellsite and cloud share insights. AI evaluates dysfunction likelihoods; remote ops supports crew decisions.", "integration_human_role": "Rig crew acted on warning; NOV customer success and operator RTOC personnel monitored and guided decisions.", "deployment_scope": "Egypt operator’s rig fleet; source says deployed across full fleet, with RTOC monitoring fleet operations.", "partners": ["Operator undisclosed"], "disclosed_results": [{"metric": "additional drilling without bit replacement", "value": "1,000 ft", "period": "January 2025 Egypt case example", "attribution": "Vendor-reported", "source_ids": ["RS053"]}, {"metric": "trip cost avoided", "value": "approximately $75,000", "period": "January 2025 Egypt case example", "attribution": "Vendor-reported estimate", "source_ids": ["RS053"]}], "economics": "NOV estimated that the alert response avoided a trip costing approximately $75,000; this is a vendor estimate for one case, not an audited saving.", "timeline": [{"date": "2025-01", "event": "Bit Balling Belief alert prompted crew response in the Egypt example.", "source_ids": ["RS053"]}, {"date": "2025-06-27", "event": "NOV published details of the case and Egypt remote operations center.", "source_ids": ["RS053"]}], "limitations": ["The operator, exact well count and independent validation are not disclosed.", "The 1,000 ft and approximately $75,000 figures are vendor-reported for one intervention."], "assessment": "Analyst inference: Specific data signals, alert, field response and avoided cost make this a strong vendor case study; operator identity and independent validation are absent.", "sources": [{"id": "RS053", "title": "NOV supports record bit run and Egypt’s first remote operations center", "url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center", "publisher": "NOV", "date": "2025-06-27", "type": "company announcement", "locator": "DBA functionality, field result and RTOC", "review_status": "full text reviewed", "supports": ["Drilling Beliefs & Analytics uses Bayesian models/decision trees on well parameters to warn of dysfunction at rig edge.", "In one Egypt well, Bit Balling Belief alert prompted crew response; ROP recovered and an extra 1,000 ft drilled without a trip, with NOV estimating ~$75,000 avoided cost.", "Customer crew and remote operations center acted on alerts; company says deployed across operator’s Egypt rig fleet."], "original_id": "nov_egypt"}], "field_sources": {"business_problem": ["RS053"], "workflow": ["RS053"], "technology_data": ["RS053"], "integration_human_role": ["RS053"], "deployment_scope": ["RS053"], "economics": ["RS053"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-016", "source_ids": ["RS053"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-158"]}, {"key": "nov-aiq-pilot", "existing_ids": ["AI-114"], "related_existing_ids": [], "company": "NOV", "title": "AIQ collaboration for edge-enabled drilling AI pilots", "department": "Drilling & completions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Collaboration announced Jan 2026; pilot planning stage", "practical_relevance": "Strategic agreement is an AI roadmap and test program, not a field deployment yet.", "business_problem": "Develop AI workflows for live drilling operations with secure access to equipment and operational data.", "workflow": "NOV and AIQ plan targeted PoCs and pilots; Max edge platform captures data, processes it at edge and integrates AI workflows.", "technology_data": "NOV drilling equipment/Max IIoT with AIQ industrial AI; no models or use-case performance specified.", "integration_human_role": "Future real-time operational decision support; details on crew interaction not disclosed.", "deployment_scope": "Targeted live-setting pilots, no customer or rig named.", "partners": ["AIQ"], "disclosed_results": [], "economics": "No contract or economic terms disclosed.", "timeline": [{"date": "2026-01-12", "event": "NOV-AIQ collaboration announced", "source_ids": ["R083"]}], "limitations": ["Announcement specifies future PoC/pilot, not completed deployment."], "assessment": "Analyst inference: Strategic agreement is an AI roadmap and test program, not a field deployment yet.", "sources": [{"id": "R083", "title": "NOV and AIQ to Advance AI-Enabled Drilling Optimization", "url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "publisher": "NOV", "date": "2026-01-12", "type": "company announcement", "locator": "Collaboration scope", "review_status": "full text reviewed", "supports": ["NOV and AIQ plan proof-of-concept/pilot AI drilling solutions in live settings.", "NOV Max platform supports secure data capture, edge processing and workflow integration; announcement is plan, not completed deployment."], "original_id": "nov_aiq"}], "field_sources": {"business_problem": ["R083"], "workflow": ["R083"], "technology_data": ["R083"], "integration_human_role": ["R083"], "deployment_scope": ["R083"], "economics": ["R083"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-017", "source_ids": ["R083"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-114"]}, {"key": "nabors-corva-predictive", "existing_ids": ["AI-115"], "related_existing_ids": [], "company": "Nabors", "title": "RigCLOUD Powered by Corva drilling intelligence", "department": "Drilling & completions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Expanded strategic alliance announced Apr 2025", "practical_relevance": "Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.", "business_problem": "Give drilling crews and remote centers real-time analytics and predictive insight across connected rig fleets.", "workflow": "RigCLOUD’s edge/cloud infrastructure connects Corva AI applications; optional SmartROS integration adds rig automation and real-time rigsite applications.", "technology_data": "Corva AI analytics and apps, edge-cloud data, SmartROS automation optional.", "integration_human_role": "Drillers/site managers/directional drillers access insights at rig and remote operations; automation depends on rig controls and user setup.", "deployment_scope": "Marketed across Nabors and third-party contractor rigs; no new customer rollout named in alliance expansion.", "partners": ["Corva AI"], "disclosed_results": [], "economics": "No new 2025 deployment economics disclosed. Older 2023 Predictive Drilling case reported 36% ROP gain and 9.7% vibration reduction; those results are outside window and are not treated as current-window deployment.", "timeline": [{"date": "2025-04-10", "event": "RigCLOUD-Corva collaboration expanded", "source_ids": ["RS054"]}], "limitations": ["Expanded partnership in window; cited product performance page does not date its case example, and Nabors first disclosed metrics in 2023."], "assessment": "Analyst inference: Platform expands a real digital distribution/automation partnership, but specific 2025 adoption or outcomes are not reported. Keep older performance claims out of the current-window results.", "sources": [{"id": "RS054", "title": "Nabors and Corva Expand Strategic Alliance to Accelerate Digital Innovation", "url": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "publisher": "Nabors", "date": "2025-04-10", "type": "company announcement", "locator": "RigCLOUD Powered by Corva scope", "review_status": "full text reviewed", "supports": ["Expanded alliance integrates RigCLOUD edge/cloud platform with Corva AI analytics and apps for real-time drilling data, predictive insights and optimization.", "Operators and contractors can use the tools at rigsite and remotely; SmartROS controls are optional. Announcement describes solution capability, not specific named customer rollout."], "original_id": "nabors_corva"}, {"id": "RS055", "title": "Predictive Drilling - Nabors", "url": "https://www.nabors.com/drilling-automation/predictive-drilling/", "publisher": "Nabors", "date": "Undated; accessed 2026-09-27", "type": "company product documentation", "locator": "Success in Action and availability", "review_status": "full text reviewed", "supports": ["Nabors reports the Predictive Drilling ROP optimizer combines AI/ML with SmartROS/RigCLOUD for remote autodriller setpoint control.", "First deployment reported 36% higher average ROP, 9.7% lower vibration and 80% lateral footage utilization; historic case source (2023) is outside window."], "original_id": "nabors_trial"}], "field_sources": {"business_problem": ["RS054", "RS055"], "workflow": ["RS054", "RS055"], "technology_data": ["RS054", "RS055"], "integration_human_role": ["RS054", "RS055"], "deployment_scope": ["RS054", "RS055"], "economics": ["RS054", "RS055"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-018", "source_ids": ["RS054", "RS055"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-115"]}, {"key": "hp-flexrobotics", "existing_ids": ["AI-116"], "related_existing_ids": [], "company": "Helmerich & Payne", "title": "FlexRobotics rig-floor automation; AI-enabled monitoring add-on deployment unconfirmed", "department": "HSE & emissions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Base automation deployed on a Permian rig; AI-enabled zone-monitoring add-on available in product documentation, customer deployment not confirmed.", "practical_relevance": "A concrete rig-floor robotics deployment reduces exposure to tubular handling work; AI evidence is limited to an optional monitoring capability whose use at the deployed rig is unconfirmed.", "business_problem": "Reduce personnel exposure and standardize tubular handling and connection work on rig floor.", "workflow": "Three robotic arms automate pipe tailing, cleaning/dope/alignment and racking. H&P separately lists AI-enabled zone monitoring as an add-on camera feature; it does not link that feature to the reported customer deployment.", "technology_data": "FlexRobotics uses robotic arms and automation software; product documentation describes AI-enabled zone monitoring on an add-on camera. AI methods, sensor data and use at the customer site are not disclosed.", "integration_human_role": "Automation performs pipe handling to keep personnel away from rig-floor red zones. The sources do not describe operator approvals/interventions or AI alert review; use of the add-on at the reported rig is unconfirmed.", "deployment_scope": "H&P reports one Permian rig deployment for an unnamed supermajor across multiple pads. The product page does not establish AI-add-on deployment at that site.", "partners": ["NOV", "Yaskawa"], "disclosed_results": [], "economics": "No quantified customer result disclosed. Product-page connection-time statement is a product capability, not a verified customer deployment outcome.", "timeline": [{"date": "2026-02-04", "event": "H&P reports first FlexRobotics system deployment for a supermajor", "source_ids": ["RS056"]}, {"date": "2026-05-06", "event": "H&P earnings say deployment is expanding", "source_ids": ["RS056"]}], "limitations": ["Earnings source confirms base robotics/automation deployment, not AI use in the deployed robot.", "AI-enabled zone monitoring is an ancillary add-on in product documentation; customer deployment of that feature is not established.", "Operator identity and deployment-specific performance/economics are not disclosed."], "assessment": "Analyst inference: include the add-on as commercial AI product capability, while classifying the proven field deployment as robotics/automation. The available sources do not establish that the AI add-on was installed at the customer rig.", "sources": [{"id": "RS056", "title": "Helmerich & Payne, Inc. Announces Fiscal First Quarter Results", "url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "publisher": "H&P", "date": "2026-02-04", "type": "earnings release", "locator": "FlexRobotics deployment", "review_status": "full text reviewed", "supports": ["H&P reports FlexRobotics deployed on a Permian rig for a supermajor across multiple pads; growing customer interest.", "Rig-floor system is collaborative robotics for tubular handling. H&P product page says AI-enabled zone monitoring is an ancillary feature; core handling automation is not AI."], "original_id": "hp_q1"}, {"id": "RS057", "title": "FlexRobotics Hands-Free Rig Floor", "url": "https://www.hpinc.com/technologies/flexrobotics", "publisher": "H&P", "date": "Undated; accessed 2026-09-27", "type": "company product documentation", "locator": "Description, operation and AI-enabled zone monitoring", "review_status": "full text reviewed", "supports": ["Three robot arms handle pipe tailing, cleaning/alignment and racking. H&P lists 1:40 slip-to-slip connection times and red-zone personnel exposure reduction.", "System was developed with NOV and uses Yaskawa robot manipulators. Cameras support AI-enabled zone monitoring."], "accessed_date": "2026-09-27", "original_id": "hp_product"}], "field_sources": {"business_problem": ["RS056", "RS057"], "workflow": ["RS056", "RS057"], "technology_data": ["RS056", "RS057"], "integration_human_role": ["RS056", "RS057"], "deployment_scope": ["RS056", "RS057"], "economics": ["RS056", "RS057"]}, "verification": "verified primary for base automation deployment; AI add-on is product capability only", "corrections_to_existing": "Separate customer deployment of base robotics/automation from the separately listed AI-enabled camera add-on; deployment of the add-on is unconfirmed.", "record_ids": ["AI-116"], "cohort": "services", "packet_id": "services-019", "source_ids": ["RS056", "RS057"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise."}, {"key": "patterson-cortex-rex", "existing_ids": ["AI-117"], "related_existing_ids": [], "company": "Patterson-UTI", "title": "Cortex machine-learning autodriller and REX rig alerts", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Nearly all active rigs equipped with Cortex apps by Q4 2025", "practical_relevance": "This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.", "business_problem": "Improve repeatability of rig drilling performance and detect operating issues early.", "workflow": "Cortex advanced ML autodriller applies machine learning to drilling parameters; REX cloud early-alert monitoring gives field operations insight. By year-end, nearly all rigs had Cortex automation applications.", "technology_data": "Machine-learning autodriller and cloud alerting platform; data features/model detail and exact rig count not stated.", "integration_human_role": "Drillers receive operating setpoints/alerts; source does not detail control mode, override or alert actioning.", "deployment_scope": "Patterson-UTI U.S. rig fleet; nearly all rigs for broader Cortex automation suite.", "partners": ["Internal teams"], "disclosed_results": [], "economics": "No AI-specific performance or economic result; Q2 2025 management cited growing customer demand/revenue for drilling automation products.", "timeline": [{"date": "2025-07-30", "event": "Q2 results detail Cortex ML autodriller and REX alerts", "source_ids": ["R086"]}, {"date": "2026-02-04", "event": "Q4 results report Cortex automation applications on nearly all rigs", "source_ids": ["RS058"]}], "limitations": ["Company’s broad “automation” rig penetration is not AI-specific; machine-learning claim applies to Cortex autodriller."], "assessment": "Analyst inference: This is scalable, named AI-enabled product use, but year-end “nearly all” measures the automation suite, not a claim that every rig used ML autodrilling continuously.", "sources": [{"id": "R086", "title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended June 30, 2025", "url": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "publisher": "Patterson-UTI", "date": "2025-07-30", "type": "earnings release", "locator": "Drilling services remarks", "review_status": "full text reviewed", "supports": ["Cortex platform enables a machine-learning autodriller; REX is a cloud-based early-alert field monitoring system.", "Company cites increased demand and revenue from drilling automation technologies, but no separate AI return metric."], "original_id": "pten"}, {"id": "RS058", "title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025", "url": "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm", "publisher": "Patterson-UTI", "date": "2026-02-04", "type": "earnings release", "locator": "Drilling Services", "review_status": "full text reviewed", "supports": ["Patterson-UTI said nearly all rigs were equipped with proprietary Cortex automation applications by year-end 2025.", "Management cited continued development of automation software and demand for performance-enhancing products; this is scale of automation suite, not AI-specific model usage."], "original_id": "pten_q4"}], "field_sources": {"business_problem": ["R086", "RS058"], "workflow": ["R086", "RS058"], "technology_data": ["R086", "RS058"], "integration_human_role": ["R086", "RS058"], "deployment_scope": ["R086", "RS058"], "economics": ["R086", "RS058"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-020", "source_ids": ["R086", "RS058"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-117"]}, {"key": "liberty-forge-stimcommander", "existing_ids": ["AI-118"], "related_existing_ids": [], "company": "Liberty Energy", "title": "Forge LLM orchestration for StimCommander frac control", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Forge launched in Q3 2025; StimCommander use already operating", "practical_relevance": "A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.", "business_problem": "Coordinate rate and pressure settings across frac assets and improve speed/efficiency of hydraulic stimulation.", "workflow": "StimCommander uses AI-driven automated rate and pressure control; Forge LLM adds intelligent asset orchestration and continuous AI optimization.", "technology_data": "Cloud LLM plus automated controls and frac fleet telemetry; detailed architecture and data scope not disclosed.", "integration_human_role": "System automates setpoint/time-to-rate control; role of frac engineers and safety constraints not described in release.", "deployment_scope": "Liberty frac fleets; specific customer/site count not disclosed.", "partners": ["Internal Liberty teams"], "disclosed_results": [{"metric": "time to desired fluid injection rate", "value": "65% improvement", "period": "Q3 2025 company report", "attribution": "Company-reported; StimCommander automation, not Forge alone", "source_ids": ["R087"]}, {"metric": "hydraulic efficiency", "value": "5%-10% improvement", "period": "Q3 2025 company report", "attribution": "Company-reported; StimCommander automation, not Forge alone", "source_ids": ["R087"]}], "economics": "Liberty reports 65% faster time to desired injection rate and 5-10% better hydraulic efficiency. These relate to fleet automation/StimCommander; no causal contribution from Forge alone is quantified. DigiPrime >30% maintenance savings is hardware result, not AI.", "timeline": [{"date": "2025-10-16", "event": "Forge launched and described as augmenting StimCommander", "source_ids": ["R087"]}], "limitations": ["Metrics are company-reported and are not attributed to Forge specifically; keep digiPrime pump savings separate."], "assessment": "Analyst inference: A named LLM is connected to an operating completions control system, while the performance metrics are reported for automation overall, not necessarily the LLM.", "sources": [{"id": "R087", "title": "Liberty Energy Inc. Announces Third Quarter 2025 Financial and Operational Results", "url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "publisher": "Liberty Energy", "date": "2025-10-16", "type": "earnings release", "locator": "Forge and StimCommander paragraphs", "review_status": "full text reviewed", "supports": ["Liberty launched Forge LLM for intelligent asset orchestration integrated with StimCommander rate/pressure control.", "Company reports automated control improved time to target fluid injection rate by 65% and hydraulic efficiency by 5-10%; maintenance savings >30% on digiPrime are pump technology results, not AI."], "original_id": "liberty"}], "field_sources": {"business_problem": ["R087"], "workflow": ["R087"], "technology_data": ["R087"], "integration_human_role": ["R087"], "deployment_scope": ["R087"], "economics": ["R087"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-021", "source_ids": ["R087"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-118"]}, {"key": "precision-alpha-ai", "existing_ids": [], "related_existing_ids": [], "company": "Precision Drilling", "title": "Alpha drilling analytics and predictive maintenance", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "78 rigs equipped with Alpha digital technology at end-2024; report published Apr 2025", "practical_relevance": "This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.", "business_problem": "Use rig/drilling data to improve performance, adapt drilling in real time and plan equipment maintenance before failures.", "workflow": "Alpha suite aggregates drilling, equipment, geology and environmental data; AI algorithms identify patterns/trends, adjust operations and support predictive maintenance scheduling.", "technology_data": "Digital automation and analytics suite with AI; vendor detail/model/connection architecture not in report.", "integration_human_role": "Drilling staff use recommendations and automation; independent impact or autonomy boundary not detailed.", "deployment_scope": "78 rigs in North America at end-2024.", "partners": ["Internal Precision Drilling teams"], "disclosed_results": [], "economics": "No quantified AI-only uptime, downtime, cost or performance results.", "timeline": [{"date": "2024-12-31", "event": "Precision reports Alpha digital technology deployed on 78 rigs.", "source_ids": ["RS059"]}, {"date": "2025-04", "event": "2024 ESG report published.", "source_ids": ["RS059"]}], "limitations": ["Company-reported or vendor case study; no independent validation identified in reviewed sources."], "assessment": "Analyst inference: This is useful fleet-scale evidence of AI capability, but disclosures bundle it with automation/analytics and do not separate AI’s effects.", "sources": [{"id": "RS059", "title": "Precision Drilling 2024 ESG Report", "url": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "publisher": "Precision Drilling", "date": "2025-04", "type": "sustainability report", "locator": "Alpha Suite technologies", "review_status": "excerpt reviewed", "supports": ["Precision reports 78 rigs with Alpha digital technology at end-2024.", "Alpha suite combines digital automation and analytics with AI algorithms for drilling performance pattern recognition, real-time adjustments and predictive maintenance; report does not break out AI-specific results."], "original_id": "precision"}], "field_sources": {"business_problem": ["RS059"], "workflow": ["RS059"], "technology_data": ["RS059"], "integration_human_role": ["RS059"], "deployment_scope": ["RS059"], "economics": ["RS059"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-022", "source_ids": ["RS059"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-159"]}, {"key": "saipem-ai-schedule", "existing_ids": [], "related_existing_ids": [], "company": "Saipem", "title": "AI Project Schedule Onshore for EPC risk forecasting", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "AI engineering initiative launched 2024; 13 use cases explored by 2025", "practical_relevance": "Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.", "business_problem": "Anticipate construction delays and bottlenecks and improve EPC project planning decisions using prior project data.", "workflow": "Project Schedule Onshore analyzes historical project data with predictive models to estimate construction risk and delays for engineering/procurement/construction planning.", "technology_data": "Predictive models on historical project data; features, training set and validation metrics not disclosed.", "integration_human_role": "Planning teams use risk signals to make faster decisions; not autonomous schedule control.", "deployment_scope": "Saipem engineering/projects; number of projects or teams using the tool not disclosed.", "partners": ["Internal Saipem teams"], "disclosed_results": [], "economics": "No quantified schedule accuracy or financial return disclosed.", "timeline": [{"date": "2024", "event": "Saipem launched its two-year AI for Engineering initiative.", "source_ids": ["RS060"]}, {"date": "2025", "event": "Saipem named 13 use cases and AI Project Schedule Onshore.", "source_ids": ["RS060"]}], "limitations": ["Company-reported or vendor case study; no independent validation identified in reviewed sources."], "assessment": "Analyst inference: Named working application inside a 13-use-case engineering program provides practical scope. Adoption and performance evidence remain light.", "sources": [{"id": "RS060", "title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "publisher": "Saipem", "date": "2026-04", "type": "sustainability report", "locator": "Harnessing AI for engineering section", "review_status": "excerpt reviewed", "supports": ["Saipem’s AI for Engineering is a two-year cross-functional effort launched 2024; 13 use cases explored by 2025.", "AI Project Schedule Onshore analyzes historical project data with predictive models for EPC construction risks/delays and bottlenecks; 12 additional apps were planned for 2026."], "original_id": "saipem"}], "field_sources": {"business_problem": ["RS060"], "workflow": ["RS060"], "technology_data": ["RS060"], "integration_human_role": ["RS060"], "deployment_scope": ["RS060"], "economics": ["RS060"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-023", "source_ids": ["RS060"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-160"]}, {"key": "transocean-office-ai", "existing_ids": [], "related_existing_ids": [], "company": "Transocean", "title": "AI tools for office and administrative productivity", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "AI tools introduced by 2025; no use in offshore drilling stated", "practical_relevance": "This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.", "business_problem": "Reduce effort in high-volume office and administrative tasks.", "workflow": "Company says it has introduced AI tools for unspecified high-volume administrative tasks and is exploring more productivity use cases.", "technology_data": "AI tool types and workflow data are not disclosed; governance committee oversees tool selection, security, training and usage.", "integration_human_role": "Corporate users; human decision processes implied but details not disclosed.", "deployment_scope": "Office/administrative workflows only; Transocean explicitly says AI is not currently used in offshore drilling operations.", "partners": ["Internal Transocean teams"], "disclosed_results": [], "economics": "No productivity, time or cost metric disclosed.", "timeline": [{"date": "2026-02-19", "event": "2025 Form 10-K reports AI use in office tasks and no AI use in offshore drilling operations.", "source_ids": ["RS061"]}], "limitations": ["Company-reported or vendor case study; no independent validation identified in reviewed sources."], "assessment": "Analyst inference: This is important negative-scope evidence: AI adoption exists, while offshore drilling operations remain outside stated AI use. Automated drilling systems and robotics in fleet should not be mislabeled AI.", "sources": [{"id": "RS061", "title": "Transocean Ltd. 2025 Form 10-K", "url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "publisher": "Transocean", "date": "2026-02-19", "type": "annual report", "locator": "Technology and AI discussion", "review_status": "full text reviewed", "supports": ["Transocean explicitly says AI is not currently used in offshore drilling operations; it has introduced AI tools for high-volume office/administrative tasks and is exploring productivity use cases.", "Company describes AI governance committee; fleet automation/robotics systems are separate from AI."], "original_id": "transocean"}], "field_sources": {"business_problem": ["RS061"], "workflow": ["RS061"], "technology_data": ["RS061"], "integration_human_role": ["RS061"], "deployment_scope": ["RS061"], "economics": ["RS061"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-024", "source_ids": ["RS061"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-161"]}, {"key": "hall-logix", "existing_ids": ["AI-016"], "related_existing_ids": [], "company": "Halliburton", "title": "LOGIX autonomous drilling", "department": "Drilling & completions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Commercial platform; 2025 case study documents a European operator pilot", "practical_relevance": "Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.", "business_problem": "Improve well placement consistency, penetration rate and well delivery in complex directional drilling.", "workflow": "Combines automated steering and drilling-parameter recommendations; commands may be sent to the bottom-hole assembly or shown as advice for directional drillers. Remote center and rig crews view drilling data and advisory outputs.", "technology_data": "Physics-based models and machine learning on drilling data; LOGIX, iCruise RSS and LWD sensors.", "integration_human_role": "The case study describes full closed-loop automation, with commands either automated or advisory; engineering/rig personnel retain operational authority as configured.", "deployment_scope": "Commercial product with a case study of three automated wells over two years for an unnamed European operator.", "partners": ["Halliburton; unnamed European operator; iCruise"], "disclosed_results": [{"metric": "Average ROP increase across four sections", "value": "43.6%", "period": "Three-well pilot over two years", "attribution": "Vendor-reported", "source_ids": ["RS062"]}, {"metric": "Well delivery time improvement", "value": ">30%", "period": "Three-well pilot over two years", "attribution": "Vendor-reported", "source_ids": ["RS062"]}], "economics": "Case study reports >30% improvement in well delivery time and 43.6% average ROP increase across four well sections; Halliburton-reported, three wells. No AI-only financial return.", "timeline": [{"date": "2025-Q1", "event": "Halliburton investor presentation includes LOGIX as autonomous drilling product.", "source_ids": ["S07"]}, {"date": "Undated; accessed 2026-09-27", "event": "Halliburton case study reports European operator pilot outcomes.", "source_ids": ["RS062"]}], "limitations": ["Case study operator and field not named. Vendor-reported results; no independent corroboration.", "The existing 2025 investor presentation is an excerpted product description; separate case study gives additional, vendor-reported workflow details."], "assessment": "Analyst inference: Named workflow plus field-performance evidence strengthen an otherwise generic capability claim, though customer and independent validation are absent.", "sources": [{"id": "S07", "title": "Halliburton Q1 2025 Investor Presentation", "url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "publisher": "Halliburton", "date": "2025-Q1", "type": "investor presentation", "locator": "Slides 11 and 19", "review_status": "full text reviewed", "supports": ["LOGIX is presented as an autonomous drilling platform combining physics-based models and machine learning."], "original_id": "hall_logix_presentation"}, {"id": "RS062", "title": "Enhanced well delivery with LOGIX automation and remote operations", "url": "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop", "publisher": "Halliburton", "date": "Undated; accessed 2026-09-27", "type": "partner case study", "locator": "Overview, solution, results", "review_status": "full text reviewed", "supports": ["An unnamed European operator piloted closed-loop directional drilling; remote center and rig personnel view data/advisories.", "Across three wells, Halliburton reports 43.6% average ROP increase and >30% well-delivery-time improvement."], "original_id": "hall_logix_case"}], "field_sources": {"business_problem": ["S07", "RS062"], "workflow": ["S07", "RS062"], "technology_data": ["S07", "RS062"], "integration_human_role": ["S07", "RS062"], "deployment_scope": ["S07", "RS062"], "economics": ["S07", "RS062"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-025", "source_ids": ["S07", "RS062"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-016"]}, {"key": "hall-ds365", "existing_ids": ["AI-017"], "related_existing_ids": ["AI-016"], "company": "Halliburton", "title": "DecisionSpace 365 (DS365.ai) cloud E&P software", "department": "Drilling & completions", "secondary_departments": [], "category": "Commercial AI product", "stage": "Commercial product suite presented in 2025; no distinct AI module deployment verified", "practical_relevance": "This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.", "business_problem": "Give exploration and production teams cloud software and AI/ML-assisted insight across subsurface, drilling and production decisions.", "workflow": "Investor material describes DS365.ai within cloud E&P services; it does not enumerate a specific model workflow or customer deployment for the AI module.", "technology_data": "DecisionSpace 365 subscription services, iEnergy cloud environment, open architecture; AI/ML functionality described at product-family level.", "integration_human_role": "Not described for the AI module in the reviewed investor presentation.", "deployment_scope": "Commercial software capability; no named customer or deployment level established.", "partners": [], "disclosed_results": [], "economics": "Not disclosed. Product revenue, AI-specific use and returns not disclosed.", "timeline": [{"date": "2025-Q1", "event": "Investor presentation identifies DS365.ai within DecisionSpace 365 cloud services.", "source_ids": ["S07"]}], "limitations": ["Source is investor presentation excerpt and offers little technical detail.", "Do not infer a particular AI task, closed-loop control, or customer adoption from product-suite naming."], "assessment": "Analyst inference: This is a credible commercial product family but the public evidence is much less specific than LOGIX or named customer applications.", "sources": [{"id": "S07", "title": "Halliburton Q1 2025 Investor Presentation", "url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "publisher": "Halliburton", "date": "2025-Q1", "type": "investor presentation", "locator": "Slides 11 and 19", "review_status": "full text reviewed", "supports": ["DecisionSpace 365 is described as a cloud E&P services portfolio with AI and machine learning spanning subsurface, drilling and production decisions."], "original_id": "hall_ds365"}, {"id": "S07", "title": "Halliburton Q1 2025 Investor Presentation", "url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "publisher": "Halliburton", "date": "2025-Q1", "type": "investor presentation", "locator": "Slides 11 and 19", "review_status": "full text reviewed", "supports": ["DecisionSpace 365 is described as cloud E&P services with AI/ML spanning subsurface, drilling and production decisions."], "original_id": "hall_logix_presentation"}], "field_sources": {"business_problem": ["S07"], "workflow": ["S07"], "technology_data": ["S07"], "integration_human_role": ["S07"], "deployment_scope": ["S07"], "economics": ["S07"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-026", "source_ids": ["S07"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-017"]}, {"key": "hall-shape-collaboration", "existing_ids": ["AI-018"], "related_existing_ids": [], "company": "Halliburton", "title": "Halliburton–Shape Digital asset performance collaboration", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Commercial AI product", "stage": "Strategic collaboration announced July 2026; no customer deployment disclosed", "practical_relevance": "The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.", "business_problem": "Connect subsurface and surface information so teams can make coordinated production, reliability, energy-efficiency and safety decisions.", "workflow": "Landmark Digital Field Solver reservoir, well and production-network models provide context for Shape Lighthouse, Aura and Reef surface AI/engineering tools, aiming at predictive asset-level decisions through the production lifecycle.", "technology_data": "Trusted operational data, domain models, Shape Digital applied AI portfolio, integrated asset and production models.", "integration_human_role": "Decision-support and predictive insights described; autonomous execution or decision authority not specified.", "deployment_scope": "Collaboration announced; no named customer, asset, deployment schedule, or field scale in source.", "partners": ["Shape Digital"], "disclosed_results": [], "economics": "No customer outcomes or separate commercial terms disclosed.", "timeline": [{"date": "2026-07-21", "event": "Strategic collaboration announced in Halliburton Q2 results and release.", "source_ids": ["RS063"]}], "limitations": ["Announced collaboration and intended workflow only; no named deployment or metrics.", "The announcement does not establish AI performance or autonomy."], "assessment": "Analyst inference: The specific connected asset workflow is more informative than a generic AI alliance, but implementation and measurable results remain unproven.", "sources": [{"id": "RS063", "title": "Halliburton and Shape Digital form strategic collaboration to advance digital asset performance management", "url": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management", "publisher": "Halliburton", "date": "2026-07-21", "type": "company announcement", "locator": "Collaboration scope and connected production planning", "review_status": "full text reviewed", "supports": ["The collaboration combines Landmark Digital Field Solver models with Shape Lighthouse, Aura and Reef to connect subsurface and surface intelligence for predictive asset decisions."], "original_id": "hall_shape"}], "field_sources": {"business_problem": ["RS063"], "workflow": ["RS063"], "technology_data": ["RS063"], "integration_human_role": ["RS063"], "deployment_scope": ["RS063"], "economics": ["RS063"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-027", "source_ids": ["RS063"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-018"]}, {"key": "baker-cordant-epam", "existing_ids": ["AI-020"], "related_existing_ids": [], "company": "Baker Hughes", "title": "Cordant Asset Strategy AI with EPAM", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Commercial AI product", "stage": "Partner-described case / feature; current product is commercial", "practical_relevance": "A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.", "business_problem": "Reduce slow data preparation and specialist effort in asset reliability strategy development, including analysis of engineering documents and unstructured text.", "workflow": "Partner webinar describes use of generative AI and text processing to support Cordant Asset Strategy and engineering document analysis in asset-performance workflows.", "technology_data": "Azure AI services and Azure Data Factory integration within SAP S/4HANA environments, according to partner webinar description.", "integration_human_role": "Reliability engineers use outputs for strategy development; decision approval and generated-content validation details not published in reviewed source description.", "deployment_scope": "Partner-described use case; no customer/site or scale identified in the webinar page reviewed.", "partners": [], "disclosed_results": [], "economics": "No measured time saved, accuracy, cost or ROI disclosed.", "timeline": [{"date": "2025-04-15", "event": "EPAM webinar describes Baker Hughes Cordant Asset Strategy and Azure AI case.", "source_ids": ["S10"]}], "limitations": ["Evidence is partner webinar synopsis, not a detailed customer deployment record.", "No named operator, outcome metric or model governance details disclosed."], "assessment": "Analyst inference: A practical AI assist inside a known maintenance workflow is plausible, but public evidence supports feature/case scope rather than measured use.", "sources": [{"id": "S10", "title": "Speed to Value in Asset Performance with Azure AI: a case study on EPAM and Baker Hughes", "url": "https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes", "publisher": "EPAM", "date": "2025-04-15", "type": "partner webinar description", "locator": "Description and key learnings", "review_status": "excerpt reviewed", "supports": ["EPAM describes generative AI and unstructured text processing for Cordant Asset Strategy, including engineering document analysis and Azure integration."], "original_id": "baker_epam"}], "field_sources": {"business_problem": ["S10"], "workflow": ["S10"], "technology_data": ["S10"], "integration_human_role": ["S10"], "deployment_scope": ["S10"], "economics": ["S10"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-028", "source_ids": ["S10"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-020"]}, {"key": "baker-c3-partnership", "existing_ids": ["AI-021"], "related_existing_ids": [], "company": "Baker Hughes", "title": "Baker Hughes–C3 AI enterprise AI joint venture renewal", "department": "Production & field operations", "secondary_departments": [], "category": "Strategy only", "stage": "Multi-year joint venture agreement renewed through June 2028 in May 2025", "practical_relevance": "Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.", "business_problem": "Continue joint development, co-selling and scaling of enterprise AI applications across energy and industrial asset operations.", "workflow": "The renewal extends the existing joint venture and product/commercial collaboration; announcement provides no newly named workflow or operator deployment.", "technology_data": "C3 AI enterprise application software combined with Baker Hughes energy and industrial domain capabilities.", "integration_human_role": "Customer decision workflows and human controls not specified in the renewal announcement.", "deployment_scope": "Commercial alliance renewal; no deployment count or operational rollout detail in this disclosure.", "partners": [], "disclosed_results": [], "economics": "Agreement value, revenue allocation and investment split not disclosed.", "timeline": [{"date": "2025-05-28", "event": "C3 AI and Baker Hughes announced multi-year renewal/expansion through June 2028.", "source_ids": ["RS064"]}], "limitations": ["Partnership evidence is strong; operational workflow, deployment scale and results are not detailed in renewal announcement.", "Category reflects strategic partnership rather than a newly verified deployment."], "assessment": "Analyst inference: Renewal demonstrates continued commercial commitment but does not independently prove customer adoption or AI value.", "sources": [{"id": "RS064", "title": "C3 AI and Baker Hughes Renew and Expand Joint Venture Agreement", "url": "https://ir.c3.ai/node/10916/pdf", "publisher": "C3 AI", "date": "2025-05-28", "type": "company announcement", "locator": "Agreement scope", "review_status": "full text reviewed", "supports": ["C3 AI and Baker Hughes renewed and expanded their joint venture agreement for multiple years through June 2028."], "original_id": "c3_baker_renewal"}], "field_sources": {"business_problem": ["RS064"], "workflow": ["RS064"], "technology_data": ["RS064"], "integration_human_role": ["RS064"], "deployment_scope": ["RS064"], "economics": ["RS064"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-029", "source_ids": ["RS064"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-021"]}, {"key": "slb-agora-qualcomm", "existing_ids": ["AI-046"], "related_existing_ids": [], "company": "SLB", "title": "Agora edge AI and Qualcomm collaboration", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "MoU announced June 2026; product capability, no named deployment tied to the MoU", "practical_relevance": "Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.", "business_problem": "Support real-time decisions in remote production environments where latency, constrained connectivity, power and continuity matter.", "workflow": "Combine low-power edge AI processing with SLB Agora edge AI and IIoT solutions for applications at wells, facilities and production systems.", "technology_data": "Qualcomm low-power compute and AI; SLB Agora edge AI/IoT and production solutions.", "integration_human_role": "MoU does not specify the degree of automation, supervision or safety interlocks for future applications.", "deployment_scope": "Collaboration aims at energy operations; no customer or installed scope disclosed.", "partners": [], "disclosed_results": [], "economics": "No contract value, deployment costs or quantified economic result.", "timeline": [{"date": "2026-06-09", "event": "SLB and Qualcomm announce MoU for energy edge AI solutions.", "source_ids": ["RS065"]}], "limitations": ["MoU and intended application; not evidence of deployment or customer results.", "Do not conflate SLB historical Agora deployments with this collaboration."], "assessment": "Analyst inference: Edge placement addresses a real operational constraint; evidence currently supports a partnership and intended product direction rather than adoption.", "sources": [{"id": "RS065", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "publisher": "SLB", "date": "2026-06-09", "type": "company announcement", "locator": "Collaboration scope", "review_status": "full text reviewed", "supports": ["The MoU combines Qualcomm low-power edge computing/AI with SLB Agora for real-time decisions across wells, facilities and production systems."], "original_id": "slb_qualcomm"}], "field_sources": {"business_problem": ["RS065"], "workflow": ["RS065"], "technology_data": ["RS065"], "integration_human_role": ["RS065"], "deployment_scope": ["RS065"], "economics": ["RS065"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-030", "source_ids": ["RS065"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-046"]}, {"key": "slb-production-studio", "existing_ids": ["AI-048"], "related_existing_ids": [], "company": "SLB", "title": "Intelligent Production Studio", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched at SLB Production and Recovery Summit, Q2 2026; demonstrator and training facility", "practical_relevance": "This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.", "business_problem": "Let operators test integrated production systems and assess how digital workflows behave before committing to field deployment.", "workflow": "Physical autonomous well pad runs real three-phase oil, water and gas flows across multiple streams; demonstrations include surveillance/anomaly detection, robotic inspection, autonomous chemical injection, digital twins and closed-loop optimization.", "technology_data": "AI, edge intelligence, connected equipment, digital twins, robotic inspection, sensors and control systems integrated on operating infrastructure.", "integration_human_role": "Visitors evaluate and plan potential configuration; source describes a demonstrator rather than customer-operated autonomous assets.", "deployment_scope": "Houston demonstration/training environment; no external field deployment implied.", "partners": [], "disclosed_results": [], "economics": "No standalone economics or measured customer outcomes disclosed.", "timeline": [{"date": "2026-Q2", "event": "Studio launched; more than 100 customers/partners visited by June 2026.", "source_ids": ["RS066"]}], "limitations": ["Studio demonstrations must not be presented as production deployment at customer assets.", "No public independent validation of any production uplift attributable to Studio."], "assessment": "Analyst inference: This is a hands-on evaluation and integration environment that can reduce design uncertainty, but the evidence is about demonstration capability, not customer outcomes.", "sources": [{"id": "RS066", "title": "Inside the Future of Production Operations", "url": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio", "publisher": "SLB", "date": "2026-06-18", "type": "company feature", "locator": "Autonomous pad and visitor details", "review_status": "full text reviewed", "supports": ["SLB says the Houston studio combines AI, automation and connected systems around an operating autonomous pad managing three-phase flow; over 100 customers and partners visited."], "original_id": "slb_studio_launch"}, {"id": "RS067", "title": "Intelligent Production Studio", "url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "publisher": "SLB", "date": "Undated; accessed 2026-09-27", "type": "company product documentation", "locator": "Inside the studio and FAQ", "review_status": "full text reviewed", "supports": ["Studio demonstrates AI surveillance, digital twins, robotic inspection, autonomous chemical injection and closed-loop workflows; this is demonstration scope."], "original_id": "slb_studio_product"}], "field_sources": {"business_problem": ["RS066", "RS067"], "workflow": ["RS066", "RS067"], "technology_data": ["RS066", "RS067"], "integration_human_role": ["RS066", "RS067"], "deployment_scope": ["RS066", "RS067"], "economics": ["RS066", "RS067"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-031", "source_ids": ["RS066", "RS067"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-048"]}, {"key": "slb-nvidia-ai-factory", "existing_ids": ["AI-105"], "related_existing_ids": [], "company": "SLB", "title": "SLB–NVIDIA AI Factory for Energy and modular AI infrastructure", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Collaboration expanded March 2026; design and development announced", "practical_relevance": "Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.", "business_problem": "Provide energy companies with infrastructure and model environments to scale AI across operational data and workflows.", "workflow": "Three tracks: modular design partnership for NVIDIA DSX AI factories; develop AI Factory for Energy reference environment; optimize processing across SLB digital platforms.", "technology_data": "NVIDIA AI infrastructure, Omniverse libraries and Nemotron open models; SLB Delfi/Lumi digital platforms and domain data/models.", "integration_human_role": "No operating workflow or human control details disclosed; reference environment remains in development.", "deployment_scope": "Commercial infrastructure design and AI reference environment ambitions; no completed customer deployment named in announcement.", "partners": [], "disclosed_results": [], "economics": "No deal value, capex, customer economics or production outcome disclosed.", "timeline": [{"date": "2026-03-25", "event": "SLB and NVIDIA announce expanded collaboration in three workstreams.", "source_ids": ["RS068"]}], "limitations": ["The data-center modular design work is AI-demand infrastructure; separate from operator AI adoption.", "AI Factory for Energy is announced development, not established deployment."], "assessment": "Analyst inference: Strategically material platform investment, but should be tracked apart from implemented operational AI because the announced factory is under development.", "sources": [{"id": "RS068", "title": "SLB Industrializes AI for the Energy Industry with NVIDIA", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/", "publisher": "SLB", "date": "2026-03-25", "type": "company announcement", "locator": "Three strategic elements", "review_status": "full text reviewed", "supports": ["SLB/NVIDIA announced modular DSX data-center design, an energy AI reference environment using generative and agentic AI, and accelerated computing for SLB digital platforms."], "original_id": "slb_nvidia"}], "field_sources": {"business_problem": ["RS068"], "workflow": ["RS068"], "technology_data": ["RS068"], "integration_human_role": ["RS068"], "deployment_scope": ["RS068"], "economics": ["RS068"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-032", "source_ids": ["RS068"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-105"]}, {"key": "slb-tachyus", "existing_ids": ["AI-106"], "related_existing_ids": [], "company": "SLB", "title": "Agreement to acquire Tachyus reservoir modeling and optimization software", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Acquisition completed in Q2 2026; operational integration into Delfi/Lumi not yet specifically verified.", "practical_relevance": "A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.", "business_problem": "Enable faster tactical reservoir management in response to changing operations and improve waterflood, production and recovery decisions.", "workflow": "Aqueon evaluates thousands of reservoir scenarios in minutes for waterflood management, pressure forecasting, production optimization, saltwater disposal and EOR; post-close SLB plans Delfi/Lumi integration for closed-loop reservoir/production workflows.", "technology_data": "Machine learning combined with reservoir physics; Aqueon models reservoir behavior under operating conditions.", "integration_human_role": "Reservoir and production teams adjust field strategy; post-acquisition decision autonomy is not specified.", "deployment_scope": "SLB announced a proposed software acquisition; Aqueon vendor says deployed across 7,500+ wells globally.", "partners": [], "disclosed_results": [{"metric": "Wells using Aqueon", "value": ">7,500", "period": "At May 2026 announcement", "attribution": "Vendor-reported", "source_ids": ["RS069"]}], "economics": "Terms not disclosed; no AI-specific customer ROI reported.", "timeline": [{"date": "2026-05-29", "event": "SLB announces agreement to acquire Tachyus.", "source_ids": ["RS069"]}, {"date": "2026-07-24", "event": "SLB Q2 results confirm the Tachyus acquisition completed during the quarter.", "source_ids": ["RS070"]}, {"date": "2026-07-24", "event": "SLB Q2 results confirm the Tachyus acquisition completed during the quarter.", "source_ids": ["RS070"]}], "limitations": ["SLB plans Delfi/Lumi integration; acquisition close is verified, but completed technical integration and customer deployment are not.", "7,500+ wells is vendor-reported installed scope, not independent outcome validation."], "assessment": "Analyst inference: A strategically relevant physics-plus-ML workflow complements strategic simulators with faster operational scenario analysis; value depends on integration and field adoption.", "sources": [{"id": "RS069", "title": "SLB to Acquire Tachyus", "url": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update", "publisher": "SLB", "date": "2026-05-29", "type": "company announcement", "locator": "Aqueon workflows, deployment and planned integration", "review_status": "full text reviewed", "supports": ["SLB describes Aqueon as combining machine learning and reservoir physics for operational reservoir-management tasks and reports deployment across 7,500+ wells.", "SLB plans post-close integration into Delfi and Lumi; close remained subject to customary conditions at announcement."], "original_id": "slb_tachyus"}, {"id": "RS070", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "publisher": "SLB", "date": "2026-07-24", "type": "company announcement", "locator": "Acquisition completion", "review_status": "full text reviewed", "supports": ["SLB reported that it completed the Tachyus acquisition during Q2 2026."], "original_id": "slb_tachyus_completed"}], "field_sources": {"business_problem": ["RS069"], "workflow": ["RS069"], "technology_data": ["RS069"], "integration_human_role": ["RS069"], "deployment_scope": ["RS069"], "economics": ["RS069"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-033", "source_ids": ["RS069", "RS070"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-106"]}, {"key": "slb-sp-global-upstream", "existing_ids": ["AI-107"], "related_existing_ids": [], "company": "SLB", "title": "Agreement to acquire S&P Global Energy upstream software portfolio and co-develop AI models", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Acquisition completed September 1, 2026; AI model and platform integration remains prospective.", "practical_relevance": "The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.", "business_problem": "Extend SLB digital subsurface coverage into widely used U.S. onshore/unconventional software workflows and connect planning tools with AI.", "workflow": "Portfolio includes Kingdom, Petra, Harmony Enterprise and other engineering tools; separate collaboration plans AI models using S&P upstream data with Lumi and Tela.", "technology_data": "Existing geoscience/petroleum engineering software and upstream data; proposed integration with SLB Lumi and Tela.", "integration_human_role": "Existing user workflows to be maintained; AI decision authority or workflow design not yet disclosed.", "deployment_scope": "Agreement covers software portfolio; transaction not yet closed as of stated expected schedule.", "partners": [], "disclosed_results": [], "economics": "Terms not disclosed; no AI model or customer economics disclosed.", "timeline": [{"date": "2026-04-24", "event": "SLB announces definitive acquisition agreement and separate AI model collaboration.", "source_ids": ["RS071"]}, {"date": "2026-09-01", "event": "S&P Global confirms divestiture completion to SLB.", "source_ids": ["RS072"]}], "limitations": ["Acquisition is complete, but the announced AI model collaboration and integration work remain prospective in the reviewed evidence.", "Do not frame the acquired software installed base as SLB AI adoption."], "assessment": "Analyst inference: The acquisition buys workflow distribution and an installed subsurface user base, while the AI strategy remains prospective until post-close products and adoption are documented.", "sources": [{"id": "RS071", "title": "SLB Advances Digital Strategy with Acquisition of S&P Global Energy’s Upstream Software Portfolio", "url": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition", "publisher": "SLB", "date": "2026-04-24", "type": "company announcement", "locator": "Transaction scope and planned AI collaboration", "review_status": "full text reviewed", "supports": ["SLB signed a definitive agreement to acquire S&P Global Energy upstream geoscience/petroleum software; separately, the parties plan to build AI models using Lumi/Tela and S&P data.", "The announced close was expected H2 2026 or early 2027, subject to approvals."], "original_id": "slb_spglobal"}, {"id": "RS072", "title": "S&P Global Completes Divestiture of Upstream Energy Software Portfolio to SLB", "url": "https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "publisher": "S&P Global", "date": "2026-09-01", "type": "company announcement", "locator": "Acquisition completion", "review_status": "full text reviewed", "supports": ["S&P Global announced completion of the upstream geoscience and petroleum engineering software portfolio divestiture to SLB on September 1, 2026.", "S&P Global said it would continue distributing its proprietary data through the divested workflow tools; financial terms were not disclosed."], "original_id": "slb_spglobal_completed"}], "field_sources": {"business_problem": ["RS071"], "workflow": ["RS071"], "technology_data": ["RS071"], "integration_human_role": ["RS071"], "deployment_scope": ["RS071"], "economics": ["RS071"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "services", "packet_id": "services-034", "source_ids": ["RS071", "RS072"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-107"]}, {"key": "hall-sekal-acquisition", "existing_ids": ["AI-111"], "related_existing_ids": ["AI-016"], "company": "Halliburton", "title": "Halliburton acquisition of Sekal drilling automation", "department": "Drilling & completions", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Acquired April 1, 2026; LOGIX/Sekal product integration announced", "practical_relevance": "Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.", "business_problem": "Expand drilling automation portfolio and combine directional drilling, wellbore hydraulics and rig equipment control.", "workflow": "Acquisition combines LOGIX remote operations and automation with Sekal DrillTronics autonomy platform and services. The companies had earlier deployed an automated on-bottom system for Equinor on the Norwegian Continental Shelf.", "technology_data": "LOGIX, DrillTronics and rig automation control system; model/control details described in pre-acquisition Equinor pilot source.", "integration_human_role": "System optimized within safe limits and enabled single-button parameter optimization in pilot; precise human override/control boundary is not fully specified.", "deployment_scope": "Acquisition completed; workflow shown in 2025 Equinor well deployment.", "partners": [], "disclosed_results": [], "economics": "Terms not disclosed in reviewed announcement.", "timeline": [{"date": "2025-02-26", "event": "Halliburton and Sekal reported integrated automated on-bottom drilling system for Equinor.", "source_ids": ["RS074"]}, {"date": "2026-04-01", "event": "Halliburton completed acquisition of Sekal from Sumitomo.", "source_ids": ["RS073"]}], "limitations": ["Acquisition terms not disclosed.", "Do not imply announced acquisition itself proves full portfolio rollout; Equinor deployment predates acquisition."], "assessment": "Analyst inference: Acquisition of technology with an operator deployment already demonstrated gives the portfolio a stronger practical basis than strategy alone, although broader post-deal scaling is not yet shown.", "sources": [{"id": "RS073", "title": "Halliburton Acquires Sekal AS From Sumitomo Corporation to Accelerate Drilling Automation Services", "url": "https://ir.halliburton.com/node/36266/pdf", "publisher": "Halliburton", "date": "2026-04-01", "type": "company announcement", "locator": "Acquisition and portfolio combination", "review_status": "full text reviewed", "supports": ["Halliburton acquired Sekal from Sumitomo and said it will combine LOGIX automation/remote operations with DrillTronics automation."], "original_id": "hall_sekal"}, {"id": "RS074", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "url": "https://ir.halliburton.com/node/35181/pdf", "publisher": "Halliburton", "date": "2025-02-26", "type": "partner case study", "locator": "Integrated system and Equinor deployment", "review_status": "full text reviewed", "supports": ["Halliburton, Sekal and rig automation delivered an integrated closed-loop automated on-bottom drilling system for an Equinor well in Norway."], "original_id": "hall_equinor"}], "field_sources": {"business_problem": ["RS073", "RS074"], "workflow": ["RS073", "RS074"], "technology_data": ["RS073", "RS074"], "integration_human_role": ["RS073", "RS074"], "deployment_scope": ["RS073", "RS074"], "economics": ["RS073", "RS074"]}, "verification": "verified primary", "corrections_to_existing": "Acquisition completed on 2026-04-01 per Halliburton primary source; the Equinor automated drilling deployment occurred in 2025 before the acquisition.", "cohort": "services", "packet_id": "services-035", "source_ids": ["RS073", "RS074"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-111"]}, {"key": "koc-ahmadi-innovation-valley", "existing_ids": [], "related_existing_ids": ["AI-014", "AI-066"], "company": "Kuwait Oil Company", "title": "Ahmadi Innovation Valley seven-year technology evaluation and deployment contract", "department": "IT, data & enterprise transformation", "secondary_departments": ["Production & field operations", "Reservoir engineering & field development"], "category": "Strategy only", "stage": "Seven-year contract awarded June 30, 2026; planned facility construction 2026 and opening 2028", "practical_relevance": "The program creates an evaluation-to-deployment channel spanning nearly 100 initiatives, but the source does not identify a named AI workflow that can yet be assessed as a deployment.", "business_problem": "Accelerate evaluation and deployment of solutions for KOC operational and strategic priorities.", "workflow": "SLB will support applied research, technology evaluation, testing, and deployment across a multi-domain project portfolio.", "technology_data": "AI is one of several domains; IIoT, production optimization, reservoir technologies, water management and energy transition are also included. No model or data integration is specified.", "integration_human_role": "KOC, operational teams, researchers and technology providers participate; the source emphasizes local capability and knowledge transfer.", "deployment_scope": "Seven-year KOC initiative with nearly 100 projects; dedicated facility planned for Kuwait. Specific field AI installation and scale undisclosed.", "partners": ["SLB", "KOC", "Industry, academic and technology providers (not individually named)"], "disclosed_results": [], "economics": "Contract value and AI-specific returns not disclosed.", "timeline": [{"date": "2026-06-30", "event": "KOC awarded SLB a seven-year Ahmadi Innovation Valley contract.", "source_ids": ["RS075"]}, {"date": "2026-2028 planned", "event": "Facility construction planned to start in 2026 and opening planned in 2028.", "source_ids": ["RS075"]}], "limitations": ["Broad multi-domain program, not evidence of a specific AI model or operating workflow.", "Facility dates are plans, not completed milestones.", "AI-specific economics and deployments not disclosed."], "assessment": "Analyst inference: This is an enabling research-to-deployment framework with substantial potential breadth, but it should not be counted as a proven AI deployment until named projects and outcomes emerge.", "sources": [{"id": "RS075", "title": "SLB Awarded Seven-Year Contract Under Kuwait Oil Company’s Ahmadi Innovation Valley Initiative", "url": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1", "publisher": "SLB", "date": "2026-06-30", "type": "company announcement", "locator": "Contract scope, key points", "review_status": "full text reviewed", "supports": ["KOC awarded SLB a seven-year AIV contract to evaluate/test/deploy technologies; scope spans nearly 100 projects across AI, IIoT, production optimization, reservoirs, water and energy transition.", "A dedicated facility is planned, with construction expected to start 2026 and opening in 2028; the announcement names no specific AI workflow or outcome."], "original_id": "koc_ahmadi"}], "field_sources": {"business_problem": ["RS075"], "workflow": ["RS075"], "technology_data": ["RS075"], "integration_human_role": ["RS075"], "deployment_scope": ["RS075"], "economics": ["RS075"]}, "verification": "verified primary", "corrections_to_existing": "New multi-domain KOC contract, related to but not an exact match for KOC seismic records AI-014/AI-066.", "cohort": "services", "packet_id": "services-036", "source_ids": ["RS075"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-162"]}, {"key": "blackstone-enverus-agreement", "existing_ids": ["AI-051"], "related_existing_ids": [], "company": "Enverus", "title": "Enverus acquisition and reported capital deployment", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Agreement announced August 2025; Enverus included in Q1 2026 capital deployed", "practical_relevance": "Ownership transaction provides context to the energy software provider; it is not itself an AI implementation.", "business_problem": "Blackstone announced acquisition of energy data and analytics company Enverus from Hellman & Friedman and Genstar.", "workflow": "Investment/acquisition transaction only; no operating AI workflow is evidenced by the deal announcement.", "technology_data": "Not disclosed for the transaction. Enverus software is described broadly as data analytics and insights.", "integration_human_role": "Not applicable to an AI workflow.", "deployment_scope": "Enverus customer footprint is described by the acquirer announcement, but no product deployment is detailed.", "partners": ["Hellman & Friedman", "Genstar"], "disclosed_results": [], "economics": "Financial terms were not disclosed in the reviewed announcement.", "timeline": [{"date": "2025-08-06", "event": "Blackstone funds announced agreement to acquire Enverus.", "source_ids": ["S30"]}, {"date": "2026-04-23", "event": "Q1 results name Enverus among private-equity capital deployed during the quarter.", "source_ids": ["S47"]}], "limitations": ["The earnings disclosure supports capital deployment, but does not supply the exact legal closing date or a standalone Enverus transaction value.", "Investment ownership does not establish operating benefits from AI."], "assessment": "Classify as capital/ownership context, separate from Enverus ONE product evidence.", "sources": [{"id": "S30", "title": "Blackstone announces agreement to acquire Enverus", "url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "publisher": "Blackstone", "date": "2025-08-06", "type": "company announcement", "locator": "Transaction announcement", "review_status": "full text reviewed", "supports": ["Blackstone funds agreed to acquire Enverus from Hellman & Friedman and Genstar; this is a financial transaction, not an AI deployment.", "Enverus provides energy data analytics and insights used by energy companies and investors."], "accessed": "2026-09-27", "original_id": "sc-001"}, {"id": "S47", "title": "Blackstone First Quarter 2026 Earnings Release", "url": "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf", "publisher": "Blackstone", "date": "2026-04-23", "type": "Earnings release", "locator": "PDF page 15 / printed page 11, Private Equity capital deployed", "review_status": "Primary PDF section reviewed", "supports": ["Enverus named among Q1 2026 capital deployed."], "accessed": "2026-09-27", "original_id": "bx-q126"}], "field_sources": {"business_problem": ["S30"], "workflow": ["S30"], "technology_data": ["S30"], "integration_human_role": [], "deployment_scope": ["S30"], "economics": ["S30"]}, "verification": "verified primary", "corrections_to_existing": "The 2025 agreement is supplemented with the Q1 2026 capital-deployment disclosure; neither the segment total nor media valuations are assigned as a verified Enverus purchase price.", "cohort": "software_capital", "packet_id": "software_capital-001", "source_ids": ["S30", "S47"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-051"]}, {"key": "enverus-one", "existing_ids": ["AI-065"], "related_existing_ids": ["AI-051"], "company": "Enverus", "title": "Enverus ONE: governed AI energy workflow platform", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched / available 2026-04-07", "practical_relevance": "Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.", "business_problem": "Energy teams evaluate AFEs and producing assets using fragmented ownership, engineering, production and economic evidence; site-siting also requires multi-factor screening.", "workflow": "Four product Flows at launch: AFE Evaluation ingests AFE, validates ownership and benchmarks offset economics; Current Production Valuation selects wells, loads production, forecasts and economics; Project Siting ranks parcels against criteria; QuickStart makes customer documents searchable through secure Q&A.", "technology_data": "Astra proprietary energy model is paired with general frontier models and Enverus datasets; private tenancy, role-based controls and audit trails.", "integration_human_role": "Outputs are decision support; governance and audit features support human verification. Details of review gates for each Flow are not disclosed.", "deployment_scope": "Product launched for energy customers; no named customer adoption or counts disclosed in launch.", "partners": ["Blackstone funds (owner transaction)", "Astra (Enverus model)"], "disclosed_results": [{"metric": "AFE evaluation duration", "value": "Weeks to hours (vendor claim)", "period": "Launch disclosure, 2026-04-07", "attribution": "Vendor-reported", "source_ids": ["S48"]}], "economics": "Vendor says AFE review can fall from weeks to hours; no customer-specific baseline or independently verified savings.", "timeline": [{"date": "2025-08-06", "event": "Blackstone funds announced agreement to acquire Enverus from H&F and Genstar.", "source_ids": ["S30"]}, {"date": "2026-04-07", "event": "Enverus ONE launched and was available, with four Flows.", "source_ids": ["S48"]}], "limitations": ["Acquisition is reported as an agreement; this source does not establish closing.", "No independent customer case, accuracy, scale, or realized ROI disclosed."], "assessment": "Concrete workflow packaging makes energy-specific tasks queryable and governed; evidence establishes launch and features, not downstream customer value. Blackstone ownership is kept separate from the AI product.", "sources": [{"id": "S48", "title": "Enverus ONE is live: governed AI platform built for energy industry", "url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "publisher": "Enverus", "date": "2026-04-07", "type": "company announcement", "locator": "Launch, platform design and four Flows", "review_status": "full text reviewed", "supports": ["Enverus ONE launched with Astra energy model paired with general frontier AI, proprietary energy datasets and governed access controls.", "AFE Evaluation, Current Production Valuation, Project Siting and QuickStart workflows are available.", "AFE evaluation automates ingestion, ownership validation, offset economics and comparisons; vendor says weeks can become hours."], "accessed": "2026-09-27", "original_id": "sc-002"}, {"id": "S30", "title": "Blackstone announces agreement to acquire Enverus", "url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "publisher": "Blackstone", "date": "2025-08-06", "type": "company announcement", "locator": "Transaction announcement", "review_status": "full text reviewed", "supports": ["Blackstone funds agreed to acquire Enverus from Hellman & Friedman and Genstar; this is a financial transaction, not an AI deployment.", "Enverus provides energy data analytics and insights used by energy companies and investors."], "accessed": "2026-09-27", "original_id": "sc-001"}], "field_sources": {"business_problem": ["S48"], "workflow": ["S48"], "technology_data": ["S48"], "integration_human_role": ["S48"], "deployment_scope": ["S48"], "economics": ["S48"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-002", "source_ids": ["S48", "S30"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-065"]}, {"key": "quantum-digital-mission", "existing_ids": ["AI-052"], "related_existing_ids": ["AI-053", "AI-054", "AI-055"], "company": "Quantum Capital Group", "title": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Internal adoption", "stage": "Operating platform described; undated page; 2024 report published 2025", "practical_relevance": "More operationally concrete than a generic investment thesis: it describes portfolio data infrastructure and specific well-level and field optimization workflows. The publication date of the 2024 ESG report is within window, but its accelerator operating date is not. No independent portfolio-level outcomes.", "business_problem": "Private equity asset teams and portfolio operators need common data access and repeatable, fast screening/operating decisions across varied energy assets.", "workflow": "Quantum Energy Cloud harmonizes company financial, operational and market data; QIQ answers questions, automates analysis and can trigger workflows. Named tasks span acquisition well economics, geospatial well/pad capital optimization, completion-design recommendations, lift optimization, base-production anomaly detection, downtime forecasting and compliance-data pipelines.", "technology_data": "Official page names Microsoft and Databricks for QEC. The ESG excerpt says QEC had 100+ AI accelerators but operating coverage is stated as of May 2024, outside the research window.", "integration_human_role": "Tool outputs support investment and operations teams; page implies workflow trigger/decision support but does not specify approvals, human sign-off or model controls.", "deployment_scope": "Firm-wide/portfolio system claim; operating companies and deployment numbers for each named workflow are not stated.", "partners": ["Microsoft", "Databricks"], "disclosed_results": [{"metric": "AI accelerators across portfolio companies", "value": "More than 100 (company-reported)", "period": "As of May 2024; outside research window", "attribution": "Company-reported", "source_ids": ["RS076"]}], "economics": "No workflow-specific ROI, spend, adoption or production lift quantified. ESG report accelerator count is an in-use company claim at pre-window period.", "timeline": [{"date": "undated", "event": "Digital Mission page describes QEC, QIQ and AI-at-Work examples; accessed 2026-09-27, page publication date not stated.", "source_ids": ["S31"]}, {"date": "2025-05-02", "event": "2024 ESG report published; accelerator/AI description includes as-of May 2024 operating scope.", "source_ids": ["RS076"]}], "limitations": ["Digital Mission page is undated; access date is not a publication date.", "ESG PDF could not be opened; excerpt review only.", "Do not treat accelerator deployments from May 2024 as in-window proof."], "assessment": "More operationally concrete than a generic investment thesis: it describes portfolio data infrastructure and specific well-level and field optimization workflows. The publication date of the 2024 ESG report is within window, but its accelerator operating date is not. No independent portfolio-level outcomes.", "sources": [{"id": "S31", "title": "Quantum Capital Group Digital Mission", "url": "https://www.quantumcap.com/digital-mission/", "publisher": "Quantum Capital Group", "date": "undated", "type": "company webpage", "locator": "QEC, QIQ and AI at Work", "review_status": "full text reviewed", "supports": ["Quantum Energy Cloud unifies portfolio financial, operational and market data using Microsoft and Databricks.", "QIQ is a proprietary generative AI interface that answers questions, automates analyses and initiates workflows over portfolio data.", "Named workflows include single-well economics, geospatial capital allocation, completion design, artificial lift, anomaly detection and downtime prediction."], "accessed": "2026-09-27", "original_id": "sc-003"}, {"id": "RS076", "title": "2024 ESG Report", "url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "publisher": "Quantum Capital Group", "date": "2025-05-02", "type": "annual report", "locator": "AI/data infrastructure section (search excerpt; PDF open failed)", "review_status": "excerpt reviewed", "supports": ["QEC and private generative AI QIQ are described; report claims over 100 AI accelerators used by portfolio companies.", "Operational as-of date in excerpt is May 2024, prior to research window; report publication itself is in-window."], "accessed": "2026-09-27", "original_id": "sc-004"}], "field_sources": {"business_problem": ["S31"], "workflow": ["S31"], "technology_data": ["S31", "RS076"], "integration_human_role": ["S31"], "deployment_scope": ["S31", "RS076"], "economics": ["RS076"]}, "verification": "verified primary", "corrections_to_existing": "AI-052 is the exact QEC match; QIQ and workflow initiatives are distinct profiles AI-053–AI-055 and are related, not exact matches to this umbrella record.", "cohort": "software_capital", "packet_id": "software_capital-003", "source_ids": ["S31", "RS076"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-052"]}, {"key": "novi-investment", "existing_ids": ["AI-122"], "related_existing_ids": [], "company": "Novi Labs", "title": "Invictus $35m growth investment in Novi Labs", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Investment announced 2025-06-25", "practical_relevance": "Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.", "business_problem": "Novi seeks capital to expand upstream energy intelligence, product work and acquisition strategy.", "workflow": "Invictus investment funds product innovation and expansion of a machine-learning and reservoir-science platform used in upstream forecasting and capital decisions. This is an investor financing event, not itself a new operator deployment.", "technology_data": "Novi says platform combines ML, proprietary data and reservoir science; architecture and incremental funded deliverables are not detailed.", "integration_human_role": "No operating customer workflow change or human review process disclosed in financing release.", "deployment_scope": "Company/vendor platform market-wide; no deployment counts tied to the financing.", "partners": ["Invictus Growth Partners"], "disclosed_results": [{"metric": "Growth investment", "value": "$35 million", "period": "2025-06-25 announcement", "attribution": "Company-reported", "source_ids": ["R095"]}], "economics": "$35m financing disclosed; use of proceeds is high-level; no return or AI-specific spend breakdown.", "timeline": [{"date": "2025-06-25", "event": "Novi announced $35m investment from Invictus Growth Partners.", "source_ids": ["R095"]}], "limitations": ["Funding announcement gives no customer-level implementation details."], "assessment": "Include as capital formation supporting an AI product company. It does not establish an Invictus portfolio operating deployment or isolate the funded dollars to a specific model.", "sources": [{"id": "R095", "title": "Novi Labs raises $35 million from Invictus Growth Partners", "url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "publisher": "Novi Labs", "date": "2025-06-25", "type": "company announcement", "locator": "Funding and platform overview", "review_status": "full text reviewed", "supports": ["Invictus Growth Partners invested $35 million to support Novi product innovation, expansion and acquisitions.", "Novi describes platform combining machine learning, proprietary datasets and reservoir science for upstream capital decisions."], "accessed": "2026-09-27", "original_id": "sc-005"}], "field_sources": {"business_problem": ["R095"], "workflow": ["R095"], "technology_data": ["R095"], "integration_human_role": [], "deployment_scope": ["R095"], "economics": ["R095"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-004", "source_ids": ["R095"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-122"]}, {"key": "novi-intelligence", "existing_ids": [], "related_existing_ids": ["AI-122"], "company": "Novi Labs", "title": "Novi Intelligence upstream research and forecasting", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched / unveiled 2025-09-10", "practical_relevance": "A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.", "business_problem": "Analysts and operators need comparable forecasts and transparent evidence about well quality, completions, depletion and local inventory when screening assets and allocating development capital.", "workflow": "Novi Intelligence surfaces forward-looking well forecasts and proprietary diagnostics (rock quality, completion intensity, prior depletion, well density); clients can inspect underlying sourced data to check and contextualize outputs.", "technology_data": "Machine-learning forecasts over operator-sourced datasets and proprietary derived metrics; specific model classes, validation design and error rates are not disclosed.", "integration_human_role": "Client can validate underlying data; human review is described at a high level, no formal approval workflow disclosed.", "deployment_scope": "Offered to Novi customers; no customer count, basin coverage or deployment dates disclosed in launch.", "partners": [], "disclosed_results": [], "economics": "No independently tested accuracy or realized economics. Vendor’s scale claims about capital decisions should not be read as causal savings.", "timeline": [{"date": "2025-09-10", "event": "Novi Intelligence unveiled.", "source_ids": ["RS077"]}], "limitations": ["No independent benchmark, named customer deployment or workflow economics."], "assessment": "A named decision-support product tied to upstream screening and forecasting. Transparency into source data may aid challengeability, but disclosed claims do not establish forecast accuracy.", "sources": [{"id": "RS077", "title": "Novi Labs unveils Novi Intelligence", "url": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "publisher": "Novi Labs", "date": "2025-09-10", "type": "company announcement", "locator": "Novi Intelligence product", "review_status": "full text reviewed", "supports": ["Novi Intelligence joins operator-sourced data, machine learning and forecasts with proprietary metrics such as rock quality, completion intensity, depletion and well density.", "Product exposes underlying data for client validation; accuracy claims are vendor claims."], "accessed": "2026-09-27", "original_id": "sc-006"}], "field_sources": {"business_problem": ["RS077"], "workflow": ["RS077"], "technology_data": ["RS077"], "integration_human_role": ["RS077"], "deployment_scope": ["RS077"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-005", "source_ids": ["RS077"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-163"]}, {"key": "novi-north-america-platform", "existing_ids": ["AI-123"], "related_existing_ids": [], "company": "Novi Labs", "title": "Acquisitions combine upstream AI forecasting with Canada and market intelligence", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Turing announced 2026-04-29; RBN announced 2026-05-28", "practical_relevance": "Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.", "business_problem": "Cross-border upstream evaluation often omits commodity pricing, transport capacity, basis and demand, making asset economics incomplete.", "workflow": "Novi acquired Turing Analytics (EVA) to extend analytical coverage into Canadian basins, followed by RBN Energy to add commodity/midstream research. Combined thesis connects well forecasts and inventory quality with transport, commodity flows, basis and downstream demand for investment/commercial decisions.", "technology_data": "Novi says its platform combines proprietary well-level data and ML production forecasting; acquisitions bring Canadian analytics and RBN market research/expertise. Integration and technical architecture across acquisitions not documented.", "integration_human_role": "Acquisition releases describe decision support for operators, investors and commercial teams; no automated approval or human governance details.", "deployment_scope": "North American energy market; Turing states client companies cover >50% Canadian production; RBN reach claims refer to research readership/users, not AI deployment.", "partners": ["Turing Analytics", "RBN Energy"], "disclosed_results": [], "economics": "Deal terms undisclosed. No evidence yet that post-acquisition integrated product has shipped or generated returns.", "timeline": [{"date": "2026-04-29", "event": "Novi announced acquisition of Turing Analytics/EVA.", "source_ids": ["RS078"]}, {"date": "2026-05-28", "event": "Novi announced acquisition of RBN Energy.", "source_ids": ["R096"]}], "limitations": ["Acquisition prices and integration status not disclosed.", "Turing and RBN are distinct acquisitions and do not evidence a single completed AI product workflow."], "assessment": "Strategically useful convergence of upstream forecasting and market context; this is an acquisition/integration thesis. Treat the joined platform as intended unless a later release documents actual integration.", "sources": [{"id": "RS078", "title": "Novi Labs acquires Turing Analytics", "url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "publisher": "Novi Labs", "date": "2026-04-29", "type": "company announcement", "locator": "Acquisition rationale and geographic coverage", "review_status": "excerpt reviewed", "supports": ["Novi acquired Turing Analytics/EVA to combine US and Canadian upstream datasets and analytics for investment evaluation.", "Novi says Turing clients represent over 50% of Canadian oil and gas production; company-reported."], "accessed": "2026-09-27", "original_id": "sc-007"}, {"id": "R096", "title": "Novi Labs acquires RBN Energy", "url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "publisher": "Novi Labs", "date": "2026-05-28", "type": "company announcement", "locator": "Combination and market workflows", "review_status": "full text reviewed", "supports": ["RBN acquisition combines Novi upstream AI forecasts with commodity, midstream and demand intelligence.", "Decision context includes basis, transport capacity, commodity flows and end-market pricing."], "accessed": "2026-09-27", "original_id": "sc-008"}], "field_sources": {"business_problem": ["R096"], "workflow": ["RS078", "R096"], "technology_data": ["RS078", "R096"], "integration_human_role": ["R096"], "deployment_scope": ["RS078", "R096"], "economics": ["RS078", "R096"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-006", "source_ids": ["RS078", "R096"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-123"]}, {"key": "quorum-streamba", "existing_ids": ["AI-056"], "related_existing_ids": [], "company": "Quorum Software", "title": "Streamba acquisition and VOR energy-logistics integration", "department": "Supply chain & logistics", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Acquired 2026-06-25; integration/agentic roadmap described", "practical_relevance": "Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.", "business_problem": "Offshore operators manage cargo, personnel, vessels and changing plans across disconnected logistics tools; disruptions force manual cross-team replanning.", "workflow": "Quorum added Streamba’s Visibility, Orchestration and Response product to DaWinci logistics for combined personnel/cargo planning and execution. VOR centralizes data and surfaces disruptions; Quorum describes agentic systems that can suggest/perform reassignment or plan updates, with human oversight.", "technology_data": "AI-native workflow orchestration is described, but release provides little model, training data or connectivity detail.", "integration_human_role": "Human oversight is explicit for autonomous response; detail on permissions and deployment gates not disclosed.", "deployment_scope": "Quorum customer product portfolio; this release does not name a customer rollout or specify installed base integration completion.", "partners": ["Streamba"], "disclosed_results": [], "economics": "Terms undisclosed; no operational ROI quantified.", "timeline": [{"date": "2026-06-25", "event": "Quorum announced Streamba acquisition to extend energy value-chain AI platform.", "source_ids": ["S32"]}], "limitations": ["No named live customer case, measured results or technical architecture detail."], "assessment": "Specific energy logistics use case with operational substance. Distinguish acquired capability and future agentic response from proven autonomous customer deployments.", "sources": [{"id": "S32", "title": "Quorum Software acquires Streamba", "url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "publisher": "Quorum Software", "date": "2026-06-25", "type": "company announcement", "locator": "Streamba/VOR and DaWinci", "review_status": "full text reviewed", "supports": ["Streamba AI-native supply chain platform and Visibility, Orchestration and Response system unify logistics data, workflows and teams.", "Quorum says agentic capabilities can identify disruption and propose or execute plan changes with human oversight."], "accessed": "2026-09-27", "original_id": "sc-009"}], "field_sources": {"business_problem": ["S32"], "workflow": ["S32"], "technology_data": ["S32"], "integration_human_role": ["S32"], "deployment_scope": ["S32"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-007", "source_ids": ["S32"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-056"]}, {"key": "pak-lease-ai", "existing_ids": ["AI-057"], "related_existing_ids": [], "company": "PakEnergy", "title": "Embedded AI lease provision processing", "department": "Land & legal", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched 2025-09-25", "practical_relevance": "Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.", "business_problem": "Land teams process lease provisions and related research manually, potentially missing key dates/terms or delaying workflows.", "workflow": "AI feature is embedded in PakEnergy Land Management to read/process lease provision work as part of existing land workflow.", "technology_data": "Announcement describes embedded AI but not model, source-document types, validation methods or integration specifics.", "integration_human_role": "Land staff remain users; no detailed review/exception process or accuracy rates given.", "deployment_scope": "Feature within PakEnergy Land Management; availability described by vendor, no customer/site count.", "partners": [], "disclosed_results": [{"metric": "Lease provision task time", "value": "Hours to seconds (vendor claim)", "period": "Launch announcement, 2025-09-25", "attribution": "Vendor-reported", "source_ids": ["S34"]}], "economics": "PakEnergy claims task duration falls from hours to seconds; no independent timing or cost validation.", "timeline": [{"date": "2025-09-25", "event": "PakEnergy announced embedded AI lease-provision capability.", "source_ids": ["S34"]}], "limitations": ["No model or accuracy disclosure; vendor speed claim only."], "assessment": "Practical document-heavy land workflow, but public disclosure is thin on controls and measured quality.", "sources": [{"id": "S34", "title": "PakEnergy launches embedded AI-powered lease provision capability", "url": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "publisher": "PakEnergy", "date": "2025-09-25", "type": "company announcement", "locator": "Land Management lease provision", "review_status": "full text reviewed", "supports": ["Embedded AI processes lease provision research and data inside PakEnergy Land Management.", "Company says work previously taking hours can be done in seconds; no accuracy or independent outcome data disclosed."], "accessed": "2026-09-27", "original_id": "sc-010"}], "field_sources": {"business_problem": ["S34"], "workflow": ["S34"], "technology_data": ["S34"], "integration_human_role": ["S34"], "deployment_scope": ["S34"], "economics": ["S34"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-008", "source_ids": ["S34"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-057"]}, {"key": "pak-capture", "existing_ids": ["AI-058"], "related_existing_ids": [], "company": "PakEnergy", "title": "PakCAPTURE ticket and receipt extraction", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched 2026-05-20", "practical_relevance": "Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.", "business_problem": "Field drivers and back-office teams rekey receipts/paper ticket details and need to associate documents with loads rapidly.", "workflow": "Driver photographs a receipt/ticket; PakAI extracts key fields, populates pickup/drop forms and links the record to a transportation load for office access.", "technology_data": "AI extraction embedded in PakTransportation; underlying model/OCR, confidence scoring and exception handling are not disclosed.", "integration_human_role": "Driver captures record; staff can use transportation system record; no explicit human validation details.", "deployment_scope": "Feature available to PakTransportation customers; no usage numbers disclosed.", "partners": [], "disclosed_results": [], "economics": "No quantified labor saving, accuracy, or customer economics disclosed.", "timeline": [{"date": "2026-05-20", "event": "PakEnergy announced PakCAPTURE.", "source_ids": ["RS079"]}], "limitations": ["No independent accuracy or economic results."], "assessment": "Specific field-to-office workflow embedded in an existing transportation product. Evidence supports launch features, not performance.", "sources": [{"id": "RS079", "title": "PakEnergy launches AI-powered Ticket-N-Receipt Management", "url": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "publisher": "PakEnergy", "date": "2026-05-20", "type": "company announcement", "locator": "PakCAPTURE workflow", "review_status": "full text reviewed", "supports": ["Drivers photograph receipts or paper tickets; PakAI extracts data, populates pickup/drop forms and associates information with a load.", "Feature embedded in PakTransportation and available to customers."], "accessed": "2026-09-27", "original_id": "sc-011"}], "field_sources": {"business_problem": ["RS079"], "workflow": ["RS079"], "technology_data": ["RS079"], "integration_human_role": ["RS079"], "deployment_scope": ["RS079"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-009", "source_ids": ["RS079"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-058"]}, {"key": "collide-battalion", "existing_ids": ["AI-126"], "related_existing_ids": ["AI-124", "AI-125"], "company": "Battalion Oil", "title": "Battalion strategic investment and planned Collide Operations System rollout", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Strategic investment and rollout announced 2026-09-08", "practical_relevance": "High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.", "business_problem": "Small operator has more than 100 TB across contracts, land records, production history and files; unstructured data impedes production and business analysis.", "workflow": "Battalion says initial rollout will build a queryable company model across records, then target gas marketing reconciliation, field invoice review, power use/hedging and well failure diagnosis. Riggs is planned for production engineers at about day 90. Battalion retains review/control; system subscription and strategic equity investment disclosed.", "technology_data": "Collide says it uses a connected asset model and source-cited answers with deterministic petroleum calculations; Battalion release does not confirm precise model stack/data ingestion or availability dates for all planned apps.", "integration_human_role": "Company expressly describes human review/control of outputs. Operators and production engineers are intended users.", "deployment_scope": "Single named customer, Battalion; >100TB is source-company description of input corpus, not total production deployment scale.", "partners": ["Battalion Oil"], "disclosed_results": [], "economics": "Subscription cost/equity terms and realized economics not stated. Announcement anticipates reduced analysis time; no specific baseline result independently verified.", "timeline": [{"date": "2026-09-08", "event": "Battalion announced strategic investment in Collide and planned deployment.", "source_ids": ["R103"]}], "limitations": ["Do not count Collide general product claims as verified Battalion implementation.", "Multiple apps are future plans; Riggs expected around day 90, after date of press release."], "assessment": "High practical specificity because a public operator names the corpus, intended use cases and staged introduction. As of the announcement, several workflows remain planned; do not state an established production outcome.", "sources": [{"id": "R103", "title": "Battalion Oil announces strategic investment in Collide", "url": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "publisher": "Battalion Oil Corporation", "date": "2026-09-08", "type": "company announcement", "locator": "Strategic investment, intended deployment and Riggs", "review_status": "full text reviewed", "supports": ["Battalion made strategic equity investment and will deploy Collide Operations System under subscription.", "Initial plan unites more than 100 TB of records, contracts and production history; planned use cases include gas reconciliation, field invoices, power, and failure diagnosis.", "Riggs is expected to be available to production engineers around day 90; human review/control retained."], "accessed": "2026-09-27", "original_id": "sc-012"}, {"id": "RS080", "title": "Collide product: Operations System", "url": "https://collide.io/product", "publisher": "Collide", "date": "undated", "type": "company webpage", "locator": "Architecture/capability page", "review_status": "excerpt reviewed", "supports": ["Collide describes petroleum-specific connected asset model and grounded answers with source citations.", "Company describes deterministic tested calculations and deployment in customer cloud/data platform; page undated."], "accessed": "2026-09-27", "original_id": "sc-013"}], "field_sources": {"business_problem": ["R103"], "workflow": ["R103"], "technology_data": ["R103", "RS080"], "integration_human_role": ["R103", "RS080"], "deployment_scope": ["R103"], "economics": ["R103"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-010", "source_ids": ["R103", "RS080"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-126"]}, {"key": "orbital-seed", "existing_ids": [], "related_existing_ids": ["AI-127"], "company": "Applied Computing", "title": "Orbital on-edge process-industry AI and seed funding", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "£9m seed announced 2025-05-28; company says one refinery live", "practical_relevance": "Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.", "business_problem": "Process engineers seek process recommendations using sensor, engineering and chemical context while protecting sensitive plant data and limiting cloud compute costs.", "workflow": "Orbital is presented as combining language models, time-series, physics and chemical-engineering models to help industrial teams analyze/optimize plant processes; deployment includes on-edge processing.", "technology_data": "Multi-model system combining language/time-series/physics/chemistry, edge deployment. Detailed data interfaces, controls and named refinery not provided.", "integration_human_role": "Human decision process, safety interlocks and scope of closed-loop authority not disclosed.", "deployment_scope": "Company says one large refinery is live; it is unidentified and unverified externally.", "partners": [], "disclosed_results": [{"metric": "Refinery live sites", "value": "One large refinery (company says; unnamed)", "period": "2025-05-28", "attribution": "Company-reported", "source_ids": ["RS081"]}, {"metric": "Cloud AI cost comparison", "value": "75% lower (company claim)", "period": "Not specified", "attribution": "Company-reported", "source_ids": ["RS081"]}, {"metric": "Payback", "value": "Under 10 weeks (company claim)", "period": "Not specified", "attribution": "Company-reported", "source_ids": ["RS081"]}], "economics": "£9m seed. Company claims 75% lower cost than cloud AI and sub-10-week payback; potential energy reductions are forward-looking, not audited.", "timeline": [{"date": "2025-05-28", "event": "Applied Computing announced £9m seed investment and Orbital overview.", "source_ids": ["RS081"]}], "limitations": ["AI-127 appears to be later $20m Series A; no primary funding source verified in this review.", "No refinery/customer named, no model audit or safety case."], "assessment": "Potentially important edge deployment for sensitive process data, but source evidence is a funding announcement and vendor claims. Do not assert independently verified savings or plant impact.", "sources": [{"id": "RS081", "title": "Applied Computing raises £9m seed to build Orbital AI", "url": "https://appliedcomputing.com/press-release-28052025", "publisher": "Applied Computing", "date": "2025-05-28", "type": "company announcement", "locator": "Orbital architecture, financing and initial refinery", "review_status": "full text reviewed", "supports": ["Applied Computing raised £9m seed funding for Orbital, a process-industry model using language, time-series, physics and chemical engineering models.", "Company says it is on edge, can work on sensitive process data, and reports one large refinery live.", "Cost savings, energy reduction and payback cited are company claims/potential, not independently supported."], "accessed": "2026-09-27", "original_id": "sc-014"}], "field_sources": {"business_problem": ["RS081"], "workflow": ["RS081"], "technology_data": ["RS081"], "integration_human_role": [], "deployment_scope": ["RS081"], "economics": ["RS081"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-011", "source_ids": ["RS081"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-164"]}, {"key": "cognite-atlas", "existing_ids": ["AI-120"], "related_existing_ids": ["AI-049"], "company": "Cognite", "title": "Atlas AI industrial agent workbench", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Commercial AI product", "stage": "Major release 2025-09-09", "practical_relevance": "Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.", "business_problem": "Industrial staff search siloed operational/engineering information and need faster cause analysis, work preparation and handover across assets.", "workflow": "Low-code agents use templates for root cause investigation, work package generation, cross-source research, handover summaries and work-order classification.", "technology_data": "Connects live OT/IT/engineering data using industrial knowledge graph with data governance and granular access; announcement describes architecture generally.", "integration_human_role": "Industrial users configure agents; role/access governance described, but customer-side approval policy per workflow not detailed.", "deployment_scope": "Product available across industrial customers; Cognite names Aker BP and others as customers but release does not tie each named customer to each template.", "partners": [], "disclosed_results": [], "economics": "No quantified benefit metrics in the announcement.", "timeline": [{"date": "2025-09-09", "event": "Cognite announced Atlas AI major release.", "source_ids": ["RS082"]}], "limitations": ["No customer-specific rollout metrics in reviewed release."], "assessment": "Industrial data context and named operational tasks make this a practical product initiative. Avoid implying every template is deployed at named oil and gas customers.", "sources": [{"id": "RS082", "title": "Cognite Atlas AI major release accelerates industrial AI", "url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "publisher": "Cognite", "date": "2025-09-09", "type": "company announcement", "locator": "Agent workbench, templates, governance", "review_status": "full text reviewed", "supports": ["Atlas AI provides low-code agent workbench over industrial OT/IT/engineering data.", "Templates cover root-cause analysis, work package generation, cross-source research, handover summaries and work-order classification.", "Cognite names energy customers; announcement is product update not independent evaluation of customer results."], "accessed": "2026-09-27", "original_id": "sc-015"}], "field_sources": {"business_problem": ["RS082"], "workflow": ["RS082"], "technology_data": ["RS082"], "integration_human_role": ["RS082"], "deployment_scope": ["RS082"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-012", "source_ids": ["RS082"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-120"]}, {"key": "seeq-intelligence", "existing_ids": ["AI-128"], "related_existing_ids": [], "company": "Seeq", "title": "Seeq Intelligence and Agent Q for operational decisions", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched 2026-03-03; available now", "practical_relevance": "Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.", "business_problem": "Industrial knowledge and event history are siloed, making it difficult for operations personnel to investigate deviations consistently and scale expert practice.", "workflow": "Agent Q assembles current process data, historical events, previous analyses, documents and expert context into traceable investigations and prioritized recommendations. Users can build multistep custom agents, scheduled or triggered, and connect agents to external systems for context or action.", "technology_data": "Seeq uses contextualized real-time operational data, documents and analytic steps; vendor describes secure extensibility and domain-aware AI but does not identify specific LLM or customer infrastructure in release.", "integration_human_role": "Designed to augment subject-matter experts; traceable output and integration enable review. Closed-loop action is technically supported in integrations but actual controls/customer approvals unspecified.", "deployment_scope": "Available Seeq Enterprise product; oil & gas among target sectors; no named O&G customer deployment for Agent Q in launch.", "partners": [], "disclosed_results": [], "economics": "No customer-specific value quantified.", "timeline": [{"date": "2026-03-03", "event": "Seeq Intelligence unveiled and made available.", "source_ids": ["RS083"]}], "limitations": ["No named energy customer case, implementation architecture or results in launch disclosure."], "assessment": "Clear workflow product with traceability, expert context and configurable integrations; however, the release establishes commercial availability, not a named energy deployment or realized gains.", "sources": [{"id": "RS083", "title": "Introducing Seeq Intelligence", "url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "publisher": "Seeq", "date": "2026-03-03", "type": "company announcement", "locator": "Agent Q, build-your-own agents and integrations", "review_status": "full text reviewed", "supports": ["Agent Q synthesizes real-time operational context, prior analyses, historical events, documents and domain expertise into investigations and recommended actions.", "Custom agents run multistep workflows on demand, scheduled or triggered; secure extensions can retrieve customer-system context and initiate workflows."], "accessed": "2026-09-27", "original_id": "sc-016"}], "field_sources": {"business_problem": ["RS083"], "workflow": ["RS083"], "technology_data": ["RS083"], "integration_human_role": ["RS083"], "deployment_scope": ["RS083"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-013", "source_ids": ["RS083"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-128"]}, {"key": "aveva-ai-connect", "existing_ids": ["AI-129"], "related_existing_ids": [], "company": "AVEVA", "title": "AVEVA industrial AI capabilities and CONNECT roadmap", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "Commercial AI product", "stage": "Announced 2026-05-20; some releases are future roadmap", "practical_relevance": "Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.", "business_problem": "Industrial organizations face costly manual data mapping and fragmented OT/IT information that impede governed AI deployment.", "workflow": "AVEVA announced data mapping assistance, industrial knowledge graph, CONNECT Flows and AI features in engineering/operations; AI can propose data mappings with lineage and governance. HMI/SCADA edge options discussed.", "technology_data": "CONNECT, Twin Builder, knowledge graph and connectors are named. Major CONNECT release with additional AI functionality was roadmap for Q1 2027, beyond the window; 2026 Flows capabilities need release-level qualification.", "integration_human_role": "Data governance/lineage emphasized; specific review roles and operational authority are not detailed.", "deployment_scope": "Commercial industrial platform/roadmap; no oil/gas-specific customer deployment identified in release.", "partners": [], "disclosed_results": [], "economics": "No deployment cost or customer economics disclosed.", "timeline": [{"date": "2026-05-20", "event": "AVEVA announced new AI capabilities at AVEVA World; major CONNECT release planned for Q1 2027.", "source_ids": ["R106"]}], "limitations": ["No named O&G deployment or verified operating results.", "A major CONNECT AI release is planned for Q1 2027 and is out of window."], "assessment": "Relevant industrial data infrastructure and engineering capabilities, but public release combines announced, demonstrated and future roadmap items. Only capabilities explicitly available by window end should be treated as deployed.", "sources": [{"id": "R106", "title": "AVEVA announces new capabilities to embed AI", "url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "publisher": "AVEVA", "date": "2026-05-20", "type": "company announcement", "locator": "CONNECT, Flows and AI roadmap", "review_status": "full text reviewed", "supports": ["AVEVA announced AI data mapping with lineage/governance, industrial knowledge graph and Flows/connectors.", "Press release labels major CONNECT release for Q1 2027 as planned; roadmap is beyond research window and must not be counted as deployed in-window."], "accessed": "2026-09-27", "original_id": "sc-017"}], "field_sources": {"business_problem": ["R106"], "workflow": ["R106"], "technology_data": ["R106"], "integration_human_role": ["R106"], "deployment_scope": ["R106"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-014", "source_ids": ["R106"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-129"]}, {"key": "baker-hughes-sourcing", "existing_ids": [], "related_existing_ids": ["AI-021"], "company": "Baker Hughes", "title": "AI sourcing optimization across procurement data", "department": "Supply chain & logistics", "secondary_departments": [], "category": "Internal adoption", "stage": "Vendor-reported deployment; undated case describes CY2025 benefits", "practical_relevance": "A specific procurement workflow connects fragmented spend records to sourcing decisions.", "business_problem": "Fragmented procurement data and manual spreadsheet reconciliation.", "workflow": "C3 AI Sourcing Optimization integrates procurement information for sourcing decisions.", "technology_data": "Procurement records; model architecture and input-data controls are not detailed.", "integration_human_role": "Procurement staff use the application; approval controls are not established by the reviewed excerpt.", "deployment_scope": "Baker Hughes; coverage figures below describe the vendor case.", "partners": ["C3 AI"], "disclosed_results": [{"metric": "SKU scope", "value": "900,000 SKUs", "period": "Undated full-scale case; CY2025 benefit", "attribution": "Vendor-reported", "source_ids": ["RS084"]}, {"metric": "Sites covered", "value": "800+ sites", "period": "Undated full-scale case", "attribution": "Vendor-reported", "source_ids": ["RS084"]}, {"metric": "Economic benefit", "value": "Millions (amount unspecified)", "period": "CY2025", "attribution": "Vendor-reported", "source_ids": ["RS084"]}], "economics": "Exact benefit amount, costs and attribution methodology are not disclosed.", "timeline": [{"date": "2025; CY2025 result in undated case page", "event": "C3 AI case page reports Baker Hughes sourcing optimization scope and benefit.", "source_ids": ["RS084"]}], "limitations": ["Vendor-published case; no customer-primary corroboration or independent assurance reviewed.", "The publication date is unknown. CY2025 describes the outcome period, not the launch date."], "assessment": "Data coverage is substantial, but a sourcing benefit should be tested against price movements, negotiated contracts and implementation costs before attributing it entirely to AI.", "sources": [{"id": "RS084", "title": "Baker Hughes achieves bottom line savings with AI Sourcing Optimization", "url": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "publisher": "C3 AI", "date": "undated; CY2025 economics", "type": "partner case study", "locator": "AI sourcing optimization; value and scale", "review_status": "excerpt reviewed", "supports": ["Case study says C3 AI Sourcing Optimization supported 900,000 SKUs across 800+ sites and realized millions in economic benefit in CY2025.", "Workflow addresses fragmented procurement data and manual spreadsheet reconciliation.", "Vendor-published customer story; not independent assurance."], "accessed": "2026-09-27", "original_id": "sc-024"}], "field_sources": {"business_problem": ["RS084"], "workflow": ["RS084"], "technology_data": ["RS084"], "integration_human_role": ["RS084"], "deployment_scope": ["RS084"], "economics": ["RS084"]}, "verification": "verified partner", "corrections_to_existing": "The sourcing workflow is separated from the AI-021 alliance renewal.", "cohort": "software_capital", "packet_id": "software_capital-015", "source_ids": ["RS084"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-165"]}, {"key": "snowflake-energy", "existing_ids": ["AI-119"], "related_existing_ids": [], "company": "Snowflake", "title": "Snowflake Energy Solutions data and partner offering", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Commercial AI product", "stage": "Launched 2026-01-27", "practical_relevance": "The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.", "business_problem": "Energy organizations must combine IT, operational technology and IoT data while managing specialized planning and analytics tasks.", "workflow": "Energy Solutions packages Snowflake data platform with more than 30 partner solutions; partners provide specialized workloads such as SAP operational/financial data links and spatial/grid analytics. This is a platform ecosystem; release does not specify a named oil/gas AI workflow deployed by an operator.", "technology_data": "Snowflake AI Data Cloud/data sharing and partner apps; release describes cross-domain data foundation, not a novel named O&G model.", "integration_human_role": "Customer organizations control access and partner applications; detailed AI oversight workflows not disclosed.", "deployment_scope": "Energy sector offering with named major customers, but no customer-specific production AI implementation described.", "partners": [], "disclosed_results": [], "economics": "No quantified customer economics from the launch.", "timeline": [{"date": "2026-01-27", "event": "Snowflake announced Energy Solutions.", "source_ids": ["R090"]}], "limitations": ["No specific named oil/gas workflow, production deployment or ROI in reviewed announcement."], "assessment": "The announcement is material data infrastructure and ecosystem positioning; avoid claiming a customer operational AI use solely because it is listed as a customer.", "sources": [{"id": "R090", "title": "Snowflake launches Energy Solutions", "url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "publisher": "Snowflake", "date": "2026-01-27", "type": "company announcement", "locator": "Energy Solutions, partner ecosystem, SAP and spatial data", "review_status": "full text reviewed", "supports": ["Energy Solutions is an industry offering combining Snowflake data platform and partner solutions to unify IT/OT/IoT data.", "Announcement names customers but does not substantiate a named oil and gas AI production deployment by them."], "accessed": "2026-09-27", "original_id": "sc-018"}], "field_sources": {"business_problem": ["R090"], "workflow": ["R090"], "technology_data": ["R090"], "integration_human_role": ["R090"], "deployment_scope": ["R090"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-016", "source_ids": ["R090"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-119"]}, {"key": "microsoft-encino-data-agent", "existing_ids": [], "related_existing_ids": ["AI-119"], "company": "Encino Energy", "title": "Encino production-data and invoice AI workflows using Microsoft Power Platform", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Documented 2025-04-29; production agent rollout described as in progress", "practical_relevance": "A concrete customer example linking Snowflake production data to Microsoft natural-language agents. Status must remain rollout/early exploration; several adjacent Power Apps savings are workflow automation, not AI.", "business_problem": "Field operators need quick answers across multi-region production tables and downtime histories without custom SQL and dashboard changes; invoice approvals also created email/rekey work.", "workflow": "Encino describes rolling out a Copilot Studio agent connected through Snowflake connector and Dataverse semantic search to answer production and downtime questions; an earlier agent analyzes activity logs. AI Builder extracts invoice fields with confidence levels and routes them to Teams approvals.", "technology_data": "Snowflake data moves through SQL/virtual tables to Dataverse; enhanced connector was beta at time of story. AI Builder document processing, Power Automate and Teams Approvals named.", "integration_human_role": "AI invoice extraction displays confidence and human approvers handle routing; data agents answer operator questions; story presents decision support, not autonomous operational control.", "deployment_scope": "Encino oil/gas operations; story covers production data architecture and invoice process, but agent stage is rollout/planning, not mature quantified results.", "partners": ["Microsoft", "Snowflake"], "disclosed_results": [{"metric": "Invoice processing administrative time", "value": "Hundreds of hours per year for one employee (customer statement)", "period": "As described 2025-04-29", "attribution": "Vendor-reported customer statement", "source_ids": ["RS085"]}], "economics": "AI invoice app reportedly saves one employee hundreds of hours/year; attribution is customer-reported in Microsoft case. Other Power Platform savings (e.g., winter plowing) are non-AI and excluded.", "timeline": [{"date": "2025-04-29", "event": "Microsoft published Encino Power Platform case; production agent rollout and invoice processing documented.", "source_ids": ["RS085"]}], "limitations": ["Microsoft is the case publisher; no independent audit.", "Do not attribute $1.2m winter maintenance savings to AI.", "Production agent is described as rolling out/planned, not as a measured realized result.", "The invoice app’s reported time savings are not disaggregated by AI extraction versus broader Power Platform workflow automation."], "assessment": "A Microsoft-published customer case names Copilot Studio data agents and AI Builder invoice extraction. The production-data agent is still rolling out/planned; the invoice app is in use, but its broader time-saving result is not attributable to AI alone.", "sources": [{"id": "RS085", "title": "Encino Energy streamlines operations with Microsoft Power Platform", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "publisher": "Microsoft", "date": "2025-04-29", "type": "partner case study", "locator": "Copilot Studio/data agent and invoice AI Builder details", "review_status": "full text reviewed", "supports": ["Encino plans direct Snowflake to Dataverse connector and is rolling out a Copilot Studio agent for questions across production data; a separate early prototype analyzes activity logs.", "Invoice app uses AI Builder document processing with confidence display and human approval routing.", "The same story describes many Power Platform apps and savings, most unrelated to AI; do not attribute the $1.2m winter maintenance savings to AI."], "accessed": "2026-09-27", "original_id": "RS085"}], "field_sources": {"business_problem": ["RS085"], "workflow": ["RS085"], "technology_data": ["RS085"], "integration_human_role": ["RS085"], "deployment_scope": ["RS085"], "economics": ["RS085"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-017", "source_ids": ["RS085"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-166"]}, {"key": "microsoft-totalenergies-agents", "existing_ids": [], "related_existing_ids": [], "company": "TotalEnergies", "title": "MARGE maintenance report classification and BuyerCompanion procurement agent", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Customer story dated 2025-10-22; MARGE operational, BuyerCompanion France operational", "practical_relevance": "The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.", "business_problem": "Maintenance intervention reports require manual fault-cause classification; procurement requests need supplier/specification/framework agreement guidance under purchasing rules.", "workflow": "MARGE analyses and classifies intervention reports; BuyerCompanion drafts specifications, identifies suppliers/frameworks and recommends compliant purchases up to €50,000. TotalEnergies also rolled out 30,000 Microsoft 365 Copilot licenses with training and agent committee governance.", "technology_data": "MARGE and BuyerCompanion use Microsoft Copilot Studio; Witivio supports BuyerCompanion. Underlying models/data sources not disclosed.", "integration_human_role": "Human decision-making retained; governance committee reviews value, architecture, cybersecurity, compliance, redundancy. BuyerCompanion produces recommendations, not stated to autonomously award purchases.", "deployment_scope": "Copilot rollout across about 30,000 employees; MARGE deployment scope not quantified; BuyerCompanion operational in France with potential wider staff reach.", "partners": ["Microsoft", "Witivio"], "disclosed_results": [{"metric": "Fault cause identification", "value": "25%-45% improvement (company-reported)", "period": "As of customer story, 2025-10-22", "attribution": "Vendor-reported customer statement", "source_ids": ["RS086"]}, {"metric": "Savings target for selected procurement", "value": "10% target, not realized result", "period": "Future target as of 2025-10-22", "attribution": "Company-reported target", "source_ids": ["RS086"]}], "economics": "TotalEnergies reports MARGE 25-45% improved fault-cause identification; BuyerCompanion targets 10% savings on certain purchases, not realized savings. Adoption metrics (70% recommending/using weekly) are self-reported.", "timeline": [{"date": "2025-10-22", "event": "Microsoft customer story described 30k licenses and named MARGE and BuyerCompanion agents.", "source_ids": ["RS086"]}], "limitations": ["Model and data provenance are not disclosed.", "Separate MARGE measured identification improvement from BuyerCompanion savings target."], "assessment": "The strongest Microsoft energy-agent example reviewed: a specific maintenance classification workflow and procurement assistant, with human governance and some measured/target metrics. Metrics come from the customer story, not independent study.", "sources": [{"id": "RS086", "title": "TotalEnergies from Microsoft 365 Copilot to Copilot Studio", "url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "publisher": "Microsoft", "date": "2025-10-22", "type": "partner case study", "locator": "MARGE and BuyerCompanion", "review_status": "full text reviewed", "supports": ["MARGE uses Copilot Studio to classify intervention reports; company reports 25-45% improvement in fault-cause identification.", "BuyerCompanion streamlines purchases up to €50,000, operational in France, with company target of 10% savings on certain purchases.", "30,000 Copilot licenses, training and internal agent-validation committee described."], "accessed": "2026-09-27", "original_id": "sc-020"}], "field_sources": {"business_problem": ["RS086"], "workflow": ["RS086"], "technology_data": ["RS086"], "integration_human_role": ["RS086"], "deployment_scope": ["RS086"], "economics": ["RS086"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-018", "source_ids": ["RS086"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-167"]}, {"key": "ambyint-encino-infinitypl", "existing_ids": [], "related_existing_ids": [], "company": "Encino Energy", "title": "InfinityPL autonomous plunger-lift optimization at Encino", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Case study published 2025-08-05; project timeframe stated as 2023 to present.", "practical_relevance": "Named AI-driven artificial-lift deployment at 344 platform wells; disclosed impacts remain vendor-reported.", "business_problem": "Limited field teams could not consistently tune Encino’s growing plunger-lift portfolio, especially late-life wells.", "workflow": "InfinityPL monitors wells, detects anomalies and autonomously adjusts setpoints; operators watch dashboards and intervene on exceptions.", "technology_data": "Vendor describes real-time surveillance, anomaly detection and autonomous setpoint control; model and sensor details are undisclosed.", "integration_human_role": "Operators monitor performance and intervene on complex cases; setpoints run in auto mode.", "deployment_scope": "Encino Appalachian Basin; case says 344 platform wells, project timeframe 2023–present; pilot/full-scale cohort sizes are not given.", "partners": ["Ambyint"], "disclosed_results": [{"metric": "Annual decline improvement", "value": "7% vs manually optimized pilot wells", "period": "Aug 2025 case snapshot", "attribution": "Vendor-reported/customer case study", "source_ids": ["RS087"]}, {"metric": "Venting events reduced", "value": "62%", "period": "Aug 2025 full-scale case snapshot", "attribution": "Vendor-reported/customer case study", "source_ids": ["RS087"]}], "economics": "No audited dollar savings or ROI disclosed.", "timeline": [{"date": "2023", "event": "Stated project start.", "source_ids": ["RS087"]}, {"date": "2025-08-05", "event": "Case reports deployment scale and outcomes.", "source_ids": ["RS087"]}], "limitations": ["Vendor case only; raw data, cohort sizes and validation method are not supplied.", "Pilot and full-scale metrics should not be combined."], "assessment": "Analyst inference: The workflow is a named, scaled closed-loop deployment. The source supports the use case and reports improvements, but cannot independently establish impact.", "sources": [{"id": "RS087", "title": "Infinity PL Drives Efficiency Gains for Encino Energy", "url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "publisher": "Ambyint", "date": "2025-08-05", "type": "partner case study", "locator": "Customer background; solutions; key outcomes", "review_status": "full text reviewed", "supports": ["Case reports 344 Appalachian Basin wells and 2023–present timeframe.", "InfinityPL uses surveillance, anomaly detection and autonomous setpoint control with operator monitoring/intervention.", "Vendor reports 7% decline improvement in pilot wells versus manual optimization and 62% fewer venting events; validation method is absent."], "accessed": "2026-09-27", "original_id": "RS087"}], "field_sources": {"business_problem": ["RS087"], "workflow": ["RS087"], "technology_data": ["RS087"], "integration_human_role": ["RS087"], "deployment_scope": ["RS087"], "economics": ["RS087"]}, "verification": "verified partner", "corrections_to_existing": "Source URL was previously only excerpt-reviewed. Reopened full case text adds 344-well scale and specific vendor-reported results. Metrics remain attributed and unverified independently.", "cohort": "software_capital", "packet_id": "software_capital-019", "source_ids": ["RS087"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-168"]}, {"key": "imubit-preem-fcc", "existing_ids": [], "related_existing_ids": [], "company": "Preem", "title": "Closed-loop AI optimization of Preem Lysekil FCC", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Partnership announced 2025-02-19; outcome not reported", "practical_relevance": "Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.", "business_problem": "Preem seeks improved FCC yield/efficiency and lower fuel use and emissions at its Lysekil refinery.", "workflow": "Imubit Optimizing Brain is to optimize FCC process variables, improve yield profile and stabilize fuel-gas balance; announcement anticipates lower emissions/fuel consumption.", "technology_data": "Imubit describes closed-loop AI optimization; release provides no variables/sensor details, model architecture, operator console, fallback controls or integration design.", "integration_human_role": "Closed-loop capability is stated; operating authority bounds and human approval are not disclosed.", "deployment_scope": "Single named refinery and FCC unit; partnership/deployment intent disclosed, not completion status or broad rollout.", "partners": ["Preem"], "disclosed_results": [], "economics": "No quantified savings or emissions results; outcomes are stated as intended.", "timeline": [{"date": "2025-02-19", "event": "Imubit and Preem announced Lysekil FCC partnership.", "source_ids": ["RS088"]}], "limitations": ["No post-startup result or deployment completion reported in source."], "assessment": "Specific process unit and control objective make this operationally relevant. Evidence establishes a partnership/intended deployment, not realized savings or emissions reductions.", "sources": [{"id": "RS088", "title": "Preem partners with Imubit for closed-loop AI optimization", "url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "publisher": "Imubit", "date": "2025-02-19", "type": "company announcement", "locator": "Lysekil FCC deployment", "review_status": "full text reviewed", "supports": ["Imubit and Preem announced Optimizing Brain at Lysekil refinery FCC to optimize process variables, yields and fuel-gas balance.", "Fuel savings and emissions reduction were anticipated; no measured post-deployment results were given."], "accessed": "2026-09-27", "original_id": "sc-022"}], "field_sources": {"business_problem": ["RS088"], "workflow": ["RS088"], "technology_data": ["RS088"], "integration_human_role": ["RS088"], "deployment_scope": ["RS088"], "economics": ["RS088"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-020", "source_ids": ["RS088"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-169"]}, {"key": "cognite-nvidia-celanese", "existing_ids": ["AI-121"], "related_existing_ids": ["AI-120"], "company": "Cognite", "title": "NVIDIA-accelerated time-series anomaly detection with Celanese", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Adjacent process industry (chemicals)", "stage": "Integration announced 2025-10-14; Celanese use case aims at anomaly-detection deployment", "practical_relevance": "Noisy continuous sensor streams can hide equipment anomalies and delay intervention.", "business_problem": "Industrial equipment produces continuous noisy readings and performance metrics with context scattered across operational information.", "workflow": "Initial joint use case analyzes real-time time-series data to identify critical anomalies and support proactive intervention before disruption.", "technology_data": "Cognite Atlas AI and Data Fusion integrated with NVIDIA accelerated computing and software; NV-Tesseract time-series foundation model family identified.", "integration_human_role": "Agents are described as producing traceable, trustworthy results in maintenance workflows; specific alert review and intervention authority are not disclosed.", "deployment_scope": "Named customer Celanese; specialty chemicals, not oil and gas. Cognite says Celanese has scaled other use cases to 50+ sites; release does not assign NVIDIA anomaly system to all sites.", "partners": ["NVIDIA", "Celanese"], "disclosed_results": [], "economics": "No customer-specific economics or false-positive metrics disclosed.", "timeline": [{"date": "2025-10-14", "event": "Cognite announced the accelerated integration and first Celanese time-series anomaly use case.", "source_ids": ["R094"]}], "limitations": ["Initial anomaly use case is a joint aim/use case; release does not disclose production scale or measured accuracy.", "Customer is Celanese, a chemicals manufacturer; treat as adjacent process-industry evidence, not oil-and-gas operator adoption."], "assessment": "Analyst inference: The named time-series anomaly workflow is technically concrete but is an adjacent chemicals example, not evidence of oil-and-gas operator deployment.", "sources": [{"id": "R094", "title": "Cognite and NVIDIA partner to advance industrial AI performance", "url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "publisher": "Cognite", "date": "2025-10-14", "type": "company announcement", "locator": "Celanese time-series anomaly detection", "review_status": "full text reviewed", "supports": ["Atlas AI/Data Fusion integrated with NVIDIA accelerated computing and software.", "Initial Celanese project targets real-time anomaly detection in industrial time-series data.", "Celanese reports over 50 global Cognite sites, but release does not say all sites use the NVIDIA model."], "accessed": "2026-09-27", "original_id": "R094"}], "field_sources": {"business_problem": ["R094"], "workflow": ["R094"], "technology_data": ["R094"], "integration_human_role": ["R094"], "deployment_scope": ["R094"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-021", "source_ids": ["R094"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-121"]}, {"key": "collide-seed-platform", "existing_ids": ["AI-124"], "related_existing_ids": ["AI-126"], "company": "Collide", "title": "Collide $5m seed and energy-specific GenAI knowledge platform", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Seed announced 2025-04-15; platform rollout underway", "practical_relevance": "Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.", "business_problem": "Petroleum engineers and energy teams need technical knowledge from authoritative papers, internal processes and practitioner Q&A without time-consuming manual search.", "workflow": "Collide raised seed financing to scale its RAG/LLM platform, which retrieves and synthesizes technical content into energy-focused cited answers. The original product evolved from a community/knowledge hub into an enterprise AI product.", "technology_data": "Retrieval-augmented generation and large language models over a curated knowledge base combining technical papers, internal processes and Q&A.", "integration_human_role": "Answers are citation-supported; announcement emphasizes engineer knowledge work, but no formal review gates or enterprise data controls specified.", "deployment_scope": "Company says it services public/private enterprise and PE-backed E&Ps; user metric combines community and platform users, so not enterprise deployments.", "partners": [], "disclosed_results": [{"metric": "Seed funding", "value": "$5 million", "period": "2025-04-15", "attribution": "Company-reported", "source_ids": ["R101"]}, {"metric": "Estimated petroleum-engineer time savings", "value": "25%-33%", "period": "Company estimate at announcement", "attribution": "Company-reported", "source_ids": ["R101"]}, {"metric": "Community users", "value": "6,000+ (not equivalent to AI enterprise users)", "period": "2025-04-15 announcement", "attribution": "Company-reported", "source_ids": ["R101"]}], "economics": "$5m seed led by Mercury Fund. Company estimates petroleum engineers save 25%-33% of time; not independent or customer-specific.", "timeline": [{"date": "2025-04-15", "event": "Collide $5m seed and energy-specific GenAI knowledge platform disclosed.", "source_ids": ["R101"]}], "limitations": ["Funding announcement conflates community footprint and enterprise service; it does not establish 6,000 AI users.", "25%-33% savings is vendor estimate with no study design disclosed."], "assessment": "Separate the seed/product disclosure from later Battalion investment and deployment. The core value proposition is cited energy knowledge retrieval, not autonomous field operation.", "sources": [{"id": "R101", "title": "Collide secures $5m to scale first GenAI platform for energy sector", "url": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "publisher": "Collide (via Business Wire)", "date": "2025-04-15", "type": "company announcement", "locator": "Seed funding and energy knowledge platform", "review_status": "full text reviewed", "supports": ["Collide announced a $5m seed led by Mercury Fund; proceeds support platform and engineering expansion.", "Platform uses retrieval augmented generation and LLMs over curated technical papers, internal processes and Q&A to provide cited, domain-specific energy answers.", "Collide reported 6,000+ community users and estimated 25%-33% time savings for petroleum engineers; these are company estimates and community users are not equivalent to deployed enterprise users."], "accessed": "2026-09-27", "original_id": "sc-033"}], "field_sources": {"business_problem": ["R101"], "workflow": ["R101"], "technology_data": ["R101"], "integration_human_role": ["R101"], "deployment_scope": ["R101"], "economics": ["R101"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "software_capital", "packet_id": "software_capital-022", "source_ids": ["R101"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-124"]}, {"key": "collide-winn-rrc", "existing_ids": ["AI-125"], "related_existing_ids": ["AI-124", "AI-126"], "company": "Winn Resources / Collide", "title": "Winn Resources Texas RRC W-10/G-10 filing workflow", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Customer workflow disclosed 2026-03-05", "practical_relevance": "Concrete operator workflow with a clear manual process and reported result. Keep the RIGGS/Winn workflow distinct from Collide’s platform launch and Battalion investment.", "business_problem": "Monthly regulatory filings require assembling well test and production data; missing values create delay and rework.", "workflow": "Collide supports monthly W-10/G-10 filings for Winn Resources, structuring existing operational records, detecting data gaps and preparing submissions.", "technology_data": "Collide says its AI platform works over customer operational data and regulatory form requirements; detailed extraction/model validation path not disclosed in accessible primary release.", "integration_human_role": "Customer says system does most of the preparation; human filing review/approval remains, but exact signoff flow is not fully described in customer source.", "deployment_scope": "Named operator Winn Resources; Texas wells; no well count or recurring months of deployment given.", "partners": [], "disclosed_results": [{"metric": "Monthly RRC filing prep time", "value": "Vendor headline: >95%; not reconciled to the stated 2–4-hour baseline and 20-minute cycle.", "period": "2026-03-05", "attribution": "Company-reported", "source_ids": ["RS089"]}], "economics": "The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence.", "timeline": [{"date": "2026-03-05", "event": "Winn Resources Texas RRC W-10/G-10 filing workflow disclosed.", "source_ids": ["RS089", "RS090"]}], "limitations": ["Customer source syndicated through PRNewswire/StreetInsider; no independent audit.", "Official Collide case-study body could not be extracted; title confirms named case but not detailed figures.", "29 September follow-up: The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence."], "assessment": "Concrete operator workflow with a clear manual process and reported result. Keep the RIGGS/Winn workflow distinct from Collide’s platform launch and Battalion investment.", "sources": [{"id": "RS089", "title": "Winn Resources partners with Collide to turn operational data into a competitive advantage", "url": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html", "publisher": "Collide release distributed through PRNewswire / StreetInsider", "date": "2026-03-05", "type": "company announcement", "locator": "W-10/G-10 filing automation", "review_status": "full text reviewed", "supports": ["Winn says Collide automates monthly Texas RRC W-10/G-10 filings using existing well test/production data; task moves from multi-hour effort to under 30 minutes.", "Winn says workflow surfaces missing-data gaps earlier and reports >95% time reduction. This is customer-reported; no independent audit."], "accessed": "2026-09-27", "original_id": "sc-034"}, {"id": "RS090", "title": "How Winn Resources Slashed Regulatory Filing Time by 95%", "url": "https://collide.io/case-studies/winn-resources-ai-regulatory-filing-oil-gas", "publisher": "Collide", "date": "undated", "type": "partner case study", "locator": "Case-study page title; body extraction unavailable", "review_status": "excerpt reviewed", "supports": ["Collide has an official named Winn Resources case study for regulatory-filing time reduction; accessible page confirms title only."], "accessed": "2026-09-27", "original_id": "sc-035"}], "field_sources": {"business_problem": ["RS089", "RS090"], "workflow": ["RS089", "RS090"], "technology_data": ["RS089", "RS090"], "integration_human_role": ["RS089", "RS090"], "deployment_scope": ["RS089", "RS090"], "economics": ["RS089", "RS090"]}, "verification": "verified partner", "corrections_to_existing": "Follow-up packet above resolves direct PRNewswire access, identifies the approximately 50-well example, and flags the time-reduction denominator problem.", "cohort": "software_capital", "packet_id": "software_capital-023", "source_ids": ["RS089", "RS090"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-125"]}, {"key": "williams-sequent-gas-marketing-ai", "existing_ids": ["AI-103"], "related_existing_ids": [], "company": "Williams", "title": "Sequent AI-assisted gas marketing test", "department": "Trading & marketing", "secondary_departments": [], "category": "Internal adoption", "stage": "Pilot / one-month test disclosed", "practical_relevance": "A narrowly described human-versus-model test in physical gas marketing illustrates a measurable decision-support experiment. Its practical value is possible improvement in how pipeline/storage capacity and local gas positions are marketed, while the disclosure does not say this test became a live execution workflow.", "business_problem": "Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs. The model score was compared with an experienced Sequent gas trader using a hindsight “perfect execution” scoring frame.", "workflow": "Williams said a young programmer/data-analytics team built models to market energy in Dallas–Fort Worth using weather, price, supply and demand as inputs", "technology_data": "AI models built by internal data analytics staff; specific model type, features, training data and infrastructure not disclosed.", "integration_human_role": "Experienced trader remained the benchmark and human execution authority is not explicitly described; the account is a retrospective anecdote from the CEO.", "deployment_scope": "One unique DFW asset, one month; company said it was a prior-year experiment. Broader deployment is not established.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "Not disclosed; no revenue uplift, reduced cost or customer outcome quantified.", "timeline": [{"date": "2025-08-05", "event": "CEO described prior-year one-month DFW test during Q2 2025 earnings call.", "source_ids": ["RS091"]}], "limitations": ["The test is described as occurring in “last year” relative to August 2025; it may fall inside or outside the 2024-09-26 start of the review window. Exact test date is not disclosed.", "This is secondary? The transcript is issuer-hosted and primary, so the claim is company-reported; score benchmark is not independently validated.", "No evidence that trading decisions became automated or that result generalized to other markets."], "assessment": "Concrete test with a crisp comparison, but small scope and hindsight benchmark limit inference; meaningful signal of commercial AI exploration, not proof of a scaled trading system.", "sources": [{"id": "RS091", "title": "Williams Q2 2025 earnings call transcript", "url": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "publisher": "Williams", "date": "2025-08-05", "type": "earnings call transcript", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Sequent gas-marketing AI anecdote; CEO describes one-month DFW test, model score 96% vs trader 94% against a 100% hindsight benchmark; narrow test only."], "original_id": "wmb25call"}, {"id": "RS092", "title": "Williams Q2 2025 earnings event page", "url": "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025", "publisher": "Williams", "date": "2025-08-05", "type": "investor relations event page", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Confirms issuer-hosted earnings materials including transcript and presentation."], "original_id": "wmb25ir"}], "field_sources": {"business_problem": ["RS091"], "workflow": ["RS091"], "technology_data": ["RS091"], "integration_human_role": ["RS091"], "deployment_scope": ["RS091"], "economics": ["RS091"]}, "verification": "verified primary", "corrections_to_existing": "Corrects existing secondary transcript attribution: Williams itself hosts the earnings transcript and reports one-month AI model score 96% versus trader 94% in a single DFW test; does not substantiate broad scaled trading or realized financial value.", "cohort": "midstream", "packet_id": "midstream-001", "source_ids": ["RS091", "RS092"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-103"]}, {"key": "tc-autonomous-pipeline-ml", "existing_ids": [], "related_existing_ids": [], "company": "TC Energy", "title": "Autonomous Pipeline commercial decision support", "department": "Trading & marketing", "secondary_departments": [], "category": "Internal adoption", "stage": "In use; scale and launch date not disclosed", "practical_relevance": "Supports daily/weekly gas transportation and Park-and-Loan (PAL) decisions by giving commercial teams real-time operational insight across assets.", "business_problem": "TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities.", "workflow": "TC Energy says the tool helps commercial staff identify, plan and execute daily/weekly asset strategies and capture incremental transportation and PAL opportunities", "technology_data": "Company identifies data analytics and machine learning; specific model, feeds, system architecture and measured accuracy are not disclosed.", "integration_human_role": "Commercial teams use the real-time insights to inform strategy and decisions; the report does not state autonomous execution.", "deployment_scope": "Company-wide commercial use is suggested by plural “commercial teams,” but assets, users, geographies and adoption rate are not disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "Incremental transportation/PAL opportunities are described as a value pathway; no quantified revenue, margin or cost outcome is reported.", "timeline": [{"date": "2025", "event": "2025 Sustainability Report describes the tool as an operational capability; launch date not given.", "source_ids": ["RS093"]}], "limitations": ["Report is primary but does not give publication month or initial deployment date.", "No financial attribution or evidence that model outputs directly execute trades.", "Distinguish from RAMP predictive maintenance: the report calls RAMP “advanced algorithms” but does not specify ML/AI, so it is not counted as an AI initiative."], "assessment": "A practical internal commercial workflow because it connects asset telemetry to recurring daily/weekly transport decisions; disclosed detail is thin on data inputs, realized uplift and governance.", "sources": [{"id": "RS093", "title": "TC Energy 2025 Report on Sustainability", "url": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "publisher": "TC Energy", "date": "2025", "type": "annual report", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Autonomous Pipeline analytics and machine learning for commercial daily/weekly asset strategies and transport/PAL opportunities; RAMP described as advanced algorithms only."], "original_id": "tc25"}, {"id": "RS094", "title": "TC Energy 2026 Management Information Circular", "url": "https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf", "publisher": "TC Energy", "date": "2026-03-05", "type": "proxy statement", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["AI governance, training, board oversight; governance does not evidence operating deployment."], "original_id": "tc26mic"}], "field_sources": {"business_problem": ["RS093"], "workflow": ["RS093"], "technology_data": ["RS093"], "integration_human_role": ["RS093"], "deployment_scope": ["RS093"], "economics": ["RS093"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-002", "source_ids": ["RS093", "RS094"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-170"]}, {"key": "tc-pipeline-iq", "existing_ids": [], "related_existing_ids": [], "company": "TC Energy", "title": "Pipeline IQ engineering-document verification agents", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Current partner-reported deployment; date of launch unknown", "practical_relevance": "Targets document-heavy pipeline construction and maintenance review: validating whether engineering records are complete and mutually consistent before human compliance review.", "business_problem": "PIPER compares engineering documentation and material test reports with weld maps and pressure-testing documentation, including heat-number traceability and completeness checks. SCOUT reportedly searches/discovers information across millions of pages. Engineers make the final compliance review.", "workflow": "PIPER compares engineering documentation and material test reports with weld maps and pressure-testing documentation, including heat-number traceability and completeness checks", "technology_data": "AWS describes an agentic platform built with Bedrock, Step Functions, Textract, Anthropic Claude through Bedrock, Strands Agents and AgentCore.", "integration_human_role": "Engineers review final outputs for compliance; AWS case says users get auditable traceability and control.", "deployment_scope": "Vendor case reports over 100 active users; company-wide, asset-specific breakdown and deployment start are not dated.", "partners": ["Amazon Web Services", "Anthropic"], "disclosed_results": [], "economics": "AWS/TC-reported 85% review-time reduction; up to $18.7m maintenance capital avoidance over five-year planning horizon is expected, not realized. Early tests on midsize documents up to 120 pages reportedly reached 95–100% accuracy.", "timeline": [{"date": "undated", "event": "AWS case study describes active users and current platform; publication and launch dates are not provided.", "source_ids": ["RS095"]}], "limitations": ["Partner case study is vendor-hosted and claims/results are partner-reported, not independent.", "Undated page and absent launch chronology prevent strict confirmation that rollout began within the 24-month window.", "Accuracy figure applies to early testing on midsize documents, not all live documents.", "Expected capital avoidance is a forecast, not realized savings."], "assessment": "Strong practical AI workflow and concrete engineering control, but implementation period is undated and all performance/economic metrics are vendor/partner-reported; do not count the metrics as independently verified.", "sources": [{"id": "RS095", "title": "TC Energy Pipeline IQ on AWS", "url": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "publisher": "Amazon Web Services", "date": "undated", "type": "partner case study", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["PIPER reviews engineering documentation/material test reports against weld maps and pressure-test docs; SCOUT discovery; >100 users, 85% review-time reduction, expected $18.7m five-year maintenance-capital avoidance. Page is undated, date/start of deployment not established."], "publication_date": "undated; not stated on page", "original_id": "tcaws"}], "field_sources": {"business_problem": ["RS095"], "workflow": ["RS095"], "technology_data": ["RS095"], "integration_human_role": ["RS095"], "deployment_scope": ["RS095"], "economics": ["RS095"]}, "verification": "verified partner", "corrections_to_existing": "Existing public evidence supports named agent workflows and vendor-reported usage/results; date of launch remains unverified.", "cohort": "midstream", "packet_id": "midstream-003", "source_ids": ["RS095"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-171"]}, {"key": "enbridge-ml-integrity-prioritization", "existing_ids": [], "related_existing_ids": [], "company": "Enbridge", "title": "Machine-learning prioritization of pipeline deformation features", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Development disclosed in 2025; subsequent operational use not established", "practical_relevance": "Helps integrity teams triage large volumes of in-line inspection and field observations to focus engineering review on potentially injurious deformations.", "business_problem": "The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation.", "workflow": "The 2025 report describes development of an ML tool to prioritize potentially injurious deformation features from ILI and field data; prioritized features proceed to engineering assessment and mitigation", "technology_data": "Machine learning applied to ILI/field inspection data; model design, thresholds and validation not disclosed.", "integration_human_role": "Engineers assess prioritized features and determine mitigation; model supports triage rather than replacing integrity decisions.", "deployment_scope": "Company pipeline integrity program; locations, asset coverage, production status and user count not disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No AI-attributable cost, risk reduction or integrity outcome quantified.", "timeline": [{"date": "2025", "event": "Sustainability report describes tool as under development.", "source_ids": ["RS096"]}, {"date": "2026", "event": "Customer handbook describes ML predictive analysis being used in complex defect assessment to prioritize repairs/resource allocation.", "source_ids": ["RS097"]}], "limitations": ["2025 status is development; 2026 handbook supports ML use in assessment but does not establish that it is the identical tool.", "No external validation, model performance or savings disclosed.", "Annual integrity spend is not attributable to this ML workflow."], "assessment": "Good operational fit because prediction narrows an established inspection-to-engineering workflow; evidence supports development and later general ML use, but public material does not link the same named model to measured outcomes.", "sources": [{"id": "RS096", "title": "Enbridge 2025 Sustainability Report", "url": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "publisher": "Enbridge", "date": "2025", "type": "sustainability report", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Describes developing machine-learning prioritization of potentially injurious deformation features from in-line inspection and field data; engineers assess and mitigate priority features."], "original_id": "enb25"}, {"id": "RS097", "title": "Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability", "url": "https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "publisher": "Enbridge", "date": "2026", "type": "company technical publication", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["States ML predictive analysis in complex pipeline defect assessment to prioritize repairs and allocate resources; no quantified outcome."], "original_id": "enb26"}, {"id": "RS098", "title": "Enbridge 2026 AI Policy", "url": "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152", "publisher": "Enbridge", "date": "2026", "type": "company policy", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Company policy describes AI governance and general AI uses; does not by itself date or quantify named operating workflow."], "original_id": "enbmic"}], "field_sources": {"business_problem": ["RS096", "RS097"], "workflow": ["RS096", "RS097"], "technology_data": ["RS096"], "integration_human_role": ["RS096", "RS097"], "deployment_scope": ["RS096", "RS097"], "economics": ["RS097"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-004", "source_ids": ["RS096", "RS097", "RS098"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-172"]}, {"key": "oneok-virtual-river-gauges", "existing_ids": [], "related_existing_ids": [], "company": "ONEOK", "title": "AI virtual river gauges for extreme-weather response", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployment reported; initial date not disclosed", "practical_relevance": "Aims to help pipeline operators locate where flooding or water-related exposure may threaten pipeline assets when physical river-gauge coverage is sparse.", "business_problem": "The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response.", "workflow": "The system estimates river conditions across a large network of locations; teams use the map to identify elevated risk and prioritize field assessment during extreme weather response", "technology_data": "ONEOK describes physics-based machine learning combining gravity, river force, pipeline characteristics and terrain/water information; provider and detailed data sources not named.", "integration_human_role": "AI output is a risk-triage layer followed by field assessment; no autonomous operational control described.", "deployment_scope": "Company reports 48,872 virtual gauges and interaction with over 100,000 terrain/water locations; in-service footprint and geographic distribution not described.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No avoided outage, response-time or financial metric disclosed.", "timeline": [{"date": "undated", "event": "Current innovation page reports deployed gauge count and weather-response use; initial deployment date not given.", "source_ids": ["RS099"]}], "limitations": ["Page is undated; accessed 2026-09-27 does not prove an in-window launch or update.", "Gauge count and method are company-reported, provider and validation absent.", "No disclosed operational outcomes or comparison to physical gauges."], "assessment": "One of the clearest upstream/midstream-adjacent practical AI cases, with explicit physics-informed method, scale and human triage; time-window status remains uncertain because only an undated current page was reviewed.", "sources": [{"id": "RS099", "title": "ONEOK Innovation Spotlight", "url": "https://sustainability.oneok.com/innovation-spotlight", "publisher": "ONEOK", "date": "undated", "type": "company sustainability webpage", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Reports AI-powered virtual river gauges and 48,872 deployed; weather-response risk triage. Undated page does not establish launch date or confirm in-window deployment."], "publication_date": "undated; not stated on page", "original_id": "oneok"}, {"id": "RS100", "title": "ONEOK Q4 2025 earnings release", "url": "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638", "publisher": "ONEOK", "date": "2026", "type": "earnings release", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Reviewed alongside company materials; no specific AI workflow found in release."], "original_id": "oneokearn"}], "field_sources": {"business_problem": ["RS099"], "workflow": ["RS099"], "technology_data": ["RS099"], "integration_human_role": ["RS099"], "deployment_scope": ["RS099"], "economics": ["RS099"]}, "verification": "verified primary", "corrections_to_existing": "Existing in-window deployment date could not be verified; do not imply the 48,872 gauge count is a 2026 increase.", "cohort": "midstream", "packet_id": "midstream-005", "source_ids": ["RS099", "RS100"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-173"]}, {"key": "p66-ml-turnaround-planning", "existing_ids": ["AI-100"], "related_existing_ids": [], "company": "Phillips 66", "title": "Machine learning in refinery turnaround planning", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "In use; launch date not disclosed", "practical_relevance": "Supports disciplined turnaround work planning so more maintenance can be prepared outside the outage window, potentially reducing duration and lost throughput.", "business_problem": "CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact. Management referenced lowering turnaround forecast mid-year and beating it, but did not attribute the forecast outcome to ML.", "workflow": "CEO said tools such as ML help move work outside the turnaround and plan/execute more effectively, shortening duration and financial impact", "technology_data": "Machine learning is named; training data, models, software, scope of work and integration not disclosed.", "integration_human_role": "Turnaround planners and operating/maintenance teams use tools in planning; no autonomous work approvals or workforce changes disclosed.", "deployment_scope": "Company turnaround program; no refinery count, units or adoption coverage disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No savings attributable to ML. Management cited lower turnaround forecast/spend and shorter-duration goals without a quantified AI-specific bridge.", "timeline": [{"date": "2026-02-04", "event": "CEO discussed ML use on Q4 2025 earnings call.", "source_ids": ["R063"]}], "limitations": ["Company-wide turnaround financial performance cannot be attributed specifically to ML.", "No model details, rollout date, refinery coverage or measured before/after result disclosed."], "assessment": "This is practical industrial decision support with human planners in the loop, but public evidence supports the use case only at a high level; no causal performance claim should be made.", "sources": [{"id": "R063", "title": "Phillips 66 Q4 2025 earnings call transcript", "url": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "publisher": "Phillips 66", "date": "2026-02-04", "type": "earnings call transcript", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["CEO describes machine learning tools in turnaround planning, moving work outside turnaround to shorten duration/financial impact; company says forecast/spend improved, does not isolate AI causation or savings."], "original_id": "p66q425"}], "field_sources": {"business_problem": ["R063"], "workflow": ["R063"], "technology_data": ["R063"], "integration_human_role": ["R063"], "deployment_scope": ["R063"], "economics": ["R063"]}, "verification": "verified primary", "corrections_to_existing": "Original existing claim that tools including ML were used is confirmed by issuer transcript. The earnings/spend improvement is not proven to have been caused by AI.", "cohort": "midstream", "packet_id": "midstream-006", "source_ids": ["R063"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-100"]}, {"key": "p66-company-ai-program", "existing_ids": ["AI-101"], "related_existing_ids": [], "company": "Phillips 66", "title": "Company AI program and frontline solution replication", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Broad program described as active; use cases scaling", "practical_relevance": "Creates a repeatable path for frontline engineers to identify data-heavy operating problems and adapt solutions across functions rather than relying only on isolated pilots.", "business_problem": "CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years. Refining leadership describes finding opportunities in a large set of cost-out work.", "workflow": "CEO described dedicated program staff finding and copying use cases across the organization; engineers capture/extract data and deploy solutions in days or weeks rather than months/years", "technology_data": "AI program includes AI and ML; specific LLM/model/vendor/data architecture not stated. Company notes ML has long been used for maintenance/turnarounds.", "integration_human_role": "Frontline engineers are central to discovering, building and deploying solutions; program staff support use-case reuse. Human-centered framing; no autonomous operating authority disclosed.", "deployment_scope": "Company-wide aspiration and multiple functions described, but no count of use cases, facilities, users or deployed models disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No quantified AI-specific savings; references to cost-out projects are not equivalent to AI-attributed savings.", "timeline": [{"date": "2026-08", "event": "CEO and refining leadership described AI normalization and frontline adoption on Q2 2026 earnings call.", "source_ids": ["R064"]}], "limitations": ["Broad program description, without use-case inventory or adoption metric.", "“Days and weeks” is a qualitative deployment-speed comparison, not independently measured performance.", "Cost-out portfolio performance is not attributed to AI."], "assessment": "Enterprise enablement and frontline participation are meaningful deployment signals, but absent case-level evidence the program itself is an adoption architecture, not proof of realized value.", "sources": [{"id": "R064", "title": "Phillips 66 Q2 2026 earnings call transcript", "url": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "publisher": "Phillips 66", "date": "2026-08", "type": "earnings call transcript", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["CEO describes broad AI program, copying use cases, frontline engineers extracting data and deploying solutions more quickly; leadership says ML long used for maintenance/turnarounds."], "original_id": "p66q226"}], "field_sources": {"business_problem": ["R064"], "workflow": ["R064"], "technology_data": ["R064"], "integration_human_role": ["R064"], "deployment_scope": ["R064"], "economics": ["R064"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-007", "source_ids": ["R064"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-101"]}, {"key": "valero-early-ml-applications", "existing_ids": ["AI-102"], "related_existing_ids": [], "company": "Valero Energy", "title": "Early AI/ML applications and operational-data foundation", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "A few applications reported; scale unclear", "practical_relevance": "A data foundation and a few selectively deployed applications could support operational availability, but public details do not identify the specific decisions or refinery systems affected.", "business_problem": "On the Q3 2025 call, the refining SVP reportedly said the company was evaluating uses for tangible value and had deployed techniques in a few applications; he emphasized high-quality standardized operating data as a prerequisite.", "workflow": "On the Q3 2025 call, the refining SVP reportedly said the company was evaluating uses for tangible value and had deployed techniques in a few applications; he emphasized high-quality standardized operating data as a prerequisite", "technology_data": "Machine learning/artificial intelligence named; data standards are described as foundational, with no architecture or model detail.", "integration_human_role": "Operational teams are implied users; transcript does not clarify roles, control boundaries or human review.", "deployment_scope": "A few unspecified applications; no site or business-unit scope disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No outcome or economics quantified.", "timeline": [{"date": "2025-10-23", "event": "Q3 earnings call discussion captured in a third-party transcript; issuer event page confirms call date.", "source_ids": ["R065", "RS101"]}], "limitations": ["Transcript is third-party hosted, not a company-primary transcript; attributed wording is secondary-only.", "Official webcast page confirms event, not the reported AI claims.", "Existing details about availability, model scope and data history have not been verified against official transcript/audio."], "assessment": "The company appears cautious and data-readiness-focused, but this profile remains a secondary-transcript lead until the original webcast/audio is checked; do not elevate the reported “few applications” into a verified deployment footprint.", "sources": [{"id": "R065", "title": "Valero Q3 2025 earnings call transcript (The Motley Fool)", "url": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "publisher": "The Motley Fool", "date": "2025-10-23", "type": "secondary transcript", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["SVP says a few AI/ML applications deployed or evaluated for tangible value; stresses operational-data quality. Not issuer-hosted; primary webcast page separately confirms call date."], "original_id": "vlo"}, {"id": "RS101", "title": "Valero Q3 2025 earnings call event", "url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "publisher": "Valero", "date": "2025-10-23", "type": "investor relations event page", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Confirms official call event, not transcript content."], "original_id": "vloir"}], "field_sources": {"business_problem": ["R065"], "workflow": ["R065"], "technology_data": ["R065"], "integration_human_role": ["R065"], "deployment_scope": ["R065"], "economics": ["R065"]}, "verification": "secondary only", "corrections_to_existing": "Do not treat AI-102 as primary verified. Earnings transcript source is third party; issuer event page only confirms call date.", "cohort": "midstream", "packet_id": "midstream-008", "source_ids": ["R065", "RS101"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-102"]}, {"key": "kmi-flyscan-investment", "existing_ids": ["AI-104"], "related_existing_ids": [], "company": "Kinder Morgan", "title": "Strategic investment in Flyscan AI-assisted aerial inspection", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Investment or acquisition", "stage": "Strategic investor; operational KMI rollout not established", "practical_relevance": "The investment backs development of remote-sensing inspection intended to detect leaks and right-of-way encroachment from routine aircraft patrols. It does not establish that KMI has deployed the system across its own pipeline network.", "business_problem": "Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas. KMI joined the strategic investor group in January 2025 and described the technology positively.", "workflow": "Flyscan says its image-analysis system automates visual patrol from ordinary patrol aircraft and is expanding leak detection into natural gas", "technology_data": "Vendor describes remote sensing and AI image analysis; sensors/model specifics and independent validation not disclosed.", "integration_human_role": "Inspection outputs would support follow-up investigation; KMI’s exact review or response workflow is not disclosed.", "deployment_scope": "Partnership/investment is confirmed; locations, miles, flights or active KMI use are not disclosed.", "partners": ["Flyscan Systems", "Marathon Petroleum", "Enbridge", "Hatch"], "disclosed_results": [], "economics": "Investment amount and economic return not disclosed.", "timeline": [{"date": "2025-01-14", "event": "Flyscan announced Kinder Morgan joined its strategic investor group.", "source_ids": ["RS102"]}], "limitations": ["Source is Flyscan’s own announcement; primary evidence for the partnership but vendor claims about product capability are not independent.", "KM’s Q1 2025 earnings release reviewed without further AI rollout disclosure.", "Do not conflate Flyscan’s overall fleet/operator metrics with KMI usage."], "assessment": "A concrete adjacent technology investment, but not evidence of operational deployment by KMI or quantified leak-detection improvements.", "sources": [{"id": "RS102", "title": "Flyscan announces Kinder Morgan as strategic investor", "url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "publisher": "Flyscan Systems", "date": "2025-01-14", "type": "partner announcement", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["KMI joins strategic investor group; vendor describes remote sensing and AI visual patrol/leak detection. This confirms investment/partnership, not KMI rollout."], "original_id": "fly"}, {"id": "RS103", "title": "Kinder Morgan Q1 2025 earnings release", "url": "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx", "publisher": "Kinder Morgan", "date": "2025-04-16", "type": "earnings release", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Reviewed as in-window operating/financial source; no Flyscan or internal AI deployment disclosed in this release."], "original_id": "kmq125"}], "field_sources": {"business_problem": ["RS102"], "workflow": ["RS102"], "technology_data": ["RS102"], "integration_human_role": ["RS102"], "deployment_scope": ["RS102"], "economics": ["RS102"]}, "verification": "verified partner", "corrections_to_existing": "Existing record’s investment/partnership claim is supported; operational adoption and natural-gas rollout remain unverified.", "cohort": "midstream", "packet_id": "midstream-009", "source_ids": ["RS102", "RS103"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-104"]}, {"key": "et-cloudburst-ai-campus", "existing_ids": ["AI-062"], "related_existing_ids": [], "company": "Energy Transfer", "title": "CloudBurst San Marcos natural-gas supply for AI campus", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Agreement signed; subject to customer final investment decision; projected operation not confirmation", "practical_relevance": "Energy Transfer supplies fuel infrastructure to a proposed behind-the-meter power campus serving AI/data-center load; this is an AI demand exposure, not ET’s internal AI adoption.", "business_problem": "Company agreed to supply up to 450,000 MMBtu/day firm gas to a campus outside San Marcos, Texas, supporting up to approximately 1.2 GW of direct behind-the-meter power, subject to CloudBurst FID.", "workflow": "Company agreed to supply up to 450,000 MMBtu/day firm gas to a campus outside San Marcos, Texas, supporting up to approximately 1", "technology_data": "Conventional pipeline gas supply and power generation arrangements; no AI technology is being adopted by Energy Transfer.", "integration_human_role": "CloudBurst/data-center operator and power-generation parties would use the delivered gas; ET’s disclosed role is supply.", "deployment_scope": "A single proposed campus; long-term term and projected start are in company announcement; operations depend on FID.", "partners": ["CloudBurst"], "disclosed_results": [], "economics": "No AI-related returns disclosed; gas volumes/contract economics not publicly quantified beyond delivery ceiling.", "timeline": [{"date": "2025-02-10", "event": "Firm supply agreement announced for proposed campus.", "source_ids": ["S41"]}], "limitations": ["Infrastructure that powers AI demand; category is deliberately separate from operating AI.", "Agreement is conditional on customer FID; announced volumes and projected date are not actual deliveries."], "assessment": "A material gas-demand contract with direct data-center linkage, but it is conventional energy infrastructure exposure rather than internal AI adoption.", "sources": [{"id": "S41", "title": "Energy Transfer and CloudBurst sign natural gas supply agreement", "url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "publisher": "Energy Transfer", "date": "2025-02-10", "type": "company announcement", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Up to 450,000 MMBtu/day firm gas supply for campus near San Marcos, supporting up to ~1.2 GW behind-the-meter power, contingent on customer FID; AI-demand infrastructure."], "original_id": "etcloud"}], "field_sources": {"business_problem": ["S41"], "workflow": ["S41"], "technology_data": ["S41"], "integration_human_role": ["S41"], "deployment_scope": ["S41"], "economics": ["S41"]}, "verification": "verified primary", "corrections_to_existing": "Correct category is AI-demand infrastructure; no operating AI technology workflow.", "cohort": "midstream", "packet_id": "midstream-010", "source_ids": ["S41"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-062"]}, {"key": "et-meta-ai-campus", "existing_ids": [], "related_existing_ids": ["AI-062"], "company": "Energy Transfer", "title": "Long-term gas transportation for Meta Louisiana data center", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Agreement announced; facilities planned for future service", "practical_relevance": "Dedicated pipeline transportation to a planned data-center project reflects gas infrastructure demand from AI/data-center expansion, not ET internal AI.", "business_problem": "Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service.", "workflow": "Energy Transfer and Entergy Louisiana announced firm transport capacity on the Tiger Pipeline system to support a future Meta data-center site, including a new lateral and long-term service", "technology_data": "Gas transportation assets and contract; no AI methods or digital operations described.", "integration_human_role": "Energy Transfer provides transport; Entergy serves the end customer.", "deployment_scope": "Future site and lateral construction; disclosed target service date 2028, outside or beyond in-window completion.", "partners": ["Entergy Louisiana", "Meta Platforms"], "disclosed_results": [], "economics": "Contract terms are not fully disclosed beyond service volume/term in company release; no AI-derived economics.", "timeline": [{"date": "2025-11-04", "event": "Long-term service agreement announced.", "source_ids": ["RS104"]}], "limitations": ["Company says the project supports AI/data-center demand; no evidence Energy Transfer runs the data-center AI or has operating AI use.", "Future service, not current deliveries."], "assessment": "Separate infrastructure case to preserve demand exposure without overstating it as AI adoption.", "sources": [{"id": "RS104", "title": "Entergy Louisiana and Energy Transfer sign agreement supporting Meta data center", "url": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "publisher": "Energy Transfer", "date": "2025-11-04", "type": "company announcement", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Long-term gas transportation infrastructure for future Meta site; separate AI-demand category, not internal AI."], "original_id": "etmeta"}], "field_sources": {"business_problem": ["RS104"], "workflow": ["RS104"], "technology_data": ["RS104"], "integration_human_role": ["RS104"], "deployment_scope": ["RS104"], "economics": ["RS104"]}, "verification": "verified primary", "corrections_to_existing": "Not the same specific CloudBurst campus; exact existing AI-062 only maps CloudBurst so this separate Meta disclosure is related, not exact same initiative.", "cohort": "midstream", "packet_id": "midstream-011", "source_ids": ["RS104"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-174"]}, {"key": "et-nexus-ai-campus", "existing_ids": [], "related_existing_ids": [], "company": "Energy Transfer", "title": "Nexus Hubbard Campus gas transport for behind-the-meter AI campus", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Agreement reported in Q1 2026 update; operating status not disclosed", "practical_relevance": "Gas transportation agreement supports a planned AI campus and on-site generation, creating an infrastructure link to compute demand rather than an internal AI workflow.", "business_problem": "Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation.", "workflow": "Q1 2026 company results say Energy Transfer entered long-term firm transport agreements through its Texas intrastate system to support Nexus Hubbard Campus and its on-site generation", "technology_data": "Natural-gas transportation; no AI model or operational-AI system described.", "integration_human_role": "ET is pipeline service provider; Nexus campus uses gas for power.", "deployment_scope": "Planned central Texas campus; volume, term and service date not reviewed in detail in the earnings update.", "partners": ["Nexus Data Centers"], "disclosed_results": [], "economics": "Financial contribution not disclosed.", "timeline": [{"date": "2026", "event": "Q1 2026 earnings update reports transport agreements.", "source_ids": ["RS105"]}], "limitations": ["The source confirms an agreement but detailed project timing and volumes are not disclosed in the reviewed passage.", "AI demand does not indicate internal use of AI by ET."], "assessment": "Track as commercial infrastructure opportunity only; details are less complete than the CloudBurst case.", "sources": [{"id": "RS105", "title": "Energy Transfer Q1 2026 results and updates", "url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "publisher": "Energy Transfer", "date": "2026", "type": "earnings release", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Reports agreements to transport gas to Nexus Hubbard Campus, an AI data center campus with on-site gas generation; infrastructure supporting AI demand."], "original_id": "etnexus"}], "field_sources": {"business_problem": ["RS105"], "workflow": ["RS105"], "technology_data": ["RS105"], "integration_human_role": ["RS105"], "deployment_scope": ["RS105"], "economics": ["RS105"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-012", "source_ids": ["RS105"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-175"]}, {"key": "williams-socrates-ai-demand", "existing_ids": ["AI-064"], "related_existing_ids": [], "company": "Williams", "title": "Socrates gas-to-power infrastructure for AI demand", "department": "Commercial, power & infrastructure", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Under construction in 2025; progress presented in 2026", "practical_relevance": "Williams is commercializing integrated gas supply, pipeline and generation capacity for data-center demand; it is not an AI operating workflow at Williams.", "business_problem": "Company reported breaking ground on the $1.6bn Socrates project to serve AI demand; 2026 presentation describes gas supply, pipeline and generation components and phased capacity/service milestones.", "workflow": "Company reported breaking ground on the $1", "technology_data": "Natural gas infrastructure and power generation; AI is the customer demand driver, not deployed technology.", "integration_human_role": "Williams develops and operates the energy assets; data-center customer consumes power.", "deployment_scope": "Project-level infrastructure; 2026 deck reports South phase in service and North expected later in 2026 (planned).", "partners": ["Blackstone"], "disclosed_results": [], "economics": "Project investment announced at $1.6bn; not an AI savings metric.", "timeline": [{"date": "2025-08-04", "event": "Earnings release said construction began on Socrates.", "source_ids": ["RS106"]}, {"date": "2026-08-04", "event": "Investor presentation updated project phases/status.", "source_ids": ["S43"]}], "limitations": ["Project capex and contracted economics belong to infrastructure underwriting.", "Do not classify as internal AI use or infer that every phase is operational."], "assessment": "High materiality infrastructure build responding to compute demand; its economics should be analyzed as power/gas infrastructure rather than AI productivity.", "sources": [{"id": "RS106", "title": "Williams Q2 2025 results", "url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "publisher": "Williams", "date": "2025-08-04", "type": "earnings release", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Describes $1.6bn Socrates power innovation project to serve growing AI demand; infrastructure, not internal AI."], "original_id": "wmsocrates"}, {"id": "S43", "title": "Williams Q2 2026 earnings presentation", "url": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "publisher": "Williams", "date": "2026-08-04", "type": "investor presentation", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Project details include gas supply, pipeline and generation; presented as power infrastructure for data-center load."], "original_id": "wms26pres"}], "field_sources": {"business_problem": ["RS106"], "workflow": ["RS106", "S43"], "technology_data": ["RS106"], "integration_human_role": ["S43"], "deployment_scope": ["RS106", "S43"], "economics": ["RS106"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-013", "source_ids": ["RS106", "S43"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-064"]}, {"key": "ecp-texas-hyperscale-campus", "existing_ids": ["AI-059"], "related_existing_ids": [], "company": "ECP", "title": "ECP/KKR hyperscale data-center campus co-located with Calpine power", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "AI-demand infrastructure", "stage": "Announced development; expected operations Q4 2026", "practical_relevance": "ECP’s role is investment in energy-adjacent data-center infrastructure; co-location with generation addresses power availability for hyperscale compute.", "business_problem": "ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations.", "workflow": "ECP, KKR and CyrusOne announced a 190MW Texas campus co-located with a Calpine gas plant; release says 144MW initial IT capacity and anticipated Q4 2026 operations", "technology_data": "Data-center campus, grid/power supply and natural gas generation; no internal AI tooling disclosed.", "integration_human_role": "ECP is financial sponsor; CyrusOne develops/operates data center; Calpine supplies generation.", "deployment_scope": "Single campus under development; planned capacity and schedule are forward-looking.", "partners": ["KKR", "CyrusOne", "Calpine"], "disclosed_results": [], "economics": "Reported total campus investment approaching $4bn; not AI adoption savings or ECP operating economics.", "timeline": [{"date": "2025-07-30", "event": "Joint campus development announced.", "source_ids": ["RS107"]}], "limitations": ["Announcement is company/parent primary and forward-looking; development and operating schedule may change.", "AI-demand exposure is separate from AI adoption by ECP."], "assessment": "Material infrastructure exposure tied to hyperscale/AI demand; maintain it as distinct investment category and do not present as ECP internal AI.", "sources": [{"id": "RS107", "title": "ECP and KKR announce Texas hyperscale data-center campus", "url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "publisher": "Bridgepoint Group / ECP", "date": "2025-07-30", "type": "company announcement", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["ECP/KKR/CyrusOne 190MW campus co-located with Calpine generation; AI-demand infrastructure."], "original_id": "ecp"}], "field_sources": {"business_problem": ["RS107"], "workflow": ["RS107"], "technology_data": ["RS107"], "integration_human_role": ["RS107"], "deployment_scope": ["RS107"], "economics": ["RS107"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-014", "source_ids": ["RS107"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-059"]}, {"key": "targa-ai-principal-engineer-role", "existing_ids": [], "related_existing_ids": [], "company": "Targa Resources", "title": "AI and emerging-technology operational deployment role", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Strategy only", "stage": "Recruiting / capability-building signal", "practical_relevance": "Shows an organizational role intended to evaluate, pilot and integrate AI into asset-heavy operating work, but does not establish that any particular AI workflow has been deployed.", "business_problem": "Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance.", "workflow": "Role description spans enterprise asset management, reliability, rotating equipment, mechanical integrity, corrosion, SCADA, pipeline integrity, operations, engineering and commercial workflows; responsibilities include selection, pilot testing, evaluation, integration, adoption and maintenance", "technology_data": "Posting requests AI/ML deployment experience and familiarity with operational systems; no named vendor/model/data source disclosed.", "integration_human_role": "Role coordinates with operations, reliability, engineering and commercial stakeholders; hiring intent is the only verified implementation signal.", "deployment_scope": "One role/job posting across the described organization; no rollout count or facilities disclosed.", "partners": ["None disclosed"], "disclosed_results": [], "economics": "No economics reported.", "timeline": [{"date": "2026-05-07", "event": "Job posting published.", "source_ids": ["RS108"]}], "limitations": ["Job listing is a strategy/workforce signal, not a deployed initiative.", "No named use case or operating result."], "assessment": "Useful leading indicator of planned operational AI capability, but insufficient evidence to claim production adoption.", "sources": [{"id": "RS108", "title": "Targa Principal Engineer, AI & Emerging Technology job posting", "url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "publisher": "Targa Resources", "date": "2026-05-07", "type": "job posting", "locator": null, "review_status": "full text reviewed", "accessed": "2026-09-27", "supports": ["Role coordinates AI/technology selection, pilots and adoption across operations/reliability/SCADA/integrity/commercial; intent/hiring signal, not proof of deployment."], "original_id": "targa"}], "field_sources": {"business_problem": ["RS108"], "workflow": ["RS108"], "technology_data": ["RS108"], "integration_human_role": ["RS108"], "deployment_scope": ["RS108"], "economics": ["RS108"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "midstream", "packet_id": "midstream-015", "source_ids": ["RS108"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-176"]}, {"key": "aramco-portfolio", "existing_ids": ["AI-038"], "company": "Aramco", "title": "Industrial AI deployment portfolio", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "secondary_departments": [], "stage": "Company-reported deployed portfolio; additional development pipeline", "verification": "verified primary", "partners": [], "sources": [{"id": "S22", "title": "Aramco deploys AI and supercomputing to drive digital transformation", "url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "date": "2025-11-24", "locator": "AI deployment and NVIDIA sections", "publisher": "Aramco", "type": "Company announcement / remarks", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "AR1"}], "field_sources": {"business_problem": ["S22"], "workflow": ["S22"], "technology_data": ["S22"], "integration_human_role": ["S22"], "deployment_scope": ["S22"], "economics": ["S22"]}, "disclosed_results": [{"metric": "AI-driven Technology Realized Value", "value": "$1.8 billion", "period": "2024", "attribution": "Aramco-reported portfolio measure", "source_ids": ["S22"]}], "timeline": [{"date": "2025-11-24", "event": "Portfolio scale disclosed.", "source_ids": ["S22"]}], "limitations": ["No public workflow-level cost or benefit bridge in this announcement.", "Do not add the portfolio value to the benefits of constituent initiatives."], "economics": "Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.", "practical_relevance": "Portfolio context for assessing scale; individual workflow outcomes must be evaluated separately.", "business_problem": "Downtime, maintenance failures and production efficiency.", "workflow": "AI models support facility monitoring, maintenance prediction and production optimization.", "technology_data": "More than 10 billion data points daily; NVIDIA GPU computing supports drilling plans and well placement.", "integration_human_role": "A commitment to train more than 6,000 AI developers supports adoption. Application-level approval controls are not detailed.", "deployment_scope": "Late-2025 company portfolio: 442 identified use cases, over 200 deployed solutions and more than 100 in development.", "assessment": "Broad organizational ambition. These counts are not equivalent to unique economic projects, and do not establish deployment depth at every facility.", "corrections_to_existing": "Preserves the portfolio separately from the Fadhili process-control workflow.", "cohort": "root_backfill", "packet_id": "root_backfill-001", "source_ids": ["S22"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-038"], "related_existing_ids": []}, {"key": "aramco-trv", "existing_ids": ["AI-037"], "company": "Aramco", "title": "AI contribution to Technology Realized Value", "category": "Internal adoption", "department": "Finance, investment & strategy", "secondary_departments": [], "stage": "Reported FY 2025 portfolio outcome", "verification": "verified primary", "partners": [], "sources": [{"id": "S21", "title": "CEO Opening Remarks from the 2025 Full Year Financial Results Media Call", "url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "date": "2026-03-10", "locator": "Technology Realized Value paragraph", "publisher": "Aramco", "type": "Company announcement / remarks", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "S21"}, {"id": "RS109", "title": "FY 2025 results presentation and script", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "date": "2026-03-10", "locator": "Slide 10; Technology Realized Value", "publisher": "Aramco", "type": "Investor presentation", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "RS109"}], "field_sources": {"business_problem": ["S21", "RS109"], "workflow": ["S21", "RS109"], "technology_data": ["S21", "RS109"], "integration_human_role": ["S21", "RS109"], "deployment_scope": ["S21", "RS109"], "economics": ["S21", "RS109"]}, "disclosed_results": [{"metric": "Total TRV", "value": "$5.3 billion", "period": "2025", "attribution": "Aramco; AI, digital and other technology combined", "source_ids": ["RS109"]}, {"metric": "TRV composition", "value": "$3.9bn capital avoidance; $1.0bn operating-cost reduction; $0.4bn revenue enhancement", "period": "2025", "attribution": "Aramco earnings presentation; not AI-only", "source_ids": ["RS109"]}, {"metric": "AI share", "value": "Over half in CEO media remarks; around half in the presentation", "period": "2025", "attribution": "Aramco; different published formulations", "source_ids": ["S21", "RS109"]}], "timeline": [{"date": "2026-03-10", "event": "FY 2025 results disclosed; annual TRV target of $3–5bn for 2026 onward is prospective.", "source_ids": ["RS109"]}], "limitations": ["Capital avoidance is not the same as incremental cash income.", "Third-party validation is described by the issuer; no independent assurance opinion was reviewed for this report.", "AI attribution is management-reported at aggregate level (“over half” of 2025 TRV); no project-level allocation or independent validation is disclosed."], "economics": "Aramco reports $5.3bn total 2025 Technology Realized Value and the CEO says over half was enabled by AI. This is company-reported aggregate attribution; the $3.9bn capital avoidance, $1.0bn operating-cost reduction and $0.4bn revenue enhancement are mixed-technology components, not an AI-only breakdown. Project-level validation, counterfactuals and payback are not disclosed.", "practical_relevance": "A portfolio-level measurement framework for technology value, rather than a standalone operating application.", "business_problem": "Assessing value across a broad technology portfolio.", "workflow": "The earnings materials aggregate capital avoided, operating costs reduced and revenue enhanced.", "technology_data": "Includes AI, digital and other technologies; this is not an AI-only financial statement line.", "integration_human_role": "Aramco describes external validation of TRV. The assurance work itself was not reviewed here.", "deployment_scope": "Company portfolio; no allocation to specific AI workflows is established.", "assessment": "Useful evidence of management measurement discipline. A reader should request the counterfactual baseline, validation methodology and allocation by project before using TRV in an investment model.", "corrections_to_existing": "CEO media remarks say over half; the earnings presentation says around half. Preserve both formulations, rather than present an exact AI-only dollar amount.", "cohort": "root_backfill", "packet_id": "root_backfill-002", "source_ids": ["S21", "RS109"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-037"], "related_existing_ids": []}, {"key": "aramco-humain", "existing_ids": ["AI-039"], "company": "Aramco", "title": "Proposed minority investment in HUMAIN", "category": "Investment or acquisition", "department": "Finance, investment & strategy", "secondary_departments": [], "stage": "Planned investment; no closing established in reviewed evidence", "verification": "verified primary", "partners": [], "sources": [{"id": "RS110", "title": "Aramco announces fourth quarter and full-year 2025 results", "url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "date": "2026-03-10", "locator": "HUMAIN investment plans; TRV highlights", "publisher": "Aramco", "type": "Company announcement / remarks", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "AR6"}], "field_sources": {"business_problem": ["RS110"], "workflow": ["RS110"], "technology_data": ["RS110"], "integration_human_role": ["RS110"], "deployment_scope": ["RS110"], "economics": ["RS110"]}, "disclosed_results": [], "timeline": [{"date": "2026-03-10", "event": "Investment plans reiterated in annual results.", "source_ids": ["RS110"]}], "limitations": ["No verified closing date or operating outcome in the reviewed release."], "economics": "Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.", "practical_relevance": "Strategic AI ecosystem investment, separate from field adoption.", "business_problem": "Expanding access to AI capabilities through ownership.", "workflow": "FY 2025 results describe plans progressing toward a significant minority interest in HUMAIN.", "technology_data": "The release does not disclose application architecture.", "integration_human_role": "Post-investment operating responsibilities are not established.", "deployment_scope": "Proposed corporate investment; transaction terms and ownership percentage are not established.", "assessment": "Potentially ambitious ownership strategy. The announcement provides no basis for treating transaction value as oilfield AI deployment spend.", "corrections_to_existing": "Retain planned status; the cited FY 2025 update does not establish closing.", "cohort": "root_backfill", "packet_id": "root_backfill-003", "source_ids": ["RS110"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-039"], "related_existing_ids": []}, {"key": "aramco-microsoft", "existing_ids": ["AI-070"], "company": "Aramco", "title": "Microsoft industrial AI and digital-talent MoU", "category": "Strategy only", "department": "IT, data & enterprise transformation", "secondary_departments": [], "stage": "Non-binding memorandum; exploratory collaboration", "verification": "verified primary", "partners": ["Microsoft"], "sources": [{"id": "R003", "title": "Aramco signs MoU with Microsoft to help advance industrial AI and digital talent transformation", "url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "date": "2026-02-12", "locator": "MoU scope and four focus areas", "publisher": "Aramco", "type": "Company announcement / remarks", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "AR4"}], "field_sources": {"business_problem": ["R003"], "workflow": ["R003"], "technology_data": ["R003"], "integration_human_role": ["R003"], "deployment_scope": ["R003"], "economics": ["R003"]}, "disclosed_results": [], "timeline": [{"date": "2026-02-12", "event": "Non-binding MoU signed.", "source_ids": ["R003"]}], "limitations": ["No implementation budget, production-user count or measured result disclosed."], "economics": "Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.", "practical_relevance": "Enablement program whose future operating deployments require separate evidence.", "business_problem": "Scaling industrial AI with data-residency and governance requirements.", "workflow": "The parties will explore Azure industrial solutions, cloud sovereignty, infrastructure optimization, an industry alliance and co-developed commercial AI systems.", "technology_data": "Microsoft Azure and potential sovereign controls; solution architecture not disclosed.", "integration_human_role": "Proposed skills work covers AI engineering, cybersecurity, data governance and product management.", "deployment_scope": "Aramco operations and the Saudi industrial ecosystem are intended areas of focus. A named production rollout is not established.", "assessment": "Broad scope, early commitment level. Track binding projects, named workflows, deployment milestones and user adoption before treating it as realized industrial transformation.", "corrections_to_existing": "Makes the non-binding and exploratory status explicit.", "cohort": "root_backfill", "packet_id": "root_backfill-004", "source_ids": ["R003"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-070"], "related_existing_ids": []}, {"key": "aramco-metabrain-plants", "existing_ids": [], "company": "Aramco", "title": "MetaBrain for Plants process-modeling system", "category": "Internal adoption", "department": "Refining, LNG & process operations", "secondary_departments": [], "stage": "Introduced at LEAP 2025; rollout scale undisclosed", "verification": "verified primary", "partners": [], "sources": [{"id": "RS111", "title": "Remarks by Ahmad O. Al-Khowaiter at LEAP 2025", "url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "date": "2025-02-09", "locator": "MetaBrain and MetaBrain for Plants paragraphs", "publisher": "Aramco", "type": "Company announcement / remarks", "review_status": "Full primary text reviewed", "supports": [], "accessed": "2026-09-27", "original_id": "AR5"}], "field_sources": {"business_problem": ["RS111"], "workflow": ["RS111"], "technology_data": ["RS111"], "integration_human_role": ["RS111"], "deployment_scope": ["RS111"], "economics": ["RS111"]}, "disclosed_results": [], "timeline": [{"date": "2025-02-09", "event": "MetaBrain for Plants introduced in executive remarks.", "source_ids": ["RS111"]}], "limitations": ["No measured accuracy, economic impact or independently evaluated results disclosed.", "Capability statements do not establish plant-wide autonomous control."], "economics": "Project-specific cost, incremental cash flow and payback are not established by the reviewed evidence.", "practical_relevance": "Reusable industrial models could reduce the effort required to build plant-specific analytics.", "business_problem": "Interpreting process behavior and reducing repeated model-development work.", "workflow": "A time-series transformer models plant processes to inform operators, engineers and scientists. Intended tasks include demand anticipation, process optimization and product-quality prediction.", "technology_data": "Large time-series datasets; distinct from the company’s industrial language model.", "integration_human_role": "Experts use model insights. The remarks do not establish autonomous execution or safety-system integration.", "deployment_scope": "No named plant, user count or deployment schedule disclosed.", "assessment": "Ambitious model reuse across process operations, with limited public implementation evidence. A pilot-to-production study would be needed to evaluate transferability across units.", "corrections_to_existing": "New workflow; related to the broader AI portfolio rather than a replacement for its record.", "related_existing_ids": ["AI-038"], "cohort": "root_backfill", "packet_id": "root_backfill-005", "source_ids": ["RS111"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-177"]}, {"key": "chevron-apolo", "existing_ids": ["AI-001"], "related_existing_ids": ["AI-002"], "company": "Chevron", "title": "APOLO prospectivity and development optimization", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "Launched Aug 2024 (pre-window); substantive Investor Day update 2025-11-12", "practical_relevance": "Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.", "business_problem": "Traditional type curves produce coarse, binary prospectivity views from limited variables, and development choices interact across spacing, completion and timing.", "workflow": "APOLO analyzes more than 50,000 wells to map prospectivity at finer resolution and recovery variation. Chevron was training it to optimize well spacing, completion design and development timing; decision support spans exploration screening and shale/tight development.", "technology_data": "Multivariate machine learning on well data; Chevron does not name training/validation design or integration details.", "integration_human_role": "Subsurface and development specialists use mapped candidates; company says still training/scaling. Human review/approval is not specified in disclosure.", "deployment_scope": "Chevron Permian/shale and tight use described; scale across portfolio is a goal. This is an in-window update about a model launched Aug 2024, so launch predates window.", "partners": [], "disclosed_results": [], "economics": "No quantified recovery uplift. Management’s broad “billions” annual AI value by decade-end is potential, not APOLO realized savings.", "timeline": [{"date": "2025-11-12", "event": "APOLO prospectivity and development optimization documented or disclosed.", "source_ids": ["S01"]}], "limitations": ["Model launch date Aug 2024 is before window; Nov 2025 investor day is qualifying update.", "No attributable APOLO production uplift or ROI disclosed."], "assessment": "Practical upstream planning tool with material geoscience scale and specific development levers. Evidence describes active model training and scaling, not proven recovery uplift.", "sources": [{"id": "S01", "title": "2025 Chevron Investor Day edited transcript", "url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "publisher": "Chevron", "date": "2025-11-12", "type": "investor presentation", "locator": "Slides 39–40; transcript pp. 21–22", "review_status": "full text reviewed", "supports": ["APOLO launched Aug 2024 (pre-window); Nov 2025 update describes >50,000 wells, finer prospectivity mapping and model training for spacing, completion and development timing.", "Chevron described an AI team and prioritization of high-value workflows including shale/tight and exploration.", "Chevron’s “billions” annual value is a potential decade-end ambition, not realized savings."], "accessed": "2026-09-27", "original_id": "mb-001"}], "field_sources": {"business_problem": ["S01"], "workflow": ["S01"], "technology_data": ["S01"], "integration_human_role": ["S01"], "deployment_scope": ["S01"], "economics": ["S01"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-001", "source_ids": ["S01"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-001"]}, {"key": "chevron-inspections", "existing_ids": ["AI-075"], "related_existing_ids": [], "company": "Chevron / DataRobot", "title": "Edge agentic assessment for robotic inspections", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Collaboration announced 2026-06-02", "practical_relevance": "Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.", "business_problem": "Aerial and terrestrial robot inspection missions require operator-by-operator permit checks, adding time and variability before work starts.", "workflow": "Safe Start agents assess operating conditions before and during robotic missions and generate plans; specialized sensor-analysis and geospatial agents coordinate. Chevron describes moving from manual point-in-time checks toward continuous assessment while inspection proceeds.", "technology_data": "DataRobot Agent Workforce Platform, NVIDIA NIM, NVIDIA AI compute, existing hard-wired gas sensors plus vision systems; edge/local assessment and integration into existing Chevron systems.", "integration_human_role": "System maintains oversight/control; human intervention is intended to decrease while approved operational standards, security, reliability and governance remain. Exact stop/go authority and fallback protocol not disclosed.", "deployment_scope": "Collaboration at Chevron facilities; no fleet/site count, rollout status beyond announced collaboration, or quantified result.", "partners": ["DataRobot", "NVIDIA"], "disclosed_results": [], "economics": "No quantified safety, time or financial outcomes. Avoid “autonomous inspections at scale” as established result.", "timeline": [{"date": "2026-06-02", "event": "Edge agentic assessment for robotic inspections documented or disclosed.", "source_ids": ["R015"]}], "limitations": ["Partner announcement; deployment scale and outcomes not independently reported.", "AI-assisted missions are not stated as human-free inspection."], "assessment": "Strong practical workflow detail and sensor/edge integration; limited to a collaboration announcement without outcome metrics.", "sources": [{"id": "R015", "title": "DataRobot and Chevron Collaborate to Advance Agentic AI for Autonomous Inspections", "url": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "publisher": "DataRobot", "date": "2026-06-02", "type": "partner case study", "locator": "Safe Start agent, robot missions and edge systems", "review_status": "full text reviewed", "supports": ["Chevron and DataRobot collaboration applies edge agents to support autonomous inspection at Chevron facilities.", "Safe Start assesses conditions before/during robotic missions using gas sensors, vision and locally running models; integrates with existing operating systems and NVIDIA NIM.", "Vendor and Chevron program manager describe reduced manual point-in-time checks and human oversight/control; no measured outcomes."], "accessed": "2026-09-27", "original_id": "mb-002"}], "field_sources": {"business_problem": ["R015"], "workflow": ["R015"], "technology_data": ["R015"], "integration_human_role": ["R015"], "deployment_scope": ["R015"], "economics": []}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-002", "source_ids": ["R015"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-075"]}, {"key": "shell-slb-collab", "existing_ids": ["AI-010"], "related_existing_ids": [], "company": "Shell / SLB", "title": "Upstream data/AI infrastructure and agentic-solutions collaboration", "department": "Production & field operations", "secondary_departments": [], "category": "Commercial AI product", "stage": "Agreement announced 2025-12-11; Q1 2026 SLB reporting repeats plans", "practical_relevance": "Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.", "business_problem": "Upstream teams often work across distinct data and workflows from subsurface to well construction and production.", "workflow": "Shell and SLB agreed to develop agentic AI solutions for upstream experts and an open infrastructure to unify data and workflows across the three domains using SLB Lumi.", "technology_data": "SLB Lumi data/AI platform; intended open and secure data environment. Specific model, data schemas and production systems not detailed.", "integration_human_role": "Framed to augment technical experts and decision-makers; no human approval gates described.", "deployment_scope": "Joint development ambition for Shell and wider industry; no named deployed workflow or user/asset scale disclosed.", "partners": ["SLB"], "disclosed_results": [], "economics": "No contract value, realized savings or results reported.", "timeline": [{"date": "2025-12-11", "event": "Upstream data/AI infrastructure and agentic-solutions collaboration documented or disclosed.", "source_ids": ["S46", "S05"]}], "limitations": ["2026-01 SLB earnings disclosure is a later update/repetition; agreement launch was 2025-12-11.", "Development plan only; no deployed agents or outcomes substantiated."], "assessment": "Material direction for connected upstream workflows, but not yet a demonstrated operational initiative in reviewed evidence.", "sources": [{"id": "S46", "title": "SLB Enters Collaboration Agreement to Accelerate New Digital Solutions", "url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "publisher": "SLB", "date": "2025-12-11", "type": "company announcement", "locator": "Shell collaboration agreement", "review_status": "full text reviewed", "supports": ["SLB and Shell agreed to develop upstream digital and AI solutions; roadmap aims for open data/AI infrastructure across subsurface, well construction and production on SLB Lumi.", "Announcement states ambition/aim, not a deployed customer workflow or realized results."], "accessed": "2026-09-27", "original_id": "mb-003"}, {"id": "S05", "title": "SLB Q4 and full-year 2025 results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "publisher": "SLB", "date": "2026-01-23", "type": "earnings release", "locator": "Digital segment customer collaborations section", "review_status": "full text reviewed", "supports": ["Q4 report repeats Shell collaboration aims and scope; this is an in-window disclosure/update, not evidence the collaboration was launched in January or deployed."], "accessed": "2026-09-27", "original_id": "mb-004"}], "field_sources": {"business_problem": ["S46"], "workflow": ["S46", "S05"], "technology_data": ["S46"], "integration_human_role": ["S46"], "deployment_scope": ["S46", "S05"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-003", "source_ids": ["S46", "S05"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-010"]}, {"key": "repsol-genai-center", "existing_ids": ["AI-032"], "related_existing_ids": ["AI-031"], "company": "Repsol", "title": "Generative AI Competence Center and employee scaling", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "In place >2 years by October 2025; report updated with 2025 counts", "practical_relevance": "Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.", "business_problem": "Repsol needs a repeatable way to identify and govern genAI cases, train staff and move selected solutions into production across a large multi-energy enterprise.", "workflow": "Competence Center develops use cases and employee skills; by 2025 company reported 60+ use cases in production, 22 scaled, and 5,000+ trained. A later consolidated report says >75 initiatives, while the October press release reports platform pilot with 22 agents/3 teams; those different counts should not be merged.", "technology_data": "Competence-center operating model and responsible AI; multi-agent platform itself is a distinct AI-031 profile. Report identifies automated daily well-data ingestion as a separate workflow detail related to the center.", "integration_human_role": "Human-in-the-loop design and supervision emphasized; 50+ people reportedly supervised agent teams in pilot; employees trained on GenAI/prompting.", "deployment_scope": "Enterprise-wide program; case counts and training numbers are reported by Repsol; pilot specific to Digitalization and Services team of 700+.", "partners": ["Accenture", "NVIDIA"], "disclosed_results": [{"metric": "GenAI use cases in production", "value": "60+; 22 scaled (Oct 2025 release)", "period": "2025-10-15", "attribution": "Company-reported", "source_ids": ["S19"]}, {"metric": "Staff trained", "value": "5,000+", "period": "2025-10-15 disclosure", "attribution": "Company-reported", "source_ids": ["S19"]}], "economics": "No dollar economics; efficiency described qualitatively.", "timeline": [{"date": "2025-10-15", "event": "Generative AI Competence Center and employee scaling documented or disclosed.", "source_ids": ["S19", "RS007"]}], "limitations": ["Case counts vary by source/date and counting definition; do not conflate competence-center cases with 22-agent pilot or later 34-agent scale.", "No independent outcomes or value quantification."], "assessment": "Useful enabling infrastructure and program scale. It documents production use and human capability-building, not a single workflow with verified returns.", "sources": [{"id": "S19", "title": "Repsol launches intelligent multi-agent platform", "url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "publisher": "Repsol", "date": "2025-10-15", "type": "company announcement", "locator": "Competence Center, platform pilot and user adoption", "review_status": "full text reviewed", "supports": ["Repsol says its GenAI Competence Center had produced 60+ use cases in production and 22 scaled; 5,000+ employees trained.", "Platform was a pilot in Digitalization and Services with 3 teams and 22 agents supporting processes for more than four months; 50+ people supervised agents.", "A separate 2025 consolidated management report later says 34 agents across three workflows; dates/scope differ, so counts should not be blended."], "accessed": "2026-09-27", "original_id": "mb-005"}, {"id": "RS007", "title": "2025 Consolidated Management Report", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "publisher": "Repsol", "date": "2026-02", "type": "annual report", "locator": "Section 5.5 Technology and digitalization", "review_status": "excerpt reviewed", "supports": ["2025 report says Competence Center promoted >75 initiatives after two years, and the multi-agent platform had 34 agents in three workflows.", "Report names automated well-data ingestion for daily drilling/geology reports, integration into corporate systems and faster exploration decision-making."], "accessed": "2026-09-27", "original_id": "mb-006"}], "field_sources": {"business_problem": ["S19"], "workflow": ["S19", "RS007"], "technology_data": ["S19", "RS007"], "integration_human_role": ["S19"], "deployment_scope": ["S19", "RS007"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-004", "source_ids": ["S19", "RS007"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-032"]}, {"key": "equinor-ai-portfolio", "existing_ids": ["AI-033"], "related_existing_ids": ["AI-034", "AI-035", "AI-036"], "company": "Equinor", "title": "AI-enabled maintenance, field planning and seismic portfolio", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Implemented cases disclosed 2026-01-07; results refer to calendar 2025 and cumulative since 2020", "practical_relevance": "Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.", "business_problem": "Equipment failures, planning search spaces and seismic interpretation volume exceed what teams can efficiently examine manually.", "workflow": "Predictive monitoring covers rotating equipment; AI generates and ranks well/development alternatives for expert review; seismic tooling accelerates interpretation. These are three distinct implementations reported together, not a single model.", "technology_data": "Company says traditional ML on operational data; predictive system spans >700 rotating machines/24,000 sensors; seismic tool processed 2m km² in 2025. Other algorithms/integration not disclosed.", "integration_human_role": "Experts focus on ranked planning options; maintenance tool flags failure/maintenance needs; geoscientists use expanded seismic interpretation. Human control and specific automation boundaries not disclosed.", "deployment_scope": "Equinor offshore platforms and land facilities; rotating asset coverage across all facilities, planning includes Johan Sverdrup phase 3, seismic use 2m km² in 2025.", "partners": [], "disclosed_results": [{"metric": "AI-attributed value", "value": "USD 130 million", "period": "Calendar 2025", "attribution": "Company-reported", "source_ids": ["S20"]}, {"metric": "Predictive maintenance value", "value": "USD 120 million", "period": "Cumulative since 2020", "attribution": "Company-reported", "source_ids": ["S20"]}, {"metric": "Johan Sverdrup phase 3 planning saving", "value": "USD 12 million", "period": "Case disclosed 2026-01-07", "attribution": "Company-reported partner saving", "source_ids": ["S20"]}, {"metric": "Seismic interpretation area", "value": "2 million km²", "period": "2025", "attribution": "Company-reported", "source_ids": ["S20"]}], "economics": "Equinor reported $130m value in 2025 and >$330m cumulative since 2020. $120m maintenance is cumulative since 2020; $12m for Johan Sverdrup is partner saving. Prevent double-counting these within 2025 total.", "timeline": [{"date": "2026-01-07", "event": "AI-enabled maintenance, field planning and seismic portfolio documented or disclosed.", "source_ids": ["S20"]}], "limitations": ["Parent total 2025 value and subcase numbers use different periods; don’t sum or treat as additive.", "No independent value audit or model architecture."], "assessment": "Substantive named use across operating reliability and subsurface work with scale and company-reported value; practical proof is stronger than generic AI strategy, but economics remain issuer claims.", "sources": [{"id": "S20", "title": "Equinor: AI delivered USD 130m value in 2025", "url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "publisher": "Equinor", "date": "2026-01-07", "type": "company announcement", "locator": "Implemented applications and value disclosure", "review_status": "full text reviewed", "supports": ["Company says AI generated USD 130m value for Equinor and partners in calendar 2025, with >USD 330m cumulative since 2020.", "Three applications: 700+ rotating machines/24,000 sensors predictive maintenance (USD 120m since 2020); well and field planning, including USD12m Johan Sverdrup phase-3 saving; seismic interpretation of 2m km2 in 2025 with tenfold capacity."], "accessed": "2026-09-27", "original_id": "mb-007"}], "field_sources": {"business_problem": ["S20"], "workflow": ["S20"], "technology_data": ["S20"], "integration_human_role": ["S20"], "deployment_scope": ["S20"], "economics": ["S20"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-005", "source_ids": ["S20"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-033"]}, {"key": "woodside-lumina", "existing_ids": ["AI-041"], "related_existing_ids": [], "company": "Woodside", "title": "Lumina enterprise agent platform and adoption governance", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Company page last updated 2026-05-27; exact Lumina release date not stated", "practical_relevance": "Practical enterprise platform with governance, but agent count alone does not establish operating value.", "business_problem": "Staff need integrated everyday AI access while enterprise use requires cybersecurity, risk controls and consistent deployment.", "workflow": "Lumina supports >50 agents for business-function tasks; Copilot handles Office workflows; ChatGPT Enterprise is sandbox for analytics/agent experimentation before agents move to Lumina. AI Centre of Excellence provides platform/tools; AI Council oversees risk, governance and priority.", "technology_data": "Enterprise agent platform; Copilot/M365 and ChatGPT Enterprise. Model/data connections for individual Lumina agents not disclosed.", "integration_human_role": "Responsible AI assessment for every use case and senior cross-functional council; platform supports staff tasks rather than autonomous operating decisions.", "deployment_scope": "Woodside enterprise-wide adoption claim; >50 agents, but no department, user count or asset-specific production scale.", "partners": [], "disclosed_results": [{"metric": "Lumina agent count", "value": "Over 50", "period": "Page updated 2026-05-27", "attribution": "Company-reported", "source_ids": ["S24"]}], "economics": "No quantitative value or cost reduction.", "timeline": [{"date": "updated 2026-05-27; page undated", "event": "Lumina enterprise agent platform and adoption governance documented or disclosed.", "source_ids": ["S24"]}], "limitations": ["Page is current and undated except last-updated stamp; use that separately from launch chronology.", "Maint Intel and Startup Advisor are separate existing exact initiatives AI-042 and AI-043 and not duplicated."], "assessment": "Practical enterprise platform with governance, but agent count alone does not establish operating value.", "sources": [{"id": "S24", "title": "Woodside Artificial Intelligence", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "publisher": "Woodside", "date": "updated 2026-05-27; page undated", "type": "company webpage", "locator": "Lumina, Copilot, ChatGPT Enterprise and governance", "review_status": "full text reviewed", "supports": ["Woodside says Lumina supports >50 agents for business tasks; Microsoft Copilot supports Office productivity; ChatGPT Enterprise is an AI sandbox before custom agents integrate with Lumina.", "Responsible AI assessment, Centre of Excellence and executive AI Council described.", "Same page includes Maint Intel and Startup Advisor, already treated as separate exact profiles AI-042/AI-043; page does not state initial deployment dates."], "accessed": "2026-09-27", "original_id": "mb-008"}], "field_sources": {"business_problem": ["S24"], "workflow": ["S24"], "technology_data": ["S24"], "integration_human_role": ["S24"], "deployment_scope": ["S24"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-006", "source_ids": ["S24"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-041"]}, {"key": "aker-cognite-rca", "existing_ids": ["AI-049"], "related_existing_ids": [], "company": "Aker BP / Cognite", "title": "Atlas AI agents for root-cause analysis and E&P workflows", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Expanded partnership disclosed 2025-09-16", "practical_relevance": "Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.", "business_problem": "Engineers investigate operational issues by manually gathering context; other processes rely on document-heavy supply-chain and equipment data workflows.", "workflow": "RCA agents identify relevant information, construct cause maps and suggest next analysis steps. Related work automates document processing for supply chain and barrier management, registers equipment, prepares predictive-maintenance data foundations and optimizes production chemistry.", "technology_data": "Atlas AI agents and low-code workbench connected to Cognite Data Fusion with live OT, IT and engineering data; exact agent models/permission setup not disclosed.", "integration_human_role": "Engineers evaluate RCA agent suggestions; partner says human insight and autonomous agents work together.", "deployment_scope": "Aker BP partnership across exploration/production; no agent/site count reported in release.", "partners": ["Cognite"], "disclosed_results": [{"metric": "Engineer time for root-cause analysis", "value": "Reduced by over 70% (partner-reported)", "period": "As reported 2025-09-16", "attribution": "Vendor-reported customer statement", "source_ids": ["S28"]}], "economics": "Aker BP/Cognite report >70% reduction in RCA analysis time; no sample size, baseline, duration or dollar conversion. Other benefits qualitative.", "timeline": [{"date": "2025-09-16", "event": "Atlas AI agents for root-cause analysis and E&P workflows documented or disclosed.", "source_ids": ["S28"]}], "limitations": ["Cognite is source publisher; productivity metric lacks sample/baseline.", "Other workflows are described as ongoing initiatives, not necessarily all live."], "assessment": "Clear analytic workflow and domain data foundation; 70% figure is meaningful but partner-reported and unbenchmarked.", "sources": [{"id": "S28", "title": "Aker BP leverages Cognite Atlas AI to pioneer AI-first future", "url": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "publisher": "Cognite", "date": "2025-09-16", "type": "partner case study", "locator": "RCA agents and additional workflows", "review_status": "full text reviewed", "supports": ["Aker BP expanded Atlas AI adoption. Described workflows include document processing for supply chain, barrier management, automatic equipment registration, data foundation for predictive maintenance and production chemistry.", "Root-cause agents identify relevant info, build cause maps and suggest steps; Cognite/Aker BP report >70% reduction in engineer RCA time."], "accessed": "2026-09-27", "original_id": "mb-009"}], "field_sources": {"business_problem": ["S28"], "workflow": ["S28"], "technology_data": ["S28"], "integration_human_role": ["S28"], "deployment_scope": ["S28"], "economics": ["S28"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-007", "source_ids": ["S28"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-049"]}, {"key": "exxon-ops-monitoring", "existing_ids": [], "related_existing_ids": ["AI-022"], "company": "ExxonMobil", "title": "AI monitoring of production bottlenecks at operating sites", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Company source dated 2025-10-23", "practical_relevance": "Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.", "business_problem": "A bottleneck or mechanical issue in one production area can constrain connected processes and reduce throughput.", "workflow": "ExxonMobil India says its AI teams work with production personnel to monitor equipment in real time, identify current or prospective bottlenecks, then coordinate with operators and production engineers to remove them.", "technology_data": "AI applied to operational/equipment data; company content also describes connecting sensor and image data, but this source does not specify specific model, field, system or sensor architecture.", "integration_human_role": "Engineers and asset operators investigate AI flags against physical operating context; source explicitly pairs AI and people.", "deployment_scope": "ExxonMobil India describes company operations at a general level; no number of sites or assets stated.", "partners": [], "disclosed_results": [], "economics": "No measured production uplift or cost result.", "timeline": [{"date": "2025-10-23", "event": "AI monitoring of production bottlenecks at operating sites documented or disclosed.", "source_ids": ["RS112"]}], "limitations": ["Company page describes practice at high level and does not verify legacy Q3 2025 transcript’s “Permian” scope.", "No data on bottlenecks addressed or realized output."], "assessment": "Provides official corroboration for an operational AI monitoring workflow behind the broad legacy operating-site claim, but not for specific Permian deployment or outcomes.", "sources": [{"id": "RS112", "title": "ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence", "url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "publisher": "ExxonMobil", "date": "2025-10-23", "type": "company webpage", "locator": "Production bottleneck monitoring", "review_status": "full text reviewed", "supports": ["ExxonMobil India describes using AI with production staff to monitor equipment in real time and identify current or prospective bottlenecks affecting production.", "Engineers and asset operators use domain knowledge to investigate and remove bottlenecks; no basin/site counts or metrics."], "accessed": "2026-09-27", "original_id": "mb-010"}], "field_sources": {"business_problem": ["RS112"], "workflow": ["RS112"], "technology_data": ["RS112"], "integration_human_role": ["RS112"], "deployment_scope": ["RS112"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-008", "source_ids": ["RS112"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-178"]}, {"key": "exxon-tt-seismic", "existing_ids": ["AI-069"], "related_existing_ids": [], "company": "ExxonMobil Trinidad and Tobago / SLB", "title": "AI-enabled horizon interpretation in frontier seismic imaging", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Two-year award reported 2026-07-24", "practical_relevance": "Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.", "business_problem": "Sparse legacy 2D seismic data makes ultra-deepwater frontier imaging and prospect evaluation difficult.", "workflow": "SLB awarded two-year data processing scope to generate first 3D image over nearly 6,000 km²; broader processing includes FWI, velocity modeling, source separation and multicomponent streamer data, with AI-enabled horizon interpretation.", "technology_data": "SLB Omega geophysical processing software, geophysical imaging suite and AI-enabled horizon interpretation; source does not explain model architecture/training data.", "integration_human_role": "Output is intended to help ExxonMobil evaluate area; interpretation remains an exploration decision-support workflow, no autonomous decision claim.", "deployment_scope": "Named ExxonMobil Trinidad and Tobago Deepwater block; 6,000 km², two-year contract.", "partners": ["SLB"], "disclosed_results": [], "economics": "Contract value and commercial discoveries/returns not disclosed.", "timeline": [{"date": "2026-07-24", "event": "AI-enabled horizon interpretation in frontier seismic imaging documented or disclosed.", "source_ids": ["S27"]}], "limitations": ["AI is a component of wider seismic processing contract; source does not isolate its contribution.", "No prospect or discovery result yet."], "assessment": "Concrete frontier imaging project with a two-year and geographic scope; do not conflate award/analysis with exploration success.", "sources": [{"id": "S27", "title": "SLB Q2 2026 results: ExxonMobil Trinidad & Tobago seismic project", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "partner/customer announcement", "locator": "Digital/technology projects section", "review_status": "full text reviewed", "supports": ["ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "accessed": "2026-09-27", "original_id": "mb-011"}], "field_sources": {"business_problem": ["S27"], "workflow": ["S27"], "technology_data": ["S27"], "integration_human_role": ["S27"], "deployment_scope": ["S27"], "economics": ["S27"]}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-009", "source_ids": ["S27"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-069"]}, {"key": "petronas-aingineer", "existing_ids": [], "related_existing_ids": ["AI-079"], "company": "PETRONAS / AIngineer / Microsoft", "title": "Industrial AI copilot integrated with engineering technical references", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Collaboration announced 2025-06-10; source says technical use cases and prior deployments, without naming sites", "practical_relevance": "Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.", "business_problem": "Engineers in complex process industries need rapid access to technical standards, databases and accumulated know-how when diagnosing reliability and maintenance issues.", "workflow": "AIngineer copilot is embedded in everyday engineering workflow and connects technical databases/standards to provide answers and insight; collaboration with Microsoft supports cloud scaling/commercialization. PETRONAS says it has collaborated on multiple technical use cases.", "technology_data": "LLM plus machine learning algorithms; integration with technical databases/standards; Microsoft cloud infrastructure; named model versions/security architecture not disclosed.", "integration_human_role": "Designed to augment engineers’ decision-making and preserve institutional knowledge; release does not specify approval or safety controls for recommendations.", "deployment_scope": "Energy/heavy industry product; PETRONAS use cases mentioned but no asset, user count or independent customer deployment details.", "partners": ["Microsoft"], "disclosed_results": [], "economics": "No realized cost/diagnostic-time outcome quantified; lower maintenance costs and faster diagnosis are stated aims.", "timeline": [{"date": "2025-06-10", "event": "Industrial AI copilot integrated with engineering technical references documented or disclosed.", "source_ids": ["RS113"]}], "limitations": ["MoU and commercialization collaboration does not establish the extent of PETRONAS operational rollout.", "No named technical use case or measured outcome."], "assessment": "Concrete industrial copilot/product that merits a separate initiative from the exploratory Microsoft MoU (AI-079); implementation proof remains limited.", "sources": [{"id": "RS113", "title": "PETRONAS-incubated AIngineer joins forces with Microsoft", "url": "https://www.petronas.com/bm/node/1668", "publisher": "PETRONAS", "date": "2025-06-10", "type": "company announcement", "locator": "AIngineer industrial engineer copilot", "review_status": "full text reviewed", "supports": ["AIngineer is a PETRONAS-incubated industrial AI copilot using LLMs and ML and integrating technical databases and standards.", "PETRONAS says it has technical use cases and prior deployments in energy/government; this release does not name a customer asset or quantify results.", "Microsoft collaboration includes cloud scaling and prospective marketplace commercialization."], "accessed": "2026-09-27", "original_id": "mb-013"}], "field_sources": {"business_problem": ["RS113"], "workflow": ["RS113"], "technology_data": ["RS113"], "integration_human_role": ["RS113"], "deployment_scope": ["RS113"], "economics": ["RS113"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-010", "source_ids": ["RS113"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-179"]}, {"key": "akerbp-copilot", "existing_ids": ["AI-131"], "related_existing_ids": [], "company": "Aker BP / Microsoft", "title": "Copilot rollout, employee agents and Knowledge Agents", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Microsoft customer story dated 2025-06-23; rollout began with 300 licenses in 2023", "practical_relevance": "Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.", "business_problem": "Employees need more efficient knowledge retrieval and routine writing/meeting work while safe operations require governance and support for broad AI adoption.", "workflow": "Aker BP expanded Copilot and Copilot Studio access, training, help desk and champion network. Knowledge Agents answer repeated document questions and route complex requests to experts; employees and central specialist team build agents.", "technology_data": "Microsoft 365 Copilot and Copilot Studio; knowledge agents use generative AI, low-code capabilities, topics and human-in-the-loop features; detailed corpus/integration not stated.", "integration_human_role": "Agents route complex queries to experts; central team supports makers; security ratings and governance precede wider adoption.", "deployment_scope": "Company-wide rollout planned for 2025; 300 starting licenses. User activity and agent-live counts are from Microsoft customer case.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Initial Copilot licenses", "value": "300", "period": "Early deployment; 2023 pilot context", "attribution": "Customer statement in vendor case", "source_ids": ["R126"]}, {"metric": "Regular active users", "value": "96% over preceding four weeks", "period": "As described 2025-06-23", "attribution": "Vendor-reported customer metric", "source_ids": ["R126"]}], "economics": "No quantified labor-savings/economic outcome.", "timeline": [{"date": "2025-06-23", "event": "Copilot rollout, employee agents and Knowledge Agents documented or disclosed.", "source_ids": ["R126"]}], "limitations": ["Microsoft-published customer narrative; 96% metric lacks denominator and independent audit.", "License expansion plan is not proof all licenses were delivered."], "assessment": "Practical corporate workflow and adoption program with named knowledge agents and explicit human-routing/governance details.", "sources": [{"id": "R126", "title": "Aker BP builds AI-first future with Microsoft 365 Copilot and AI agents", "url": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "publisher": "Microsoft", "date": "2025-06-23", "type": "partner case study", "locator": "Employee Copilot rollout, governance, Knowledge Agents", "review_status": "full text reviewed", "supports": ["Aker BP started with 300 Copilot licenses and planned all-staff licensing in 2025; trained employees, set up help desk and champion/accountability network.", "Employee/small-team-built agents, including Knowledge Agents, answer repetitive documentation questions and route harder queries; several agents live.", "Microsoft case quotes 96% active users in prior four weeks; vendor-published customer metric, not independent audit."], "accessed": "2026-09-27", "original_id": "mb-014"}], "field_sources": {"business_problem": ["R126"], "workflow": ["R126"], "technology_data": ["R126"], "integration_human_role": ["R126"], "deployment_scope": ["R126"], "economics": ["R126"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-011", "source_ids": ["R126"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-131"]}, {"key": "santos-xecta", "existing_ids": ["AI-083"], "related_existing_ids": [], "company": "Santos / Xecta", "title": "Integrated production-system model for Eastern Australia and PNG", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Xecta source index lists deployment in June/July 2025; exact release date not available", "practical_relevance": "Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.", "business_problem": "Continuous production optimization needs a joined view from reservoir through facility inlet across geographically distributed assets.", "workflow": "Xecta says Santos implemented an Integrated Production System Model to generate ongoing optimization insights across Eastern Australia and Papua New Guinea; legacy profile also claims expansion from Cooper Basin and five-year agreement, not independently verified in reviewed Xecta index.", "technology_data": "Specific model/data architecture could not be reviewed because linked release was unavailable; Xecta index names the integrated production model.", "integration_human_role": "Decision rights/control automation are not disclosed by accessible primary page.", "deployment_scope": "Named deployment in Eastern Australia and PNG, but no operating well/facility counts found.", "partners": ["Xecta"], "disclosed_results": [], "economics": "No quantitative production/economic results in accessible primary listing.", "timeline": [{"date": "June/July 2025 listing; exact publication date unavailable", "event": "Integrated production-system model for Eastern Australia and PNG documented or disclosed.", "source_ids": ["RS114"]}], "limitations": ["Official vendor news index reviewed; exact article body and date inaccessible.", "Claims about initial Cooper Basin rollout, 5-year contract and Kelvin.ai autonomous control not reverified."], "assessment": "Retain as a named customer/model disclosure with limited verification; obtain the underlying dated release before making stronger claims.", "sources": [{"id": "RS114", "title": "Xecta news and resources index", "url": "https://xecta.com/news-resources/", "publisher": "Xecta", "date": "June/July 2025 listing; exact publication date unavailable", "type": "company webpage", "locator": "Santos Eastern Australia and PNG digital oilfield listing", "review_status": "excerpt reviewed", "supports": ["Xecta index identifies Santos integrated production-system model/digital oilfield across Eastern Australia and PNG.", "Underlying dated technical release did not open; current confirmation limited to vendor page index."], "accessed": "2026-09-27", "original_id": "mb-015"}], "field_sources": {"business_problem": ["RS114"], "workflow": ["RS114"], "technology_data": ["RS114"], "integration_human_role": [], "deployment_scope": ["RS114"], "economics": []}, "verification": "verified partner", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-012", "source_ids": ["RS114"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-083"]}, {"key": "total-digital-factory", "existing_ids": ["AI-030"], "related_existing_ids": ["AI-029"], "company": "TotalEnergies", "title": "Digital Factory portfolio of AI and digital solutions", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "2025-06-12 disclosure, with five-year factory anniversary", "practical_relevance": "Useful company-level production of use cases; reported solution count is not an operational results metric.", "business_problem": "Business units need centrally developed solutions for earth sciences, machinery, industrial asset performance, renewable production and customer energy use.", "workflow": "Digital Factory’s 300 developers/data/digital experts built over 100 solutions, 60 employing technologies from machine learning to generative AI. Mistral collaboration spawned separate planned joint lab use cases; those should remain AI-029, not be collapsed into this portfolio profile.", "technology_data": "Solution-level architectures are heterogeneous; source mentions ML-to-GenAI range but no platform, models or data integration details.", "integration_human_role": "Central digital teams develop solutions for business users; customer/industrial operating decisions remain unspecified.", "deployment_scope": "Company-wide factory; counts from June 2025 company announcement.", "partners": [], "disclosed_results": [{"metric": "Digital Factory team", "value": "300 developers, data scientists and digital experts", "period": "2025-06-12", "attribution": "Company-reported", "source_ids": ["RS115"]}, {"metric": "Factory solutions", "value": "100+ total; 60 use ML-to-GenAI methods", "period": "2025-06-12", "attribution": "Company-reported", "source_ids": ["RS115"]}], "economics": "No AI-only economic value quantified.", "timeline": [{"date": "2025-06-12", "event": "Digital Factory portfolio of AI and digital solutions documented or disclosed.", "source_ids": ["RS115"]}], "limitations": ["Portfolio counts do not establish active use, adoption or business value per solution.", "Specific Mistral joint lab is separate AI-029 and launch describes tests/plans, not production."], "assessment": "Useful company-level production of use cases; reported solution count is not an operational results metric.", "sources": [{"id": "RS115", "title": "TotalEnergies to collaborate with Mistral AI", "url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "publisher": "TotalEnergies", "date": "2025-06-12", "type": "company announcement", "locator": "Digital Factory baseline and joint AI lab", "review_status": "full text reviewed", "supports": ["TotalEnergies says its Digital Factory had 300 digital experts and developed >100 solutions, 60 using technologies from ML to GenAI.", "The announcement creates a distinct Mistral joint innovation lab for researcher assistant, industrial asset decision support, customer and renewable use cases; it says those use cases will be designed/tested, not deployed."], "accessed": "2026-09-27", "original_id": "mb-016"}], "field_sources": {"business_problem": ["RS115"], "workflow": ["RS115"], "technology_data": ["RS115"], "integration_human_role": ["RS115"], "deployment_scope": ["RS115"], "economics": ["RS115"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "majors_backfill", "packet_id": "majors_backfill-013", "source_ids": ["RS115"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-030"]}, {"secondary_departments": ["Refining, LNG & process operations"], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "indianoil-vishleshan-2", "existing_ids": [], "company": "Indian Oil Corporation (IndianOil)", "title": "AI/ML predictive maintenance across refineries and operating locations", "department": "Maintenance & reliability", "stage": "IndianOil’s FY2024-25 annual report says it is deploying AI and machine learning for predictive maintenance; site-specific rollout status is not disclosed.", "practical_relevance": "A concrete reliability workflow is identified, but the public report does not name facilities, equipment classes or achieved downtime reduction.", "business_problem": "Improve asset reliability and reduce downtime at refineries and operating locations.", "workflow": "IndianOil says it is deploying AI and machine learning for predictive maintenance. The report does not describe specific alarms, predictions or maintenance decisions.", "technology_data": "AI/ML is stated; data sources, model types and interfaces are not disclosed.", "integration_human_role": "Human review and work-order integration are not disclosed.", "deployment_scope": "Refineries and operating locations; site count and rollout stage are not disclosed.", "economics": "No measured uptime, downtime or financial impact disclosed.", "timeline": [{"date": "2025 (FY2024-25 annual report)", "event": "IndianOil reported ongoing deployment of AI/ML predictive maintenance.", "source_ids": ["RS116"]}], "limitations": ["The report gives a broad ongoing-deployment statement with no named site, asset class, model, status by asset or outcome data."], "assessment": "Analyst inference: predictive maintenance is operationally relevant, but disclosure is too broad to evaluate technical maturity or results.", "sources": [{"id": "RS116", "title": "IndianOil – Intellectual Capital, Integrated Annual Report 2024-25", "url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "publisher": "Indian Oil Corporation", "date": "2025 (FY2024-25)", "type": "annual report webpage", "locator": "Leveraging Data and AI for Operational Efficiency", "review_status": "excerpt reviewed", "supports": ["IndianOil says it is deploying AI and machine learning for predictive maintenance in refineries and operating locations to improve reliability and reduce downtime."], "original_id": "regional-001"}], "field_sources": {"business_problem": ["RS116"], "workflow": ["RS116"], "technology_data": ["RS116"], "integration_human_role": [], "deployment_scope": ["RS116"], "economics": []}, "cohort": "regional", "packet_id": "regional-001", "source_ids": ["RS116"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-180"], "related_existing_ids": []}, {"secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "bpcl-refinery-reliability-and-safety-ai", "existing_ids": [], "company": "Bharat Petroleum Corporation (BPCL)", "title": "Refinery predictive alerts and AI video safety analytics", "department": "Maintenance & reliability", "stage": "FY2024-25: video analytics implemented at all three refineries; predictive analytics initiated across all three, extending a successful Kochi trial", "practical_relevance": "The projects put AI into two concrete refinery workflows: early warning for critical-equipment degradation and video-based safety/compliance monitoring. BPCL distinguishes implemented video coverage from the predictive system’s implementation stage.", "business_problem": "Reduce the chance that critical equipment failure or performance degradation causes lost production, and detect field safety hazards/non-compliance.", "workflow": "A cloud-based predictive-analytics project is being implemented across Mumbai, Kochi and Bina refineries to issue early alerts for identified critical equipment; this extended a successful Kochi trial. Separately, Ultra Critical Video Analytics applies AI algorithms to field video for non-compliance and safety-hazard detection and integrates with BPCL’s Work Permit System.", "technology_data": "Cloud analytics and predictive alerts are disclosed for the equipment workflow; video-feed algorithms and Work Permit System integration for UCVA. Models, sensors, alarm thresholds and cybersecurity architecture are not disclosed.", "integration_human_role": "The source describes alerts and work-permit integration but not automated control actions, escalation rules or human review responsibilities.", "deployment_scope": "Three refineries (Mumbai, Kochi, Bina). UCVA was implemented across all three; predictive reliability analytics was initiated across all three after a Kochi trial.", "economics": "No AI-attributed cost, uptime, incident or return metrics disclosed. BPCL also describes conventional Operations Driven Reliability, excluded from the AI workflow here.", "timeline": [{"date": "FY2024-25", "event": "Predictive analytics initiated across three refineries, extending Kochi trial; UCVA implemented across all three.", "source_ids": ["RS117"]}], "limitations": ["No measured reliability, safety or financial outcome is given. Predictive analytics was described as an initiated implementation, so full operational rollout is not established."], "assessment": "Analyst inference: combining asset alerts with existing work-permit controls is a practical maintenance-and-safety application, but the public report supports deployment status more strongly than achieved impact.", "sources": [{"id": "RS117", "title": "Annual Report 2024-25", "url": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "publisher": "Bharat Petroleum", "date": "2025 (FY2024-25)", "type": "annual report", "locator": "Digital initiatives, p. 135", "review_status": "excerpt reviewed", "supports": ["AI predictive analytics for asset reliability/performance was initiated across Mumbai, Kochi and Bina refineries, extending a successful Kochi trial; cloud predictive alerts target failures/degradation in identified critical equipment.", "AI video algorithms detect non-compliance and safety hazards from field video; Ultra Critical Video Analytics was implemented across all three refineries and integrated with the in-house Work Permit System."], "original_id": "regional-002"}], "field_sources": {"business_problem": ["RS117"], "workflow": ["RS117"], "technology_data": ["RS117"], "integration_human_role": ["RS117"], "deployment_scope": ["RS117"], "economics": []}, "cohort": "regional", "packet_id": "regional-002", "source_ids": ["RS117"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-181"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "hpcl-visakh-lcmax-digital-suite", "existing_ids": [], "company": "Hindustan Petroleum Corporation (HPCL)", "title": "AI-enabled LC-Max digital suite at Visakh residue-upgradation facility", "department": "Refining, LNG & process operations", "stage": "Deployed with the facility commissioning announced January 2026", "practical_relevance": "This is a named AI-enabled suite tied to a major residue-hydrocracking unit. It combines monitoring, predictive analytics and optimization; public disclosure does not separate the suite’s contribution from the physical process technology’s performance.", "business_problem": "Stabilize and optimize operation of the high-complexity residue-upgradation process.", "workflow": "At commissioning of the Visakh RUF, HPCL said the digital suite integrates real-time monitoring, predictive analytics based on proprietary thermodynamic models and AI-driven optimization.", "technology_data": "Proprietary thermodynamic models and AI are named. Sensor coverage, model architecture, control integration and operating mode are not described.", "integration_human_role": "Human approval, safeguards, alert routing and whether optimization is advisory or closed-loop are not disclosed.", "deployment_scope": "The RUF LC-Max Digital Suite at HPCL’s Visakh Refinery.", "economics": "The announcement reports physical-unit capacity and conversion/yield characteristics, but does not attribute these outcomes to AI. AI-specific economics are not disclosed.", "timeline": [{"date": "2026-01-05", "event": "HPCL announced RUF commissioning and deployment of its LC-Max Digital Suite.", "source_ids": ["RS118"]}], "limitations": ["AI contribution to process performance is not isolated; model/data/control details and quantified AI-specific outcomes are absent."], "assessment": "Analyst inference: the integration of predictive analytics with process optimization in a newly commissioned complex unit is operationally relevant, though public evidence does not establish autonomy or incremental AI value.", "sources": [{"id": "RS118", "title": "HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery", "url": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "publisher": "Hindustan Petroleum", "date": "2026-01-05", "type": "company announcement", "locator": "paragraphs 24-28", "review_status": "full text reviewed", "supports": ["At commissioning of the Visakh residue-upgradation facility, HPCL said its LC-Max Digital Suite integrates real-time monitoring, predictive analytics using proprietary thermodynamic models and AI-driven optimization.", "The source reports RUF physical-process conversion/yield figures separately; it does not attribute those outcomes to AI."], "original_id": "regional-003"}], "field_sources": {"business_problem": ["RS118"], "workflow": ["RS118"], "technology_data": ["RS118"], "integration_human_role": [], "deployment_scope": ["RS118"], "economics": ["RS118"]}, "cohort": "regional", "packet_id": "regional-003", "source_ids": ["RS118"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-182"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "ril-crude-feedstock-optimizer", "existing_ids": [], "company": "Reliance Industries Limited (RIL)", "title": "AI-driven crude-feedstock blend selection for O2C", "department": "Refining, LNG & process operations", "stage": "Described by the chairman as an existing proprietary tool in the June 2026 shareholder statement", "practical_relevance": "Crude selection and blending are directly linked to refinery economics and operating constraints. Reliance names a deployed decision-support tool but does not disclose its operating scope or measured value.", "business_problem": "Select efficient crude blends for the Oil-to-Chemicals system.", "workflow": "Reliance says its proprietary AI-driven feedstock-optimization tool helps select efficient crude blends; a broader statement says AI process optimization improves O2C yield and reduces energy use.", "technology_data": "AI and proprietary tool are identified, but model type, crude assays, unit constraints, planning integration and data refresh are not disclosed.", "integration_human_role": "The source says the tool helps select blends; final decision authority or operator approval is not described.", "deployment_scope": "Reliance O2C operations are referenced, with no site or asset count disclosed.", "economics": "Company states the broader O2C AI optimization is improving yield and reducing energy, but gives no quantity, period-specific baseline or AI-specific financial impact.", "timeline": [{"date": "2026-06", "event": "Chairman’s statement describes AI feedstock optimization and O2C process optimization.", "source_ids": ["RS119"]}], "limitations": ["The public statement lacks technical detail, deployment scale, quantified results and attribution methodology."], "assessment": "Analyst inference: blend selection is a high-value, repeatable refinery planning use case; public disclosure establishes existence of the tool but not its effect size or integration depth.", "sources": [{"id": "RS119", "title": "Chairman’s Statement, 49th Annual General Meeting", "url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "publisher": "Reliance Industries", "date": "2026-06", "type": "shareholder statement", "locator": "O2C business performance", "review_status": "excerpt reviewed", "supports": ["Reliance says its proprietary AI-driven feedstock-optimization tool helps select efficient crude blends.", "The same statement says AI-driven O2C process optimization improves yield and reduces energy; no quantified result, control design or site scope is given."], "original_id": "regional-004"}], "field_sources": {"business_problem": ["RS119"], "workflow": ["RS119"], "technology_data": ["RS119"], "integration_human_role": ["RS119"], "deployment_scope": ["RS119"], "economics": ["RS119"]}, "cohort": "regional", "packet_id": "regional-004", "source_ids": ["RS119"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-183"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "pttgc-smart-plant-ai", "existing_ids": [], "company": "PTT Global Chemical (PTTGC / GC)", "title": "Smart Plant AI: plant-wide optimization and predictive maintenance", "department": "Maintenance & reliability", "stage": "GC reported the Smart Plant workflows in July 2026; site-level rollout is unspecified.", "practical_relevance": "Analyst inference: these workflows target plant-wide operating efficiency and equipment reliability.", "business_problem": "Improve production efficiency and equipment reliability.", "workflow": "Plant-wide optimization uses AI and digital plant models in real-time operating optimization. AI/ML analyze equipment-condition data for anomaly/failure prediction; AI vision, drones and robotics support inspection.", "technology_data": "AI, digital plant models, equipment analytics, computer vision and robotics are named.", "integration_human_role": "Approval, maintenance-system integration and closed-loop control are not disclosed.", "deployment_scope": "GC production complex; site count not stated.", "economics": "No AI-workflow-specific savings or results disclosed.", "timeline": [{"date": "2026-07-20", "event": "GC detailed Smart Plant AI initiatives in a company announcement.", "source_ids": ["RS120"]}], "limitations": ["Rollout counts and measured outcomes are not disclosed."], "assessment": "Analyst inference: coupling complex-wide optimization with condition-based failure prediction is a relatively mature operational portfolio; the evidence is company-reported and lacks measurable technical/financial proof.", "sources": [{"id": "RS120", "title": "GC Advances Holistic Optimization Through Digital Technology", "url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "publisher": "PTT Global Chemical", "date": "2026-07-20", "type": "company announcement", "locator": "Smart Plant / Supply Chain & Operations Optimization", "review_status": "full text reviewed", "supports": ["GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale."], "original_id": "regional-005"}], "field_sources": {"business_problem": ["RS120"], "workflow": ["RS120"], "technology_data": ["RS120"], "integration_human_role": [], "deployment_scope": ["RS120"], "economics": ["RS120"]}, "cohort": "regional", "packet_id": "regional-005", "source_ids": ["RS120"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-184"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [], "partners": [], "verification": "verified primary", "key": "pttgc-hero-heat-integration", "existing_ids": [], "company": "PTT Global Chemical (PTTGC / GC)", "title": "HERO AI-enabled heat-exchanger-network energy optimization", "department": "Refining, LNG & process operations", "stage": "Industrial-scale deployment reported by GC in July 2026.", "practical_relevance": "Analyst inference: a defined AI/simulation application to process energy efficiency.", "business_problem": "Optimize heat integration to reduce process energy use.", "workflow": "HERO uses AI and process simulation to optimize heat integration across production processes.", "technology_data": "AI and process simulation; data and control details are undisclosed.", "integration_human_role": "Approval and automatic-control status are undisclosed.", "deployment_scope": "GC says deployed at industrial scale; sites are not named.", "economics": "No HERO-specific energy or financial result disclosed.", "timeline": [{"date": "2026-07-20", "event": "GC reported industrial-scale deployment of HERO.", "source_ids": ["RS120"]}], "limitations": ["Deployment and “first in Southeast Asia” claims are company-reported; no baseline or quantified result."], "assessment": "Analyst inference: AI plus process simulation applied to heat integration is a clearly scoped industrial workflow, but effect size and replication remain unverified.", "sources": [{"id": "RS120", "title": "GC Advances Holistic Optimization Through Digital Technology", "url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "publisher": "PTT Global Chemical", "date": "2026-07-20", "type": "company announcement", "locator": "Smart Plant / Supply Chain & Operations Optimization", "review_status": "full text reviewed", "supports": ["GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale."], "original_id": "regional-005"}], "field_sources": {"business_problem": ["RS120"], "workflow": ["RS120"], "technology_data": ["RS120"], "integration_human_role": [], "deployment_scope": ["RS120"], "economics": []}, "cohort": "regional", "packet_id": "regional-006", "source_ids": ["RS120"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-185"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [{"metric": "well indicators interpreted", "value": "more than 10,000 in about 2 minutes, versus more than 3 days manually", "period": "July 2025 disclosure", "attribution": "Company-reported", "source_ids": ["RS121"]}, {"metric": "additional production per well", "value": "average 1 tonne of oil/day per well", "period": "July 2025 disclosure; evaluation basis not given", "attribution": "Company-reported", "source_ids": ["RS121"]}, {"metric": "wells analyzed / optimization candidates", "value": "more than 2,000 analyzed; over 500 ranked for planned operating-mode optimization", "period": "1H 2026", "attribution": "Company-reported", "source_ids": ["RS122"]}], "partners": [], "verification": "verified primary", "key": "rosneft-echotools", "existing_ids": [], "company": "Rosneft", "title": "EchoTools AI well-diagnostic interpretation and operating-mode recommendations", "department": "Production & field operations", "stage": "Introduced July 2025; rollout and operating results updated in August 2026", "practical_relevance": "EchoTools converts acoustic well-survey interpretation into faster well diagnostics and candidate operating-mode changes. The 2026 disclosure indicates use on thousands of wells and a ranked optimization queue; Rosneft reports production gains but no independent validation.", "business_problem": "Speed up interpretation of well annulus fluid-level/acoustic data and identify operating-mode changes that may increase production.", "workflow": "AI determines fluid level and sound speed in the well annulus; these values feed bottomhole-pressure calculations and recommendations for well operating mode. The August 2026 update describes screening over 2,000 wells and ranking over 500 for planned optimization.", "technology_data": "AI software; model details, telemetry integration, uncertainty handling and production-system interfaces are not disclosed.", "integration_human_role": "The software recommends operating adjustments and identifies candidate wells; the public source does not describe approval, field execution or safety-control responsibilities.", "deployment_scope": "Rosneft reports over 2,000 wells analyzed in 1H 2026 and over 500 shortlisted for planned mode optimization; broader rollout scope is not stated.", "economics": "Rosneft reports additional oil output but gives no realized financial value, deployment cost or independent verification.", "timeline": [{"date": "2025-07-17", "event": "EchoTools AI workflow publicly described; this is the initial disclosure identified in the screen.", "source_ids": ["RS121"]}, {"date": "2026-08-20", "event": "Company reported first-half 2026 rollout/update metrics.", "source_ids": ["RS122"]}], "limitations": ["All performance and rollout figures are company-reported; methods, baselines and independent validation are not disclosed.", "The average daily-per-well uplift is not tied to a stated measurement period or cohort definition."], "assessment": "Analyst inference: among screened regional cases, EchoTools provides one of the clearest workflows and implementation metrics; evidence supports decision assistance and prioritization, not autonomous operation.", "sources": [{"id": "RS121", "title": "Rosneft improves well research methods (EchoTools)", "url": "https://rosneft.ru/press/news/item/222577", "publisher": "Rosneft", "date": "2025-07-17", "type": "company announcement", "locator": "EchoTools", "review_status": "excerpt reviewed", "supports": ["AI interprets fluid level and sound speed in the well annulus; the derived values support bottomhole-pressure calculation and selection of an operating mode.", "Rosneft says the software processes over 10,000 well indicators in about two minutes versus over three days manually, and recommendations yield an average additional 1 tonne/day per well."], "original_id": "regional-006"}, {"id": "RS122", "title": "Rosneft increases production through innovative software (EchoTools update)", "url": "https://www.rosneft.ru/press/news/item/224545/", "publisher": "Rosneft", "date": "2026-08-20", "type": "company announcement", "locator": "EchoTools rollout update", "review_status": "excerpt reviewed", "supports": ["Rosneft reports that in 1H 2026 EchoTools analyzed information from more than 2,000 wells and generated a ranking of over 500 candidate wells for operating-mode optimization.", "This is a later rollout update to the July 2025 named workflow, not its launch date."], "original_id": "regional-007"}], "field_sources": {"business_problem": ["RS121"], "workflow": ["RS121", "RS122"], "technology_data": ["RS121"], "integration_human_role": ["RS121"], "deployment_scope": ["RS122"], "economics": ["RS121"]}, "cohort": "regional", "packet_id": "regional-007", "source_ids": ["RS121", "RS122"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-186"], "related_existing_ids": []}, {"secondary_departments": [], "category": "Internal adoption", "disclosed_results": [{"metric": "routine workload reduction", "value": "70-80%", "period": "July 2025 company disclosure", "attribution": "Company-reported", "source_ids": ["RS123"]}], "partners": [], "verification": "verified primary", "key": "rosneft-llm-geology-assistant", "existing_ids": [], "company": "Rosneft", "title": "LLM virtual assistant for technical text search and documentation", "department": "Reservoir engineering & field development", "stage": "Company reported active use in July 2025", "practical_relevance": "This is a named internal GenAI workflow applied to domain work from geology through field-development design. The disclosed functionality remains knowledge-work assistance rather than autonomous engineering decisions.", "business_problem": "Reduce routine effort in finding, checking and formatting technical text information.", "workflow": "Employees submit simple text queries to a company-developed LLM assistant used for tasks ranging from geology to field-development design; Rosneft says it helps search, verify and format text information.", "technology_data": "Large language models are stated. Model, corpus, retrieval architecture, access controls and validation process are not disclosed.", "integration_human_role": "The assistant responds to user text queries; final engineering decisions remain unspecified. Human review should be assumed only as an analyst inference, not an explicit company control statement.", "deployment_scope": "Used across unspecified company tasks including geology and field-development design; user count and business-unit rollout are not disclosed.", "economics": "No spend, productivity baseline, headcount effect or financial return disclosed.", "timeline": [{"date": "2025-07-28", "event": "Rosneft announced the deployed LLM-based virtual assistant.", "source_ids": ["RS123"]}], "limitations": ["Company-reported workload reduction has no method or independent validation; use volume and controls are undisclosed."], "assessment": "Analyst inference: useful internal productivity application, but less operationally consequential and less auditable than the company’s well-diagnostic deployment.", "sources": [{"id": "RS123", "title": "Rosneft introduces new technologies using artificial intelligence", "url": "https://rosneft.ru/press/news/item/222604/", "publisher": "Rosneft", "date": "2025-07-28", "type": "company announcement", "locator": "LLM virtual assistant", "review_status": "excerpt reviewed", "supports": ["Company-developed LLM-based virtual assistant is used for text search, verification and formatting tasks, with applications from geology to field-development design.", "Rosneft claims 70-80% lower routine workload for these tasks; no independent validation or financial return is disclosed."], "original_id": "regional-008"}], "field_sources": {"business_problem": ["RS123"], "workflow": ["RS123"], "technology_data": ["RS123"], "integration_human_role": ["RS123"], "deployment_scope": ["RS123"], "economics": ["RS123"]}, "cohort": "regional", "packet_id": "regional-008", "source_ids": ["RS123"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-187"], "related_existing_ids": []}, {"key": "permian-resources-ai-subsurface-placement", "existing_ids": [], "company": "Permian Resources", "title": "AI-driven subsurface insight and precision wellbore placement", "department": "Reservoir engineering & field development", "secondary_departments": ["Exploration & geoscience", "Drilling & completions"], "category": "Internal adoption", "stage": "Described in Q3 2025 earnings presentation; operating scale and asset-specific deployment are not quantified.", "practical_relevance": "A concrete subsurface-to-well-placement workflow intended to expand play boundaries. The source does not establish its model design, deployment extent, or causal contribution to well outcomes.", "business_problem": "Expanding productive play boundaries requires subsurface interpretation and precise wellbore placement.", "workflow": "Permian Resources says it uses AI-driven subsurface insights with precision wellbore placement. The presentation places the claim in a discussion of technical execution, next to its Haley development update, but does not specify a decision sequence, geological inputs, or whether AI recommendations are advisory or automated.", "technology_data": "The company labels the insight AI-driven. It does not disclose model class, training data, geophysical/seismic inputs, validation approach, software, or external technology partners in the cited slide.", "integration_human_role": "Not disclosed. The source does not say who reviews the insight, how geoscientists or drilling engineers approve placement, or whether the system controls execution.", "deployment_scope": "Company presentation refers to play boundary expansion and discusses Permian operations. Number of wells/acreage or specific AI-assisted placement outcomes is not disclosed.", "partners": [], "disclosed_results": [], "economics": "No AI-specific economics disclosed. Haley wells were reported to outperform offsets by ~45%, but the presentation does not attribute this result to the AI workflow; it is excluded from AI results.", "timeline": [{"date": "2025-11-05", "event": "Q3 2025 investor presentation disclosed AI-driven subsurface insights and precision wellbore placement.", "source_ids": ["RS124"]}], "limitations": ["Single company presentation disclosure, with no model, data, validation, or deployment-count detail.", "Do not attribute adjacent Haley outperformance to AI."], "assessment": "Analyst inference: the described workflow is decision-support for subsurface interpretation and wellbore placement, potentially useful for testing play extensions. Practical value and operational maturity cannot be established from the slide alone.", "sources": [{"id": "RS124", "title": "Q3 2025 Earnings Presentation", "url": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "publisher": "Permian Resources", "date": "2025-11-05", "type": "investor presentation", "locator": "Slide 5, “Leading Execution, Cutting Edge Technology Drive Robust Results”", "review_status": "full text reviewed", "supports": ["Permian Resources said it was leveraging AI-driven subsurface insights and precision wellbore placement to expand play boundaries.", "The slide places this statement alongside its Haley development discussion; it separately reports Haley early-Q3 2025 startup and approximately 45% offset outperformance, without attributing that result to AI."], "accessed": "2026-09-27", "original_id": "a1-001"}], "field_sources": {"business_problem": ["RS124"], "workflow": ["RS124"], "technology_data": ["RS124"], "integration_human_role": ["RS124"], "deployment_scope": ["RS124"], "economics": ["RS124"]}, "verification": "verified primary", "corrections_to_existing": "", "cohort": "round2/additions-1", "packet_id": "round2-additions-1-001", "source_ids": ["RS124"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-188"], "related_existing_ids": []}, {"key": "kmg-refinery-ai-predictive-maintenance", "existing_ids": [], "company": "KazMunayGas", "title": "AI predictive-maintenance rollout across Kazakhstan refineries", "department": "Maintenance & reliability", "secondary_departments": ["Refining, LNG & process operations", "Refining, LNG & process operations"], "category": "Internal adoption", "stage": "2025 pilots at Atyrau and Pavlodar; a full-scale rollout wave began in 2026; ANPZ subsidiary detail describes a pilot launched in October (year not stated; its article is dated December 2025)", "practical_relevance": "This is a named asset-health workflow with subsidiary-level equipment examples and a parent-company rollout plan. It is more concrete than a generic digital program, but outcome data remain limited to an identified anomaly rather than verified avoided downtime or savings.", "business_problem": "Detect early equipment anomalies from operating data so refinery teams can investigate before failures cause unplanned shutdowns.", "workflow": "AI models analyze equipment/process parameters from sensor streams to identify abnormal behavior and alert staff. At ANPZ, the pilot covered equipment around catalytic cracking and CCR; the refinery reported a compressor anomaly was found in time to address it during scheduled work.", "technology_data": "KMG describes AI analysis of high-volume operating data from thousands of installed sensors for early anomaly detection. ANPZ says its wider plant digital ecosystem includes Honeywell PHD, MES and equipment monitoring systems; the release does not explicitly state the predictive model was integrated with each platform.", "integration_human_role": "The release describes alerts and recommendations for personnel to check anomalies and take action; human maintenance response remains part of the workflow.", "deployment_scope": "KMG says pilots were implemented at Atyrau and Pavlodar in 2025. Its May 2026 release says a full-scale wave began at both sites, with >100 critical units per refinery planned across three stages; Shymkent began initial compressor coverage in April 2026; 10 critical units are connected at the polypropylene plant. A separate ANPZ December 2025 article says 28 critical assets were listed for its first monitoring circuit in 2026. These are different rollout scopes/stages, not completed fleetwide deployment.", "partners": ["Honeywell (ANPZ pilot)"], "disclosed_results": [], "economics": "No monetary savings, avoided downtime, predictive accuracy or failure-reduction result disclosed. KMG’s only concrete operating example is an anomaly detected before scheduled maintenance; do not convert that example into quantified avoided loss.", "timeline": [{"date": "2025", "event": "KMG reports pilot projects at Atyrau and Pavlodar; its May 2026 release says an early-warning system for delayed-coking reactors was already functioning at Atyrau.", "source_ids": ["RS125"]}, {"date": "October (year omitted; article dated 2025-12-25)", "event": "ANPZ says a pilot with Honeywell launched in October and covered catalytic-cracking compressor and CCR equipment; it reports a compressor anomaly addressed during planned maintenance.", "source_ids": ["RS126"]}, {"date": "2026-04", "event": "KMG reports the initial predictive-analytics stage for critical compressors began at Shymkent refinery.", "source_ids": ["RS125"]}, {"date": "2026", "event": "KMG reports a full-scale rollout wave began at Atyrau and Pavlodar; >100 assets at each site remain a three-stage plan.", "source_ids": ["RS125"]}, {"date": "2026", "event": "ANPZ says 28 critical assets are listed for its initial monitoring circuit.", "source_ids": ["RS126"]}], "limitations": ["KMG and its subsidiary are the reporting parties; no independent performance validation located.", "ANPZ article omits the year of the October pilot start; December 2025 publication does not justify a stronger date claim.", "ANPZ’s 28-unit initial circuit and KMG’s >100-unit phased plans are different scopes and should not be summed or conflated.", "No model accuracy, avoided downtime, savings or spend disclosed."], "assessment": "Analyst inference: The initiative appears operationally relevant because it couples anomaly detection with refinery maintenance systems and has asset-level examples. Public evidence supports pilot activity and planned expansion, not quantified reliability economics.", "sources": [{"id": "RS125", "title": "КМГ внедряет предиктивную аналитику на НПЗ", "url": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "publisher": "KazMunayGas", "date": "2026-05-22", "type": "company announcement", "locator": "Russian-language release; refinery predictive analytics program", "review_status": "full text reviewed", "supports": ["KazMunayGas reports that predictive-maintenance pilots ran in 2025 at Atyrau and Pavlodar refineries; its Atyrau delayed-coking reactor early-warning system was operating at disclosure.", "KMG says 2026 deployment waves target more than 100 critical equipment units each at Atyrau and Pavlodar; Shymkent began an initial compressor-monitoring stage in April 2026, and the polypropylene plant connected 10 equipment units.", "The release describes sensor data and AI-based anomaly detection to alert staff before failure; no accuracy, avoided-downtime or financial return is quantified."], "original_id": "R2-001"}, {"id": "RS126", "title": "AI control of equipment at Atyrau refinery: predictive analytics in action", "url": "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/", "publisher": "Atyrau Oil Refinery (ANPZ)", "date": "2025-12-25", "type": "company announcement", "locator": "AI predictive analytics pilot; connected to Honeywell PHD/MES", "review_status": "full text reviewed", "supports": ["ANPZ says a predictive-failure-analysis pilot began in October (year omitted; article is dated 2025-12-25) with Honeywell; it applies AI to operating parameters and flags anomalies for staff follow-up.", "The refinery names equipment around the catalytic-cracking and CCR units; it reports the system surfaced a compressor issue in time for planned maintenance and a separate prime-pump pressure anomaly.", "The release describes 28 critical assets as a planned 2026 monitoring scope; it does not report accuracy, avoided downtime or savings."], "original_id": "R2-002"}], "field_sources": {"business_problem": ["RS125", "RS126"], "workflow": ["RS125", "RS126"], "technology_data": ["RS125", "RS126"], "integration_human_role": ["RS126"], "deployment_scope": ["RS125", "RS126"], "economics": ["RS125", "RS126"]}, "verification": "verified primary", "corrections_to_existing": "New KMG/ANPZ workflow supported by primary sources; year of ANPZ October pilot start is unstated.", "related_existing_ids": [], "cohort": "round2/additions-2", "packet_id": "round2-additions-2-001", "source_ids": ["RS125", "RS126"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-189"]}, {"key": "qatarenergy-lng-machinery-fleet-ai", "existing_ids": [], "company": "QatarEnergy LNG", "title": "AI pilot for rotating-equipment fleet predictive maintenance", "department": "Maintenance & reliability", "secondary_departments": ["Refining, LNG & process operations"], "category": "Internal adoption", "stage": "A successful pilot is described in the 2025 strategy material; expansion across assets is a stated plan", "practical_relevance": "The disclosure identifies a practical reliability problem and proposed asset-wide scaling, but omits pilot location, vendor, deployment date and measured impact.", "business_problem": "Use legacy equipment data to anticipate rotating-equipment maintenance needs and expand a successful AI pilot across the LNG asset fleet.", "workflow": "QatarEnergy LNG describes AI solutions for machinery-fleet predictive use based on legacy rotating-equipment data. Detailed alert, work-order or maintenance-decision steps are not disclosed.", "technology_data": "AI and historical/legacy rotating-equipment data are named. Model type, sensors, integrations and system architecture are not disclosed.", "integration_human_role": "Not disclosed beyond an intended predictive-maintenance use; whether maintenance personnel review alerts or systems generate work orders is not stated.", "deployment_scope": "The 2025 magazine calls this a successful pilot and proposes expansion across all assets; it does not name pilot asset(s), number of assets, or rollout milestones.", "partners": [], "disclosed_results": [], "economics": "No financial, reliability, avoided-failure or downtime metrics disclosed.", "timeline": [{"date": "2025-07", "event": "Company magazine lists pilot expansion across assets as a strategic initiative.", "source_ids": ["RS127"]}], "limitations": ["The document indicates a pilot and future expansion, not confirmed asset-wide deployment.", "Partner, location, AI technique, economics and operating metrics are not disclosed.", "The PDF is available and searchable in full; the relevant strategy passage is brief and does not disclose technical or outcome detail."], "assessment": "Analyst inference: This is a credible but low-detail proof of a specific LNG machinery-fleet use case. It signals a test and intent to scale, not evidence of an operational fleet-wide system.", "sources": [{"id": "RS127", "title": "The Pioneer 167 — QatarEnergy LNG", "url": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "publisher": "QatarEnergy LNG", "date": "2025-07", "type": "company magazine / strategy document", "locator": "The Pioneer 167, PDF p. 5, Corporate Strategic Initiatives: AI Solutions for Machinery Fleet", "review_status": "full text reviewed", "supports": ["QatarEnergy LNG’s 2025 strategic-plan page says the company will expand a successful AI pilot across all assets.", "It identifies legacy data use for predictive purposes in rotating equipment.", "The passage provides no pilot location, vendor, asset count, rollout milestones or measured results."], "original_id": "R2-003"}], "field_sources": {"business_problem": ["RS127"], "workflow": ["RS127"], "technology_data": ["RS127"], "integration_human_role": ["RS127"], "deployment_scope": ["RS127"], "economics": ["RS127"]}, "verification": "verified primary", "corrections_to_existing": "New workflow from the full 2025 QatarEnergy LNG magazine; the source discloses a successful pilot and planned expansion, but little implementation detail.", "related_existing_ids": [], "cohort": "round2/additions-2", "packet_id": "round2-additions-2-002", "source_ids": ["RS127"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-190"]}, {"key": "ambyint-infinitypl-methane-upgrade", "existing_ids": [], "company": "Ambyint", "title": "InfinityPL methane-mitigation upgrade for plunger-lift wells", "department": "Production & field operations", "secondary_departments": ["HSE & emissions"], "category": "Commercial AI product", "stage": "2025 ERA-funded development/upgrade; Tourmaline named as planned deployment partner; field deployment status of the upgraded capabilities not verified.", "practical_relevance": "A funded, domain-specific AI upgrade targets plunger-lift inefficiency and methane venting. Disclosure is unusually concrete about algorithms, partner, target workflow and prospective scale, but does not establish achieved emissions reductions.", "business_problem": "Plunger-lift wells can vent methane when operation is inefficient; continuous manual monitoring and intervention limit coverage.", "workflow": "Upgrade Ambyint’s cloud InfinityPL platform to predict inefficiencies, detect complex anomalies and reduce continuous manual intervention. Tourmaline plans to deploy the upgraded technology across a substantial number of Alberta plunger-lift sites.", "technology_data": "ERA describes advanced machine-learning and reinforcement-learning algorithms added to the existing AI-driven cloud platform. Specific input data, model validation and control guardrails are not disclosed.", "integration_human_role": "The project aims to reduce continuous manual intervention. Sources do not specify whether the upgraded system changes setpoints autonomously or describe operator approval, override, or safety controls.", "deployment_scope": "Existing platform reported deployed by Ambyint at 17 North American producers; new ERA-funded project identifies Tourmaline plunger-lift well sites in Alberta as planned deployment area. Existing footprint is not equivalent to deployment of upgraded methane functionality.", "partners": ["Emissions Reduction Alberta", "Tourmaline Oil"], "disclosed_results": [{"metric": "ERA funding", "value": "$4.1 million", "period": "2025 award", "attribution": "ERA-reported grant amount", "source_ids": ["RS128", "RS129"]}, {"metric": "Potential methane venting reduction", "value": "up to 90%", "period": "project target/estimate, not measured outcome", "attribution": "Company- and funder-reported prospective target", "source_ids": ["RS128", "RS129"]}], "economics": "Funding is for technology development; no customer savings or realized methane reduction for the upgraded system is reported.", "timeline": [{"date": "2025-04-16", "event": "ERA announced project support and Tourmaline’s planned deployment partnership.", "source_ids": ["RS129"]}, {"date": "2025-05-09", "event": "Ambyint described the funded InfinityPL upgrade.", "source_ids": ["RS128"]}], "limitations": ["The up-to-90% figure is an expected reduction, not a verified result.", "No Tourmaline deployment date, number of wells, baseline, measurement protocol, or post-deployment result is disclosed.", "Technical and safety-control details are not disclosed.", "Sources do not establish autonomous setpoint management or describe operator approval/override controls for the upgrade."], "assessment": "Analyst inference: A specific practical AI workstream with third-party funding and a named prospective operator deployment. Evidence supports development and intent, not live rollout or quantified impact.", "sources": [{"id": "RS128", "title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction", "url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "publisher": "Ambyint", "date": "2025-05-09", "type": "company announcement", "locator": "Full page", "review_status": "full text reviewed", "supports": ["Ambyint says $4.1m from ERA supports a next-generation InfinityPL platform upgrade.", "Company says the existing platform is deployed by 17 producers, and the project is expected to reduce methane venting by up to 90%.", "Tourmaline is named as key planned deployment partner at plunger-lift sites in Alberta."], "accessed": "2026-09-27", "original_id": "P3A001"}, {"id": "RS129", "title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector", "url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "publisher": "Emissions Reduction Alberta / Government of Alberta", "date": "2025-04-16", "type": "public funding announcement", "locator": "Ambyint project description", "review_status": "full text reviewed", "supports": ["ERA confirms $4.1m funding for Ambyint’s upgrade to the InfinityPL AI platform using advanced ML and reinforcement learning.", "ERA describes prediction of inefficiencies, complex-anomaly detection and reduced manual intervention; up to-90% methane-venting reduction remains expected.", "Tourmaline plans deployment across a substantial number of Alberta plunger-lift well sites."], "accessed": "2026-09-27", "original_id": "P3A002"}], "field_sources": {"business_problem": ["RS128", "RS129"], "workflow": ["RS128", "RS129"], "technology_data": ["RS129"], "integration_human_role": ["RS128", "RS129"], "deployment_scope": ["RS128", "RS129"], "economics": ["RS128", "RS129"]}, "verification": "verified primary", "corrections_to_existing": "No exact profile match found in the round-two baseline; addition is a distinct 2025 funded upgrade, not a claim that Ambyint’s established optimization platform was first launched in 2025. Public sources state $4.1 million and do not establish autonomous setpoint management.", "cohort": "round2/additions-3", "packet_id": "round2-additions-3-001", "source_ids": ["RS128", "RS129"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-191"], "related_existing_ids": []}, {"key": "kpler-ai-chartering-email-extraction", "existing_ids": [], "company": "Kpler", "title": "AI email parsing and vessel recommendation in chartering workflow", "department": "Trading & marketing", "secondary_departments": ["Supply chain & logistics"], "category": "Commercial AI product", "stage": "AI email-parsing and chartering-agent workflow presented in October 2025; source said AI chartering agent was planned to launch in Q4 2025. Current product materials describe AI email parsing, but dated production launch of the agent was not independently confirmed.", "practical_relevance": "Targets a specific operational bottleneck in maritime energy trading: turning high-volume unstructured broker/cargo communications into searchable lists and matching vessel candidates to commercial and risk constraints.", "business_problem": "Charterers and brokers manually process email, PDFs and spreadsheets containing cargo and vessel positions; delays and scattered records impede fixture decisions.", "workflow": "AI parsing converts unstructured emails, PDFs and spreadsheets into structured cargo/position lists. A planned chartering agent recommends vessel candidates using freight economics, risk and user-defined filters. Present product materials also describe AI extraction of vessel names, IMO numbers, cargo details and laycan dates into tonnage/cargo/fixture lists.", "technology_data": "Kpler describes AI-powered parsing across email/PDF/spreadsheet inputs and market, freight and risk data; underlying model, precision validation dataset and operating controls are undisclosed. Current product page cites 10+ years of proprietary historical data and real-time market inputs, but its undated statements are not treated as verified 2025 launch evidence.", "integration_human_role": "Outputs support chartering professionals’ analysis and selection; the event description does not say the agent autonomously fixes vessels or executes trades.", "deployment_scope": "Commercial product directed to shipowners, charterers and brokers across commodity markets, including oil, refined products and LNG; number of customers/users of AI features not disclosed.", "partners": [], "disclosed_results": [], "economics": "Kpler’s current product page promotes faster analysis but no attributable customer time-saved, cost or trading-return evidence was identified.", "timeline": [{"date": "2025-10-16", "event": "Kpler event described email parsing and previewed a chartering agent planned for Q4 launch.", "source_ids": ["RS130"]}, {"date": null, "event": "Current product page describes natural-language analysis and AI extraction in chartering; page publication date not stated.", "source_ids": ["RS131"]}], "limitations": ["The Q4 2025 agent was announced as planned; launch/commercial availability is not confirmed by dated primary evidence reviewed here.", "No evidence that recommendations execute trades or replace charterer approval.", "No email-parsing accuracy figure appears in the reviewed current product page; no independent accuracy or productivity metric was identified."], "assessment": "Analyst inference: A useful AI-enabled commercial workflow with visible data handoffs and a human decision-maker. Verification is strongest for announced features; rollout, accuracy and realized economics remain open.", "sources": [{"id": "RS130", "title": "Navigating Risk & Harnessing AI Briefing", "url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "publisher": "Kpler", "date": "2025-10-16", "type": "company announcement", "locator": "Harness AI in Chartering", "review_status": "full text reviewed", "supports": ["Kpler described AI-powered parsing of emails, PDFs and spreadsheets into structured cargo and position lists.", "Kpler previewed an AI chartering agent to recommend vessel candidates based on freight economics, risk and custom filters, and said it was launching Q4."], "accessed": "2026-09-27", "original_id": "P3K001"}, {"id": "RS131", "title": "Kpler AI", "url": "https://www.kpler.com/solutions/ai", "publisher": "Kpler", "date": null, "type": "product page", "locator": "Kpler Copilot / Kpler agents / FAQ", "review_status": "full text reviewed", "supports": ["Current Kpler page describes natural-language questions, AI-assisted market analysis, and chartering/cargo/compliance agent categories.", "Page describes sources including proprietary history, AIS, satellite tracking, customs/trade flows and refinery information."], "accessed": "2026-09-27", "original_id": "P3K002"}], "field_sources": {"business_problem": ["RS130"], "workflow": ["RS130", "RS131"], "technology_data": ["RS130", "RS131"], "integration_human_role": ["RS130"], "deployment_scope": ["RS130", "RS131"], "economics": []}, "verification": "verified primary", "corrections_to_existing": "No exact Kpler profile match found in the round-two baseline. Removed unsupported 99% accuracy claim; the dated event source supports the announced workflow and planned Q4 2025 launch, while current undated product material does not establish a numeric parsing-accuracy result.", "cohort": "round2/additions-3", "packet_id": "round2-additions-3-002", "source_ids": ["RS130", "RS131"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-192"], "related_existing_ids": []}, {"key": "ifs-kodiak-material-replenisher", "existing_ids": [], "company": "Kodiak Gas Services", "title": "Material Replenisher AI agent for compression-field technicians", "department": "Supply chain & logistics", "secondary_departments": ["Maintenance & reliability"], "category": "Internal adoption", "stage": "Rolled out by 2026-04-23, according to IFS; rollout date and usage scale not specified.", "practical_relevance": "A deployed conversational agent addresses the parts-search and ordering bottleneck for Kodiak field technicians. IFS reports annual ROI and technician hours returned after rollout; methodology and independent validation are not disclosed.", "business_problem": "Kodiak says its 800 technicians support 4.5 million horsepower of compression capacity and spent time searching for parts; the reviewed sources do not specify a 150+ warehouse network.", "workflow": "Technicians find and order parts conversationally through the Material Replenisher agent. The sources do not specify nearby-stock search, exact warehouse-location inputs, or all fulfillment steps.", "technology_data": "IFS identifies Material Replenisher as an IFS Loops agent and Kodiak as an IFS Cloud customer. The reviewed sources do not specify the exact inventory or warehouse data integrations, model, or retrieval architecture.", "integration_human_role": "Technicians initiate conversational parts requests; approval thresholds, exception handling, fulfillment confirmation and other controls are not described.", "deployment_scope": "IFS reports rollout at Kodiak; it describes a 4.5-million-horsepower fleet and 800 technicians. Warehouse count, sites covered, user adoption and operational rollout date are not disclosed.", "partners": ["IFS"], "disclosed_results": [{"metric": "Potential annual ROI and technician hours returned", "value": "$3 million and 90,000 hours", "period": "2025 conditional scenario: if half of workforce engages daily for one year", "attribution": "IFS-reported customer estimate; conditional and not a measured result", "source_ids": ["RS132"]}, {"metric": "Annual ROI and technician hours returned", "value": "$3 million and 90,000 hours", "period": "IFS described these as the result following Kodiak rollout in a release published 2026-04-23; methodology/timeframe not provided", "attribution": "Vendor-reported result; not independently validated", "source_ids": ["RS133"]}], "economics": "The November 2025 IFS release framed $3m annual ROI and 90,000 hours as a conditional scenario if half the technicians used the agent daily. In April 2026, IFS said Kodiak had rolled out the agent and called the same figures the result. This later claim is vendor-reported and lacks measurement methodology, timeframe, and independent validation.", "timeline": [{"date": "2025-11-27", "event": "IFS published a conditional half-workforce daily-use scenario for $3m annual ROI and 90,000 hours; this release does not establish rollout by this date.", "source_ids": ["RS132"]}, {"date": "2026-04-23", "event": "IFS said Kodiak had rolled out Material Replenisher and described $3m annual ROI / 90,000 hours returned as the result.", "source_ids": ["RS133"]}], "limitations": ["The April 2026 result is vendor-reported; no measurement method, baseline, timeframe or independent customer validation is provided.", "The 2025 release gives the same numbers as a conditional scenario and does not date the rollout.", "Warehouse count, nearby-stock search, data-system integration details, adoption rates and control boundaries are not disclosed in the linked sources."], "assessment": "Analyst inference: IFS reports a live customer workflow for conversational parts finding and ordering. Its 2026 release presents ROI and time returned as post-rollout results, but absent methodology these remain vendor-attributed rather than independently verified outcomes.", "sources": [{"id": "RS132", "title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI", "url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "publisher": "IFS", "date": "2025-11-27", "type": "company announcement", "locator": "Kodiak customer disclosure / CIO quote", "review_status": "full text reviewed", "supports": ["IFS reported Kodiak rolled out a Material Replenisher agent for technicians to find and order parts conversationally.", "CIO’s $3m and 90,000-hour figures are conditional on half the workforce engaging once a day; the 2025 release does not establish rollout date."], "accessed": "2026-09-27", "original_id": "P3IFS001"}, {"id": "RS133", "title": "IFS Loops launches Agent Studio", "url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "publisher": "IFS", "date": "2026-04-23", "type": "company announcement", "locator": "Kodiak paragraph", "review_status": "full text reviewed", "supports": ["IFS says Kodiak rolled out Material Replenisher; technicians find and order parts through conversation.", "IFS calls $3m annual ROI and 90,000 hours returned “the result” after describing rollout; this vendor claim lacks disclosed measurement methodology or independent validation."], "accessed": "2026-09-27", "original_id": "P3IFS002"}], "field_sources": {"business_problem": ["RS132", "RS133"], "workflow": ["RS132", "RS133"], "technology_data": ["RS132", "RS133"], "integration_human_role": ["RS132"], "deployment_scope": ["RS132", "RS133"], "economics": ["RS132", "RS133"]}, "verification": "verified partner", "corrections_to_existing": "No exact Kodiak Material Replenisher record was found in the baseline. The 2025 release gives a conditional estimate; the 2026 release reports rollout and calls the figures the result. Scope and realized-value details remain vendor-reported; removed unsupported warehouse-count and system-integration specifics.", "cohort": "round2/additions-3", "packet_id": "round2-additions-3-003", "source_ids": ["RS132", "RS133"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-193"], "related_existing_ids": []}, {"key": "expand-leucipa-lucy", "existing_ids": [], "company": "Expand Energy", "title": "Leucipa production optimization and Lucy assistant pilot", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployment beginning at Operations Support Center; Lucy pilot planned", "practical_relevance": "Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.", "business_problem": "Production teams need usable operating insights and coordinated field decisions.", "workflow": "Expand describes Leucipa for production management, optimization and forecasting, plus a Lucy generative-AI chat pilot for real-time production insight.", "technology_data": "Leucipa and a conversational generative-AI interface; model architecture and validation metrics not disclosed.", "integration_human_role": "Initial use is at the Operations Support Center. Decision authority and control-system write permissions are not disclosed.", "deployment_scope": "Field Operations and Production Engineering expansion is planned; completed rollout is not established.", "partners": ["Baker Hughes"], "disclosed_results": [], "economics": "Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.", "timeline": [{"date": "2026-03-03", "event": "Operator reports initial deployment and planned team expansion.", "source_ids": ["RS134"]}], "limitations": ["A deployment start and assistant pilot are different stages.", "No quantified production uplift or cost saving is established.", "Do not equate the broader vendor platform footprint with Expand’s actual deployment."], "assessment": "Analyst inference: The practical diligence question is whether operator insights change field actions and nominated supply reliably. Ask for adoption, exception handling and measured business outcomes.", "sources": [{"id": "RS134", "title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "publisher": "Expand Energy", "date": "2026-03-03", "type": "company announcement", "locator": "Leucipa deployment and Snowflake data-platform sections", "review_status": "full text reviewed", "supports": ["Operations Support Center deployment; Lucy pilot; Snowflake drilling ML and SCADA alerting."], "original_id": "root2-exe"}], "verification": "verified primary", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS134"], "workflow": ["RS134"], "technology_data": ["RS134"], "integration_human_role": ["RS134"], "deployment_scope": ["RS134"], "economics": ["RS134"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-001", "source_ids": ["RS134"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-194"], "related_existing_ids": []}, {"key": "expand-snowflake-drilling-ml", "existing_ids": [], "company": "Expand Energy", "title": "Snowflake-hosted drilling ML and operational alerts", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Operator-reported operational data platform and model hosting", "practical_relevance": "Analyst inference: This is a distinct drilling and alerting workflow alongside production optimization. Determine how alerts are validated and what happens when model confidence is low.", "business_problem": "Legacy information and field-data demands complicate operating analysis.", "workflow": "The platform hosts real-time drilling ML, while centralized operating and SCADA data supports issue detection and Operations Support Center alerts.", "technology_data": "Snowflake-hosted data and models; no model family, training set or predictive accuracy disclosed.", "integration_human_role": "Alerts go to the support center; automatic actuation is not established.", "deployment_scope": "Operator data platform; well and asset counts are not disclosed.", "partners": ["Snowflake"], "disclosed_results": [], "economics": "Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.", "timeline": [], "limitations": ["Data integration benefits cannot be attributed entirely to AI.", "No independently measured drilling or failure-rate improvement is provided."], "assessment": "Analyst inference: This is a distinct drilling and alerting workflow alongside production optimization. Determine how alerts are validated and what happens when model confidence is low.", "sources": [{"id": "RS134", "title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "publisher": "Expand Energy", "date": "2026-03-03", "type": "company announcement", "locator": "Leucipa deployment and Snowflake data-platform sections", "review_status": "full text reviewed", "supports": ["Operations Support Center deployment; Lucy pilot; Snowflake drilling ML and SCADA alerting."], "original_id": "root2-exe"}], "verification": "verified primary", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS134"], "workflow": ["RS134"], "technology_data": ["RS134"], "integration_human_role": ["RS134"], "deployment_scope": ["RS134"], "economics": ["RS134"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-002", "source_ids": ["RS134"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-195"], "related_existing_ids": []}, {"key": "viridien-evergreen-data", "existing_ids": [], "company": "Viridien", "title": "Evergreen AI-assisted subsurface data preparation", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Commercial AI product", "stage": "Commercial solution launched", "practical_relevance": "Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.", "business_problem": "Fragmented geoscience information requires repeated manual preparation before interpretation.", "workflow": "Evergreen classifies, curates, contextualizes and validates data, then ingests and standardizes it across files, records and applications.", "technology_data": "AI and domain data science cover structured, semi-structured and unstructured geoscience data; model identities are not disclosed.", "integration_human_role": "The product connects third-party applications across on-premise, cloud and hybrid environments. Human quality-control procedures are not specified.", "deployment_scope": "Marketed to subsurface asset teams; no named customer or installed footprint disclosed.", "partners": [], "disclosed_results": [], "economics": "Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.", "timeline": [{"date": "2026-06-10", "event": "Evergreen solution launch announced.", "source_ids": ["RS135"]}], "limitations": ["Data-readiness claims are vendor statements.", "No time-saving benchmark, accuracy test or contract economics provided."], "assessment": "Analyst inference: Repeated preparation is a concrete bottleneck, but usable data requires lineage and quality controls. A buyer should test whether errors are surfaced and whether interpretations remain traceable.", "sources": [{"id": "RS135", "title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "url": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "publisher": "Viridien", "date": "2026-06-10", "type": "company announcement", "locator": "Product launch description", "review_status": "full text reviewed", "supports": ["AI-assisted classification, curation, contextualization and validation of subsurface data."], "original_id": "root2-vir"}], "verification": "verified primary", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS135"], "workflow": ["RS135"], "technology_data": ["RS135"], "integration_human_role": ["RS135"], "deployment_scope": ["RS135"], "economics": ["RS135"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-003", "source_ids": ["RS135"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-196"], "related_existing_ids": []}, {"key": "tgs-seismic-foundation-contract", "existing_ids": [], "company": "TGS", "title": "Seismic Foundation Model contract for an international operator", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Commercial AI product", "stage": "Multiyear contract award; customer unnamed", "practical_relevance": "Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.", "business_problem": "Structural and stratigraphic interpretation is the target workflow.", "workflow": "TGS will develop and deliver AI foundation-model capabilities for selected exploration and development interpretation tasks.", "technology_data": "TGS SFM learns geological patterns from the company’s global multi-client seismic library.", "integration_human_role": "The announcement positions the model as interpreter support; review controls and model-acceptance criteria are not disclosed.", "deployment_scope": "An international oil company has contracted the capability; identity, asset locations and deployment schedule are undisclosed.", "partners": [], "disclosed_results": [], "economics": "Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.", "timeline": [{"date": "2026-08-18", "event": "Multiyear IOC award announced.", "source_ids": ["RS136"]}], "limitations": ["Contract award establishes commercial adoption, not measured interpretation quality.", "No model size, benchmark, reserve impact or financial terms disclosed."], "assessment": "Analyst inference: The commercial commitment makes this more concrete than a generic model launch. Technical diligence should test transfer to unseen geology and uncertainty communication.", "sources": [{"id": "RS136", "title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "url": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "publisher": "TGS", "date": "2026-08-18", "type": "company announcement", "locator": "Opening four paragraphs", "review_status": "full text reviewed", "supports": ["Unnamed IOC contract; structural and stratigraphic interpretation; training on global multi-client seismic library."], "original_id": "root2-tgs"}], "verification": "verified primary", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS136"], "workflow": ["RS136"], "technology_data": ["RS136"], "integration_human_role": ["RS136"], "deployment_scope": ["RS136"], "economics": ["RS136"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-004", "source_ids": ["RS136"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-197"], "related_existing_ids": []}, {"key": "woodside-trion-ai-drilling", "existing_ids": [], "company": "Woodside", "title": "Trion integrated well-delivery contract with AI-enabled drilling", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Contract awarded; services scheduled for early2026 in the cited disclosure", "practical_relevance": "Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.", "business_problem": "The ultra-deepwater Trion project requires coordinated delivery of complex wells.", "workflow": "SLB’s integrated contract includes AI-enabled drilling alongside directional services, logging, fluids, cementing and completions.", "technology_data": "The specific AI model and input design are not disclosed in the earnings bullet.", "integration_human_role": "Services are to be managed through SLB Performance Live delivery centers. Driller override and autonomous-control boundaries are unspecified.", "deployment_scope": "18 ultra-deepwater wells in Mexico; delivery completion is not established by the award disclosure.", "partners": ["SLB"], "disclosed_results": [{"metric": "Contracted well-delivery scope", "value": "18 ultra-deepwater wells", "period": "Award disclosed April2025", "attribution": "Vendor-reported", "source_ids": ["RS137"]}], "economics": "Project spending, operating cost and realized incremental return are not disclosed in the reviewed evidence.", "timeline": [{"date": "2025-04-25", "event": "Award disclosed in quarterly results.", "source_ids": ["RS137"]}, {"date": "2026 (planned)", "event": "Services scheduled to begin early in the year; completion not verified here.", "source_ids": ["RS137"]}], "limitations": ["The cited April2025 release is an award and forward schedule, not a current progress certificate.", "No AI-attributable efficiency or well-quality result is disclosed."], "assessment": "Analyst inference: A named asset and well count allow follow-up against execution milestones. Separate overall integrated-service performance from the incremental contribution of AI.", "sources": [{"id": "RS137", "title": "SLB Announces First-Quarter 2025 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "publisher": "SLB", "date": "2025-04-25", "type": "earnings release", "locator": "Core — Woodside Trion contract bullet", "review_status": "excerpt reviewed", "supports": ["18-well integrated contract with AI-enabled drilling, planned early-2026 services and Performance Live delivery."], "original_id": "root2-wood"}], "verification": "verified partner", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS137"], "workflow": ["RS137"], "technology_data": ["RS137"], "integration_human_role": ["RS137"], "deployment_scope": ["RS137"], "economics": ["RS137"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-005", "source_ids": ["RS137"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-198"], "related_existing_ids": []}, {"key": "technip-dap-ai-for-all", "existing_ids": [], "company": "Technip Energies", "title": "Digital Acceleration Plan and AI for All workforce program", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Management reports program implementation and adoption; financial ambition remains a target", "practical_relevance": "Analyst inference: Combining adoption with project-delivery reform is relevant, but broad targets need a bridge to named workflows and attributable cash outcomes.", "business_problem": "The program targets project-execution productivity and value erosion.", "workflow": "The Digital Acceleration Plan applies advanced analytics and LLMs; AI for All builds employee skills for everyday use.", "technology_data": "Management identifies analytics and large language models; individual model architecture and data permissions are not provided.", "integration_human_role": "An employee-skills program accompanies deployment. Named workflow owners, assurance process and completion rates are not disclosed.", "deployment_scope": "Company-level initiative described in 2025 shareholder messages; individual site coverage undisclosed.", "partners": [], "disclosed_results": [], "economics": "Management targets €100 million in cost savings and incremental revenue by2028 across the Digital Acceleration Plan; spending and AI-only return are undisclosed.", "timeline": [{"date": "2025", "event": "Shareholder message reports implementation of AI for All and progress on the digital plan.", "source_ids": ["RS138"]}], "limitations": ["The web page does not expose an exact publication date.", "Financial target combines digital cost savings and incremental revenue; it is not AI-only realized savings.", "Do not treat the workforce program as evidence all employees are active AI users."], "assessment": "Analyst inference: Combining adoption with project-delivery reform is relevant, but broad targets need a bridge to named workflows and attributable cash outcomes.", "sources": [{"id": "RS138", "title": "Technip Energies shareholder information — 2025 chair and CEO messages", "url": "https://investors.technipenergies.com/shareholder-information", "publisher": "Technip Energies", "date": null, "type": "annual report", "locator": "Digital technology; Driving Innovation and Technological Leadership; Looking Ahead", "review_status": "excerpt reviewed", "supports": ["Digital Acceleration Plan using advanced analytics and LLMs; AI for All; €100m savings and incremental-revenue target by2028."], "original_id": "root2-ten"}], "verification": "verified primary", "corrections_to_existing": "New second-pass research entry; no earlier initiative ID duplicated.", "field_sources": {"business_problem": ["RS138"], "workflow": ["RS138"], "technology_data": ["RS138"], "integration_human_role": ["RS138"], "deployment_scope": ["RS138"], "economics": ["RS138"]}, "cohort": "round2/additions-root", "packet_id": "round2-additions-root-006", "source_ids": ["RS138"], "report_date": "2026-09-27", "evidence_note": "Source review establishes what was disclosed. Performance remains company or vendor reported unless explicitly identified otherwise.", "record_ids": ["AI-199"], "related_existing_ids": []}, {"key": "followup-ai-200", "packet_id": "followup-ai-200", "record_ids": ["AI-200"], "existing_ids": [], "related_existing_ids": [], "company": "Shaya Ecuador / SLB", "title": "Neuro autonomous geosteering", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "One Ecuador well disclosed in December 2024", "practical_relevance": "A named operating workflow with explicit source limits.", "business_problem": "", "workflow": "AI interprets live downhole measurements and changes the well trajectory.", "technology_data": "", "integration_human_role": "", "deployment_scope": "", "partners": [], "disclosed_results": [{"metric": "Reported operating evidence", "value": "25 autonomous trajectory changes across a 2,392-ft lateral.", "period": "One Ecuador well disclosed in December 2024", "attribution": "Provider-reported", "source_ids": ["FU002"]}], "economics": "", "timeline": [], "limitations": ["SLB-reported; no numeric production uplift, control comparison or fleet-scale outcome disclosed."], "assessment": "", "sources": [{"id": "FU002", "title": "SLB introduces autonomous AI geosteering", "url": "https://www.slb.com/newsroom/press-release/2024/slb-adds-ai-geosteering", "publisher": "SLB", "date": "2024-12-09", "type": "Provider press release", "locator": "Full article / relevant case section", "review_status": "full text reviewed", "supports": ["25 autonomous trajectory changes across a 2,392-ft lateral."], "accessed": "2026-09-28"}], "source_ids": ["FU002"], "field_sources": {"workflow": ["FU002"]}, "verification": "verified partner", "cohort": "Opportunity follow-up", "report_date": "2026-09-28", "evidence_note": "Source review confirms the statement, not independent performance."}, {"key": "followup-ai-201", "packet_id": "followup-ai-201", "record_ids": ["AI-201"], "existing_ids": [], "related_existing_ids": [], "company": "Chevron / OPX Ai", "title": "Kaybob integrated AI operations and lift control", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployment began 2020; live by early 2022; retrospective disclosure April 2026", "practical_relevance": "A named operating workflow with explicit source limits.", "business_problem": "", "workflow": "IOCaaS links SCADA and historian data to surveillance and bounded gas-lift actions.", "technology_data": "", "integration_human_role": "", "deployment_scope": "", "partners": [], "disclosed_results": [{"metric": "Reported operating evidence", "value": "5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.", "period": "Deployment began 2020; live by early 2022; retrospective disclosure April 2026", "attribution": "Provider-reported", "source_ids": ["FU003"]}], "economics": "", "timeline": [], "limitations": ["Vendor-authored retrospective. Facility changes and commodity effects also influenced economics; no controlled AI-only estimate."], "assessment": "", "sources": [{"id": "FU003", "title": "Field deployments of AI-based IOCaaS", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "publisher": "SPE / JPT; OPX Ai authors", "date": "2026-04-01", "type": "Vendor-authored technical case article", "locator": "Full article / relevant case section", "review_status": "full text reviewed", "supports": ["5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year."], "accessed": "2026-09-28"}], "source_ids": ["FU003"], "field_sources": {"workflow": ["FU003"]}, "verification": "verified partner", "cohort": "Opportunity follow-up", "report_date": "2026-09-28", "evidence_note": "Source review confirms the statement, not independent performance."}, {"key": "followup-ai-202", "packet_id": "followup-ai-202", "record_ids": ["AI-202"], "existing_ids": [], "related_existing_ids": [], "company": "ConocoPhillips / OPX Ai", "title": "Montney AI surveillance and flow assurance", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Four-month implementation; calendar operating dates not specified; disclosed April 2026", "practical_relevance": "A named operating workflow with explicit source limits.", "business_problem": "", "workflow": "IOCaaS monitors lift, hydrate risk and compressor condition.", "technology_data": "", "integration_human_role": "", "deployment_scope": "", "partners": [], "disclosed_results": [{"metric": "Reported operating evidence", "value": "3–4% production above forecast after four months of operation; about 5% lower LOE.", "period": "Four-month implementation; calendar operating dates not specified; disclosed April 2026", "attribution": "Provider-reported", "source_ids": ["FU003"]}], "economics": "", "timeline": [], "limitations": ["Vendor-authored partial-year comparison. A 6% production uplift was prospective, not the observed result."], "assessment": "", "sources": [{"id": "FU003", "title": "Field deployments of AI-based IOCaaS", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "publisher": "SPE / JPT; OPX Ai authors", "date": "2026-04-01", "type": "Vendor-authored technical case article", "locator": "Full article / relevant case section", "review_status": "full text reviewed", "supports": ["3–4% production above forecast after four months of operation; about 5% lower LOE."], "accessed": "2026-09-28"}], "source_ids": ["FU003"], "field_sources": {"workflow": ["FU003"]}, "verification": "verified partner", "cohort": "Opportunity follow-up", "report_date": "2026-09-28", "evidence_note": "Source review confirms the statement, not independent performance."}, {"key": "missing-sept29-ai-203", "packet_id": "missing-sept29-ai-203", "record_ids": ["AI-203"], "existing_ids": [], "related_existing_ids": [], "company": "Marathon Pipe Line (MPLX / Marathon Petroleum)", "title": "AI-assisted pipeline right-of-way surveillance", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Deployment stated; expanding across operations", "practical_relevance": "MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.", "business_problem": "", "workflow": "Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.", "technology_data": "MPL says it adopted AI-driven monitoring technology and is deploying it across operations. The disclosure does not identify the vendor, sensor/data architecture, alert handling, or whether imagery comes from aircraft, satellites, or other sources.", "integration_human_role": "", "deployment_scope": "Deployment stated; expanding across operations", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No quantified detection, incident reduction, or cost result disclosed.", "period": "2025-04-09", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-001"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-04-09", "event": "Deployment stated; expanding across operations", "source_ids": ["M29-001"]}], "limitations": ["Company interview from a senior MPL field-services/planning executive; operational claims are company-reported and lack independent performance data or specific rollout coverage."], "assessment": "MPL/MPC absent from the existing 202-record initiative register; this is an explicit operational AI use case, not the earlier 2024 general statement that MPC was deploying AI/ML.", "sources": [{"id": "M29-001", "title": "Our People: How Paige Chenevey drives success at Marathon Pipe Line", "url": "https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/", "publisher": "Marathon Pipe Line (MPLX / Marathon Petroleum)", "date": "2025-04-09", "type": "Operator interview / company news", "locator": "Paragraphs under ‘What drives MPL’s success in safety, innovation and sustainability?’, especially lines 43–44 in web text; page dates article Apr 9 2025", "review_status": "Full company page opened; article text directly accessible.", "supports": ["Monitor pipeline safety threats along rights-of-way that aerial patrol may miss, including land movement, deep tilling and leaks.", "No quantified detection, incident reduction, or cost result disclosed."], "accessed": "2026-09-29"}], "source_ids": ["M29-001"], "field_sources": {"workflow": ["M29-001"], "technology_data": ["M29-001"], "deployment_scope": ["M29-001"], "practical_relevance": ["M29-001"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-204", "packet_id": "missing-sept29-ai-204", "record_ids": ["AI-204"], "existing_ids": [], "related_existing_ids": [], "company": "ORLEN Group", "title": "Group-wide AI rollout with market-monitoring and supply-chain assistants", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Rollout / implementation underway; use cases partly prospective", "practical_relevance": "ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.", "business_problem": "", "workflow": "Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.", "technology_data": "Microsoft partnership covers consolidated licensing, secure cloud and intelligent assistants. ORLEN states solutions are being implemented and names Microsoft Copilot and Agentic AI; planned AI competence centre is intended to identify and scale higher-impact applications. This is enterprise deployment/rollout, not a single autonomous O&G control loop.", "integration_human_role": "", "deployment_scope": "Rollout / implementation underway; use cases partly prospective", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed.", "period": "2025-07-08", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-002"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-07-08", "event": "Rollout / implementation underway; use cases partly prospective", "source_ids": ["M29-002"]}], "limitations": ["Company announcement states scope and intent; no measured outcomes and no specific production asset use case attributed to the new AI assistants."], "assessment": "ORLEN absent from existing register; explicit AI rollout with O&G market and supply-chain workflows gives it an integrated-operator initiative not represented in the current corpus.", "sources": [{"id": "M29-002", "title": "ORLEN’s transformation to gain momentum with CEE’s largest AI deployment", "url": "https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment", "publisher": "ORLEN Group", "date": "2025-07-08", "type": "Operator press release", "locator": "Web lines 312–323: AI deployment objectives; Copilot and Agentic AI; use cases monitoring global oil and gas market data and streamlining supply-chain logistics; internal competence centre to identify, develop and scale projects.", "review_status": "Full English company announcement opened; exact application wording accessible.", "supports": ["Assist staff with recurring information work and data-supported decisions, including monitoring strategic oil and gas market data and supply-chain logistics; broader uses include energy operations, production, cybersecurity and corporate controls.", "More than 60,000 employees and 200+ companies describes ORLEN Group, not AI adoption. No operating or productivity outcome disclosed."], "accessed": "2026-09-29"}], "source_ids": ["M29-002"], "field_sources": {"workflow": ["M29-002"], "technology_data": ["M29-002"], "deployment_scope": ["M29-002"], "practical_relevance": ["M29-002"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-205", "packet_id": "missing-sept29-ai-205", "record_ids": ["AI-205"], "existing_ids": [], "related_existing_ids": [], "company": "Apache Corporation / APA Corporation", "title": "AI use in emissions reporting, logistics and reservoir-management decisions", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Use asserted as ongoing; specific maturity and deployment status unclear", "practical_relevance": "Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.", "business_problem": "", "workflow": "Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.", "technology_data": "The interview names three functions and says Apache works with service and high-tech partners; no product, model, operating asset, deployment stage by use case, or implementation detail is supplied.", "integration_human_role": "", "deployment_scope": "Use asserted as ongoing; specific maturity and deployment status unclear", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "None disclosed.", "period": "2024-11", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-003"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2024-11", "event": "Use asserted as ongoing; specific maturity and deployment status unclear", "source_ids": ["M29-003"]}], "limitations": ["Company interview in Egypt Oil & Gas industry supplement. It names AI functions (emissions reporting, supply-chain logistics, reservoir decisions) but gives no implementation example, product, asset by use case, maturity by use case, or measured result. These are high-level company-reported claims."], "assessment": "Apache/APA absent from current register. New candidate for upstream operator breadth, with low confidence on implementation detail; retain as a thin disclosure rather than inflate impact.", "sources": [{"id": "M29-003", "title": "Egypt Leading Oil & Gas Excellence Supplement — November 2024 (Apache interview)", "url": "https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence-1.pdf", "publisher": "Apache Corporation / APA Corporation", "date": "2024-11", "type": "Operator executive interview in industry supplement / original report", "locator": "Printed p. 14 (PDF page 14 / extracted web PDF P13), interview with Greg McDaniel, VP Apache Egypt Assets and Country Manager; technology question/answer lines 826–837. Supplement imprint: November 2024 at printed p. 3 / PDF P2; publisher publication listing corroborates.", "review_status": "Full original PDF opened; Apache interview, AI statements and November 2024 imprint directly reviewed. Publisher lists “Egypt Leading Oil & Gas Excellence – November 2024”.", "supports": ["Support emissions measurement/reporting, supply-chain logistics optimization and reservoir-management decisions with AI.", "None disclosed."], "accessed": "2026-09-29"}], "source_ids": ["M29-003"], "field_sources": {"workflow": ["M29-003"], "technology_data": ["M29-003"], "deployment_scope": ["M29-003"], "practical_relevance": ["M29-003"]}, "verification": "verified partner", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-206", "packet_id": "missing-sept29-ai-206", "record_ids": ["AI-206"], "existing_ids": [], "related_existing_ids": [], "company": "Murphy Oil Corporation", "title": "AI/ML-assisted continuous cyber monitoring and response", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "In use", "practical_relevance": "Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.", "business_problem": "", "workflow": "Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.", "technology_data": "Murphy reports internal and third-party AI/ML-backed monitoring and analytics within a layered cyber stack; AI use is explicit but it is a security workflow, not hydrocarbon production optimization.", "integration_human_role": "", "deployment_scope": "In use", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome.", "period": "2025-08-06", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-004"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-08-06", "event": "In use", "source_ids": ["M29-004"]}], "limitations": ["The AI/ML use case is cybersecurity continuous monitoring/detection and response, not hydrocarbon production optimization. The report does not isolate model role, vendors, OT scope or performance. Separate GenAI Hub/Copilot deployment is reported on p. 79."], "assessment": "Murphy absent from register. Qualifies as a practical operator AI workflow at enterprise/cyber boundary; keep segmented from core operational AI findings.", "sources": [{"id": "M29-004", "title": "2025 Murphy Sustainability Report", "url": "https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf", "publisher": "Murphy Oil Corporation", "date": "2025-08-06", "type": "Operator sustainability report", "locator": "Printed p. 78 (PDF page index P77), “Risk Management Strategy” Technology column: continuous monitoring and analytics built on advanced machine learning and AI; tools include endpoint/email protection and Managed Detection and Response. Printed p. 79 (PDF index P78) describes the Murphy Generative AI Hub and enterprise Microsoft Copilot rollout. Report printed p. 3 (PDF P2) gives publication date August 6, 2025; issuer Q2 announcement/update also dates Aug. 6 and confirms the report was published.", "review_status": "Full original 112-page issuer PDF opened; page 78 cyber AI and page 79 GenAI text reviewed. Report itself states publication date Aug. 6, 2025 on p. 3; issuer announcement corroborates.", "supports": ["Continuous cyber-threat monitoring, analytics and detection/response to safeguard company systems and data.", "No AI-attributed detection or incident metrics disclosed; report notes 24/7/365 availability for business-critical systems in separate text, not as an AI outcome."], "accessed": "2026-09-29"}], "source_ids": ["M29-004"], "field_sources": {"workflow": ["M29-004"], "technology_data": ["M29-004"], "deployment_scope": ["M29-004"], "practical_relevance": ["M29-004"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-207", "packet_id": "missing-sept29-ai-207", "record_ids": ["AI-207"], "existing_ids": [], "related_existing_ids": [], "company": "TotalEnergies", "title": "Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.", "practical_relevance": "A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.", "business_problem": "", "workflow": "Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.", "technology_data": "AI assistant built on Honeywell Experion distributed control system; it provides predictive insights to operators. The disclosure describes operator corrective action, not autonomous process control.", "integration_human_role": "", "deployment_scope": "Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions.", "period": "2025-11-11", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-005"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-11-11", "event": "Initial pilot; preliminary results disclosed, with continued pilot collaboration announced.", "source_ids": ["M29-005"]}], "limitations": ["Preliminary, company-reported; sample size is five events, with no false-alarm rate, missed-event rate, production or emissions quantity, counterfactual, deployment cost or payback. The source does not establish autonomous control authority."], "assessment": "A named refinery-unit pilot with a measurable lead time and clear human action boundary; useful follow-up evidence but too early and small-sample to support reliability or return claims.", "sources": [{"id": "M29-005", "title": "TotalEnergies and Honeywell Pilot AI-Assisted Control Room to Accelerated Shift to Industrial Autonomy", "url": "https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial", "publisher": "TotalEnergies", "date": "2025-11-11", "type": "Operator-hosted joint pilot announcement", "locator": "Lines 11–19; five event predictions and 12-minute lead at line 17", "review_status": "Full primary announcement reviewed.", "supports": ["Control-room operators use predictive insights to anticipate maintenance or upset events in a delayed coking unit and intervene before alarm incidents.", "Five potential events forecast an average of 12 minutes before alarm incidents; company says timely corrective action helped minimize downtime and flaring emissions."], "accessed": "2026-09-29"}], "source_ids": ["M29-005"], "field_sources": {"workflow": ["M29-005"], "technology_data": ["M29-005"], "deployment_scope": ["M29-005"], "practical_relevance": ["M29-005"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-208", "packet_id": "missing-sept29-ai-208", "record_ids": ["AI-208"], "existing_ids": [], "related_existing_ids": [], "company": "Snam / Engineering Group", "title": "AI extraction of gas-plant design data and equipment relationships", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Research & development", "stage": "Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.", "practical_relevance": "Adds a named midstream engineering-document workflow and its technical limitations.", "business_problem": "", "workflow": "Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.", "technology_data": "Combines OCR, language models, diagram recognition and constrained optimization. Snam supplied documents for approximately 1,500 regulating/metering plants; this is a dataset size, not a deployment count.", "integration_human_role": "", "deployment_scope": "Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics.", "period": "2025-11-05", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-006"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-11-05", "event": "Developed and evaluated; authors describe applicability as user support. Routine fleet-wide use is not established.", "source_ids": ["M29-006"]}], "limitations": ["The abstract gives different aggregate accuracy figures. Complex diagrams remain a weakness. No independently validated performance, cash saving, measured time saving or fleet rollout is established; the 11,000-plant fleet is contextual."], "assessment": "Adds a named midstream engineering-document workflow and its technical limitations.", "sources": [{"id": "M29-006", "title": "Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants", "url": "https://arxiv.org/html/2511.08609v1", "publisher": "Snam / Engineering Group", "date": "2025-11-05", "type": "Operator/provider-authored technical preprint", "locator": "Sections 3.1, 4 (Tables 1–3) and 5; abstract uses different summary accuracy figures", "review_status": "Full original preprint reviewed; not represented as a peer-reviewed deployment study.", "supports": ["Convert design PDFs, equipment lists and P&IDs into structured plant data for Snam's MGM users.", "Table 3 reports 95% component identification and 90% correct assignment to sections; these are task-specific evaluation metrics."], "accessed": "2026-09-29"}], "source_ids": ["M29-006"], "field_sources": {"workflow": ["M29-006"], "technology_data": ["M29-006"], "deployment_scope": ["M29-006"], "practical_relevance": ["M29-006"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-209", "packet_id": "missing-sept29-ai-209", "record_ids": ["AI-209"], "existing_ids": [], "related_existing_ids": [], "company": "PETRONAS", "title": "PROTEAN predictive rotating-equipment analytics", "department": "Maintenance & reliability", "secondary_departments": [], "category": "Internal adoption", "stage": "Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.", "practical_relevance": "New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.", "business_problem": "", "workflow": "Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.", "technology_data": "PETRONAS reports deploying PROTEAN across Malaysia and international assets to monitor more than 350 critical rotating equipment items.", "integration_human_role": "", "deployment_scope": "Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": ">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given.", "period": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "attribution": "Source-reported; see evidence limits", "source_ids": ["RS017"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "event": "Company describes the system as deployed in 2025 and links it to reliability and maintenance outcomes.", "source_ids": ["RS017"]}], "limitations": ["The report establishes deployment breadth, not the number of sites, time in service, model performance, false alarms, realized avoided downtime or savings. It gives no implementation cost or payback and does not quantify the claimed outcome in the text."], "assessment": "New operator-primary follow-up: named system and cross-geography scale, with operating outcome claims but no quantified post-rollout results or deployment economics.", "sources": [{"id": "RS017", "title": "PETRONAS Integrated Report 2025", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "publisher": "PETRONAS", "date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "type": "Operator integrated report", "locator": "Printed pp. 54–55 (PDF page 29), Predicting Failures Before They Happen", "review_status": "Full original report downloaded and relevant spread reviewed; printed pp. 54–55, PDF page 29.", "supports": ["Use machine learning to predict failures of critical rotating equipment and inform maintenance to reduce unplanned downtime and production losses.", ">350 critical rotating equipment items monitored; no avoided-failure, downtime, production, maintenance-spend or accuracy metric is given."], "accessed": "2026-09-29"}], "source_ids": ["RS017"], "field_sources": {"workflow": ["RS017"], "technology_data": ["RS017"], "deployment_scope": ["RS017"], "practical_relevance": ["RS017"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-210", "packet_id": "missing-sept29-ai-210", "record_ids": ["AI-210"], "existing_ids": [], "related_existing_ids": [], "company": "Diversified Energy Company (DEC)", "title": "OneDEC producer-statement automation and contract repository for revenue accounting and settlements", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Internal adoption", "stage": "Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.", "practical_relevance": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "business_problem": "", "workflow": "Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.", "technology_data": "OneDEC connects governed data in Snowflake with ServiceNow approvals and RAG agents. Human oversight and monitored/versioned outputs are described at platform level; the AI contribution to each financial workflow is not isolated.", "integration_human_role": "", "deployment_scope": "Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.", "period": "2026-05-28", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-007"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2026-05-28", "event": "Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.", "source_ids": ["M29-007"]}], "limitations": ["No quantified financial outcome or workflow-specific deployment denominator. Platform architecture does not prove every statement or contract step uses AI."], "assessment": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "sources": [{"id": "M29-007", "title": "Diversified Energy 2025 Sustainability Report", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "publisher": "Diversified Energy Company (DEC)", "date": "2026-05-28", "type": "Operator-published sustainability report; issuer release confirms report date", "locator": "Printed p. 64 (PDF physical page 63), Artificial Intelligence / AI Efficiency at Scale; issuer release dated 28 May 2026", "review_status": "Full original 85-page operator report reviewed at PDF p.63; primary investor-relations announcement verifies report publication date.", "supports": ["Producer-statement digitization and normalization supports accounting, marketing and revenue distribution. A shared contract repository supplies pricing and settlement terms. Both appear in the report’s AI efficiency discussion.", "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed."], "accessed": "2026-09-29"}], "source_ids": ["M29-007"], "field_sources": {"workflow": ["M29-007"], "technology_data": ["M29-007"], "deployment_scope": ["M29-007"], "practical_relevance": ["M29-007"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-211", "packet_id": "missing-sept29-ai-211", "record_ids": ["AI-211"], "existing_ids": [], "related_existing_ids": [], "company": "Trafigura", "title": "AI-enabled maritime voyage optimization across controlled fleet", "department": "Supply chain & logistics", "secondary_departments": [], "category": "Internal adoption", "stage": "Planned fleet rollout in July 2025; completion and active vessel count unverified.", "practical_relevance": "No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.", "business_problem": "", "workflow": "AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.", "technology_data": "Trafigura announced a ZeroNorth alliance and equity investment, with rollout intended across more than 350 controlled vessels. Human approvals and control authority are not described.", "integration_human_role": "", "deployment_scope": "Planned fleet rollout in July 2025; completion and active vessel count unverified.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population.", "period": "2025-07-09", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-008"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-07-09", "event": "Planned fleet rollout in July 2025; completion and active vessel count unverified.", "source_ids": ["M29-008"]}], "limitations": ["Future-tense announcement. Fleet size is not a verified AI usage count, and expected savings are not operating results."], "assessment": "No Trafigura initiative appeared in the full existing company register, source library, 177 research packets, or coverage packet. This is a named energy-trader workflow that combines trading logistics, shipping, and environmental performance.", "sources": [{"id": "M29-008", "title": "Trafigura announces strategic alliance with maritime technology provider ZeroNorth", "url": "https://www.trafigura.com/news-and-insights/press-releases/2025/trafigura-announces-strategic-alliance-with-maritime-technology-provider-zeronorth/", "publisher": "Trafigura", "date": "2025-07-09", "type": "Operator press release with named partner and attributed Global Head of Shipping", "locator": "Trafigura press release, 9 July 2025, paragraphs 2-5", "review_status": "Full original Trafigura press release opened and read; date and rollout phrasing verified in source.", "supports": ["AI uses weather, vessel-performance, specification and fuel data to support voyage optimization; emissions analytics and reporting are also included.", "No achieved fuel, emissions or cost saving disclosed; 350+ vessels is a rollout target population."], "accessed": "2026-09-29"}], "source_ids": ["M29-008"], "field_sources": {"workflow": ["M29-008"], "technology_data": ["M29-008"], "deployment_scope": ["M29-008"], "practical_relevance": ["M29-008"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-212", "packet_id": "missing-sept29-ai-212", "record_ids": ["AI-212"], "existing_ids": [], "related_existing_ids": [], "company": "SM Energy", "title": "Machine-learning gas-lift injection-rate optimization in Midland Basin production operations", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.", "practical_relevance": "SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.", "business_problem": "", "workflow": "A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.", "technology_data": "Management describes a compression-linked analytics platform and engineers applying learned practices. Vendor, well count, execution controls and validation method are undisclosed.", "integration_human_role": "", "deployment_scope": "Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "Management reports incremental production without a numerical uplift, baseline or measurement period.", "period": "2025-11-21", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-009"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2025-11-21", "event": "Use in Midland Basin asserted in November 2025; proposed Civitas extension remained prospective.", "source_ids": ["M29-009"]}], "limitations": ["Company-authored conference transcript; no independently measured result. Recommendations do not establish autonomous valve control. Civitas extension was prospective."], "assessment": "SM Energy is absent from the full existing register and research packets. The source names a concrete production workflow, data-learning loop, physical control lever, and actual operating basin, not a generic AI commitment.", "sources": [{"id": "M29-009", "title": "SM Energy Company — Fireside Chat Stephens Annual Investment Conference (internally generated transcript; Rule 425 filing)", "url": "https://www.sec.gov/Archives/edgar/data/893538/000089353825000160/stephensannualinvestorco.htm", "publisher": "SM Energy", "date": "2025-11-21", "type": "Company conference transcript filed with SEC under Rule 425", "locator": "Transcript p. 7, COO Beth McDonald; conference 20 November, filing 21 November 2025", "review_status": "Full SEC EDGAR transcript and issuer-hosted copy independently checked.", "supports": ["A machine-learning model learns from step-rate tests and suggests gas-lift injection rates for Midland Basin wells.", "Management reports incremental production without a numerical uplift, baseline or measurement period."], "accessed": "2026-09-29"}], "source_ids": ["M29-009"], "field_sources": {"workflow": ["M29-009"], "technology_data": ["M29-009"], "deployment_scope": ["M29-009"], "practical_relevance": ["M29-009"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-165", "packet_id": "missing-sept29-ai-165", "record_ids": ["AI-165"], "existing_ids": ["AI-165"], "related_existing_ids": [], "company": "Baker Hughes", "title": "AI sourcing optimization across enterprise procurement (implementation-detail enrichment)", "department": "Supply chain & logistics", "secondary_departments": [], "category": "Internal adoption", "stage": "Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.", "practical_relevance": "AI-165 already held a concise profile, but omitted the operational mechanics, feedback/validation controls, scale path, data footprint, user count, and the distinction between an opportunity acceptance rate and realized dollar benefit.", "business_problem": "", "workflow": "Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.", "technology_data": "Vendor describes growth from a three-site pilot in 2019 to enterprise coverage, five source systems, weekly reviews and buyer feedback. The page remains undated.", "integration_human_role": "", "deployment_scope": "Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency.", "period": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "attribution": "Source-reported; see evidence limits", "source_ids": ["RS084"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "event": "Enterprise-scale implementation per vendor case; update reports benefits realized in CY2025, though its publication date is not shown.", "source_ids": ["RS084"]}], "limitations": ["Opportunity acceptance is not a savings percentage. Benefit amount, implementation costs, attribution method and publication date remain undisclosed."], "assessment": "AI-165 already held a concise profile, but omitted the operational mechanics, feedback/validation controls, scale path, data footprint, user count, and the distinction between an opportunity acceptance rate and realized dollar benefit.", "sources": [{"id": "RS084", "title": "Baker Hughes Achieves Bottom Line Savings with AI Sourcing Optimization", "url": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "publisher": "Baker Hughes", "date": "Unverified; page is live and states realized benefits in calendar 2025. The deployment began in 2019, which is outside this review window; current source discloses a CY2025 update but supplies no publication date.", "type": "Vendor-hosted named customer case study", "locator": "C3 AI customer page: Introduction, The Solution, and Project Highlights (web lines 147-214); page says initial 3-site pilot scaled to 10, then enterprise; CY2025 value in lines 162-164", "review_status": "Full C3 AI HTML case study opened and read; all cited detail visible in page text.", "supports": ["Procurement staff review sourcing recommendations using part similarity, landed cost, price indices and tariff data, accepting or rejecting opportunities.", "Vendor reports 200+ active users, 30+ million data records and a 5–7% acceptance rate for opportunities. “Millions” of benefit in CY2025 has no exact amount or currency."], "accessed": "2026-09-29"}], "source_ids": ["RS084"], "field_sources": {"workflow": ["RS084"], "technology_data": ["RS084"], "deployment_scope": ["RS084"], "practical_relevance": ["RS084"]}, "verification": "verified partner", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-125", "packet_id": "missing-sept29-ai-125", "record_ids": ["AI-125"], "existing_ids": ["AI-125"], "related_existing_ids": [], "company": "Winn Resources", "title": "AI-supported Texas Railroad Commission monthly W-10/G-10 filings (detail enrichment; regulatory workflow with accounting adjacency)", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.", "practical_relevance": "AI-125 already captures the headline time reduction but not the source’s missing-data exception trigger, customer-built follow-up protocol, full human approval gate, range of old cycle time, precise 50-well quote, or the prospective accounting use case. The new facts expand commercial/admin coverage without claiming AI use in live revenue accounting.", "business_problem": "", "workflow": "Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.", "technology_data": "Winn describes a protocol for investigating missing-data flags. The vendor release reports a roughly 50-well cycle and retains staff review; future revenue-statement extraction remains a plan.", "integration_human_role": "", "deployment_scope": "W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count.", "period": "2026-03-05", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-010"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2026-03-05", "event": "W-10/G-10 filing workflow in use by March 2026; broader accounting and FDC workflows planned.", "source_ids": ["M29-010"]}], "limitations": ["The vendor headline claims >95% time reduction, but 20 minutes versus 2–4 hours implies about 83–92% if the denominators match. No alternative baseline, repeated-cycle study, cost or cash saving is supplied. Planned revenue extraction is not deployed accounting evidence."], "assessment": "AI-125 already captures the headline time reduction but not the source’s missing-data exception trigger, customer-built follow-up protocol, full human approval gate, range of old cycle time, precise 50-well quote, or the prospective accounting use case. The new facts expand commercial/admin coverage without claiming AI use in live revenue accounting.", "sources": [{"id": "M29-010", "title": "Winn Resources Partners with Collide to Turn Operational Data Into a Competitive Advantage", "url": "https://www.prnewswire.com/news-releases/winn-resources-partners-with-collide-to-turn-operational-data-into-a-competitive-advantage-302705190.html", "publisher": "Winn Resources", "date": "2026-03-05", "type": "Vendor-issued press release with named customer executive quotation", "locator": "PRNewswire release sourced to Collide, 5 March 2026; 'Challenge', 'Solution', and 'Looking Ahead' sections. Specific case figures and approval process in paragraphs equivalent to 50-well cycle, 2-4 hour prior process, missing test flags, human review, and planned revenue extraction.", "review_status": "Original PRNewswire release opened from the syndication copy; source attribution says Collide. Full text accessible.", "supports": ["Prepare Texas W-10/G-10 filings from operating data, flag missing well tests and route drafts for human review before submission.", "One reported cycle: approximately 50 wells filed in under 30 minutes; an executive quotes 20 minutes. Typical prior process: 2–4 hours, depending on well count."], "accessed": "2026-09-29"}], "source_ids": ["M29-010"], "field_sources": {"workflow": ["M29-010"], "technology_data": ["M29-010"], "deployment_scope": ["M29-010"], "practical_relevance": ["M29-010"]}, "verification": "verified partner", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "missing-sept29-ai-143", "packet_id": "missing-sept29-ai-143", "record_ids": ["AI-143"], "existing_ids": ["AI-143"], "related_existing_ids": [], "company": "PETRONAS Carigali", "title": "TriCipta AI third JDA: surface equipment optimization", "department": "Exploration & geoscience", "secondary_departments": [], "category": "Internal adoption", "stage": "Development agreement / solution development; no deployment or results for this third JDA reported yet. PETRONAS separately says AI.SEEK and AI-enabled Global Exploration Basin Screening & Analysis from earlier JDAs are deployed.", "practical_relevance": "Material 2026 extension of the existing TriCipta initiative into an operational workflow; appropriate as an enrichment to the existing AI-143 profile, while clearly separating planned work from deployed earlier products.", "business_problem": "", "workflow": "Third joint development agreement extends the TriCipta partnership into AI for upstream surface equipment optimization, targeting production, maintenance decisions and asset value.", "technology_data": "PETRONAS Carigali, IBM Malaysia and Tridiagonal.AI combine enterprise AI and engineering-domain AI. The source says this JDA is newly signed; it does not state that its surface optimization solutions have been delivered or deployed.", "integration_human_role": "", "deployment_scope": "Development agreement / solution development; no deployment or results for this third JDA reported yet. PETRONAS separately says AI.SEEK and AI-enabled Global Exploration Basin Screening & Analysis from earlier JDAs are deployed.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No operating or economic metric for surface equipment optimization disclosed. Prior-JDA AI.SEEK and GEB deployment is reported without quantified result.", "period": "2026-07-08", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-011"]}], "economics": "Deployment cost and cash payback are not established by this source.", "timeline": [{"date": "2026-07-08", "event": "Development agreement / solution development; no deployment or results for this third JDA reported yet. PETRONAS separately says AI.SEEK and AI-enabled Global Exploration Basin Screening & Analysis from earlier JDAs are deployed.", "source_ids": ["M29-011"]}], "limitations": ["Announcement establishes scope and partnership only. Do not transfer prior exploration products’ deployment status to the new equipment workflow or read promised measurable impact as realized results. No model/control boundaries, costs or payback stated."], "assessment": "Material 2026 extension of the existing TriCipta initiative into an operational workflow; appropriate as an enrichment to the existing AI-143 profile, while clearly separating planned work from deployed earlier products.", "sources": [{"id": "M29-011", "title": "PETRONAS Advances TriCipta AI through the third Joint Development Agreement with IBM and Tridiagonal.AI", "url": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "publisher": "PETRONAS Carigali", "date": "2026-07-08", "type": "Operator media release", "locator": "Lines 129–143; new domain and scope lines 135–140; prior deployed AI.SEEK and GEB lines 142–143", "review_status": "Full issuer media release reviewed.", "supports": ["Third joint development agreement extends the TriCipta partnership into AI for upstream surface equipment optimization, targeting production, maintenance decisions and asset value.", "No operating or economic metric for surface equipment optimization disclosed. Prior-JDA AI.SEEK and GEB deployment is reported without quantified result."], "accessed": "2026-09-29"}], "source_ids": ["M29-011"], "field_sources": {"workflow": ["M29-011"], "technology_data": ["M29-011"], "deployment_scope": ["M29-011"], "practical_relevance": ["M29-011"]}, "verification": "verified primary", "cohort": "Gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"key": "swarm2-ai-213", "packet_id": "swarm2-ai-213", "record_ids": ["AI-213"], "existing_ids": [], "related_existing_ids": [], "company": "Koch Energy Services", "title": "LLM-assisted physical gas trade capture from chat into ETRM", "department": "Trading & marketing", "secondary_departments": ["Finance, investment & strategy"], "category": "Internal adoption", "stage": "Named Koch executive reports operating use; announcement body dated 18 September 2025, webpage metadata dated 26 November.", "practical_relevance": "Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.", "business_problem": "", "workflow": "An LLM converts abbreviated trader chat into structured physical-gas trade fields for booking in Koch’s ETRM system.", "technology_data": "Source chat stays linked to the trade for checking and downstream reconciliation. An undated vendor case study describes 12–13 extracted fields and trader validation before submission.", "integration_human_role": "Traders validate and submit generated order tickets, according to the undated case study.", "deployment_scope": "Named Koch executive reports operating use; announcement body dated 18 September 2025, webpage metadata dated 26 November.", "partners": ["ipushpull"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Koch reports 25 traders, more than 35,000 trades through the platform, and approximately 75% less trade-input time.", "period": "At the September 2025 announcement; measurement period unstated", "attribution": "Customer executive statement published by vendor; not independently validated.", "source_ids": ["SW2-001", "SW2-002"]}], "economics": "Process-time improvement; no labor-cost realization, implementation cost or payback disclosed.", "timeline": [{"date": "2025-09-18", "event": "Release body: 18 September 2025. Webpage metadata: 26 November 2025. Companion implementation page is undated.", "source_ids": ["SW2-001", "SW2-002"]}], "limitations": ["Platform trade count does not establish that every trade was AI-parsed. No baseline minutes or measurement method. Later undated 90,000-trade/85% figures are not treated as dated progress."], "assessment": "Closes a named-customer gap in gas-trade capture, with a clear system-of-record interface and a reported process metric.", "sources": [{"title": "ipushpull launches LLM-enabled agent to automate physical gas trade booking via chat", "url": "https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat", "date": "2025-09-18", "type": "Vendor announcement with named customer quote", "locator": "Opening dateline; Darnell Bortz, Director of Natural Gas Trading quote", "review_status": "Full announcement reviewed by researcher; independent audit used indexed text after direct retrieval failed.", "publisher": "ipushpull", "id": "SW2-001", "accessed": "2026-09-29"}, {"title": "Koch Energy Services and ipushpull automate physical gas trade booking through chat", "url": "https://ipushpull.com/case-studies/koch-energy-services-and-ipushpull-automate-physical-gas-trade-booking-through-chat", "date": "Undated", "type": "Vendor customer case study", "locator": "Reimagining trade capture; Human-in-the-loop", "review_status": "Indexed full source reviewed; used for workflow controls, not dated outcome update.", "publisher": "ipushpull", "id": "SW2-002", "accessed": "2026-09-29"}], "source_ids": ["SW2-001", "SW2-002"], "field_sources": {"workflow": ["SW2-001", "SW2-002"], "technology_data": ["SW2-001", "SW2-002"], "deployment_scope": ["SW2-001", "SW2-002"], "practical_relevance": ["SW2-001", "SW2-002"]}, "verification": "verified partner", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Customer executive statement published by vendor; not independently validated."}, {"key": "swarm2-ai-214", "packet_id": "swarm2-ai-214", "record_ids": ["AI-214"], "existing_ids": [], "related_existing_ids": [], "company": "Sentinel Peak Resources", "title": "Intelligent Oilfield steam-injection optimization with Quantum", "department": "Production & field operations", "secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "stage": "Reported operational use before the disclosure window; included for an in-window sponsor disclosure of earlier results.", "practical_relevance": "A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.", "business_problem": "", "workflow": "An AI-based thermal-management tool supports lower steam-injection rates in California heavy-oil operations.", "technology_data": "Sentinel Peak developed the Intelligent Oilfield tool with Quantum’s digital team and applied steam analysis to acquired Seneca assets.", "integration_human_role": "Review and actuation controls are not disclosed.", "deployment_scope": "Reported operational use before the disclosure window; included for an in-window sponsor disclosure of earlier results.", "partners": ["Quantum Energy Partners"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "For 2023, the case study reports more than 3 million barrels less steam injection and over 120,000 metric tons less Scope 1 GHG.", "period": "2023 operating results reported in 2024-titled report uploaded in May 2025", "attribution": "Portfolio-company data; report explicitly says Quantum and third parties have not verified it.", "source_ids": ["RS076"]}], "economics": "No quantified cash saving, deployment cost or payback disclosed.", "timeline": [{"date": "2025-05-02", "event": "2024-titled report uploaded May 2025; filename indicates 2 May. Reported operating results relate to 2023.", "source_ids": ["RS076"]}], "limitations": ["Results predate the 24-month disclosure window and include stewardship and analysis beyond AI. AI-only causality and baseline method are unproven. May 2025 upload is established by source path; exact May 2 date is file-indicated."], "assessment": "A named PE-backed operator and a specific co-developed operating workflow, with explicit sponsor linkage.", "sources": [{"title": "Quantum Capital Group 2024 Stakeholder Report — Sentinel Peak case study", "url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "date": "2025-05-02 (file-indicated)", "type": "Sponsor report", "locator": "Printed/PDF page 91, Sentinel Peak Resources", "review_status": "Original 96-page PDF downloaded and relevant page fully reviewed; date and attribution limits retained.", "publisher": "Quantum Capital Group", "id": "RS076", "accessed": "2026-09-29"}], "source_ids": ["RS076"], "field_sources": {"workflow": ["RS076"], "technology_data": ["RS076"], "deployment_scope": ["RS076"], "practical_relevance": ["RS076"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Portfolio-company data; report explicitly says Quantum and third parties have not verified it."}, {"key": "swarm2-ai-215", "packet_id": "swarm2-ai-215", "record_ids": ["AI-215"], "existing_ids": [], "related_existing_ids": ["AI-153"], "company": "Occidental", "title": "AI-assisted prioritization of lease-operator visits in routeless operations", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Management reports approximately 40% of onshore production had transitioned to routeless operations.", "practical_relevance": "A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.", "business_problem": "", "workflow": "Automation, field sensors and AI prioritize lease-operator routes; management reports routeless operations across part of onshore production.", "technology_data": "CEO described the integrated operating model and production coverage; model, data pipeline, dispatch logic and human override arrangements are not specified.", "integration_human_role": "Not disclosed.", "deployment_scope": "Management reports approximately 40% of onshore production had transitioned to routeless operations.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Approximately 40% of onshore production covered; this is a production-share denominator, not a share of wells or personnel.", "period": "Q2 2025 earnings call", "attribution": "CEO-reported operating scope; no independently measured outcome.", "source_ids": ["SW2-003"]}], "economics": "Adjacent $150 million expected 2025 onshore cost savings cover broader initiatives and cannot be assigned to this AI workflow.", "timeline": [{"date": "2025-08-07", "event": "7 August 2025 earnings call reports existing operating coverage.", "source_ids": ["SW2-003"]}], "limitations": ["No route-level time, cost, failure-rate or production-uplift result. The contribution of AI cannot be separated from sensors and automation."], "assessment": "A concrete field-service workflow with disclosed operating coverage; helps distinguish deployed work from a general AI strategy.", "sources": [{"title": "Occidental Q2 2025 Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy2q25transcript.pdf", "date": "2025-08-07", "type": "Issuer-hosted earnings-call transcript", "locator": "Page 3, Vicki Hollub prepared remarks", "review_status": "Full 17-page company-hosted transcript reviewed.", "publisher": "Occidental", "id": "SW2-003", "accessed": "2026-09-29"}], "source_ids": ["SW2-003"], "field_sources": {"workflow": ["SW2-003"], "technology_data": ["SW2-003"], "deployment_scope": ["SW2-003"], "practical_relevance": ["SW2-003"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "CEO-reported operating scope; no independently measured outcome."}, {"key": "swarm2-ai-153", "packet_id": "swarm2-ai-153", "record_ids": ["AI-153"], "existing_ids": ["AI-153"], "related_existing_ids": [], "company": "Occidental", "title": "AI Center of Excellence and 2025 project portfolio", "department": "IT, data & enterprise transformation", "secondary_departments": ["Reservoir engineering & field development"], "category": "Internal adoption", "stage": "At year-end 2025: 14 projects deployed, four piloting and two under development.", "practical_relevance": "Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.", "business_problem": "", "workflow": "The proxy reports 20 AI projects developed in 2025. Separately, it describes reservoir-characterization AI tools integrated with geoscience platforms; their inclusion in the count and link to the Center of Excellence are not specified.", "technology_data": "The Compensation Committee discusses AI alongside technology objectives. It reports improved reservoir predictions in blind testing without publishing the test design or accuracy gain.", "integration_human_role": "Not disclosed.", "deployment_scope": "At year-end 2025: 14 projects deployed, four piloting and two under development.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "20 projects; each assigned $10–90 million in estimated annual benefit. No realized portfolio total disclosed.", "period": "2025 year-end status", "attribution": "Issuer proxy disclosure; estimated economic benefits, not independently established returns.", "source_ids": ["SW2-004"]}], "economics": "Estimates are not net cash returns. Deployment costs, overlap between benefits and project-level realization remain undisclosed; do not sum the range into a portfolio valuation.", "timeline": [{"date": "2026-03-19", "event": "Proxy furnished 19 March 2026; underlying project status at 2025 year-end.", "source_ids": ["SW2-004"]}], "limitations": ["“Deployed” is not defined by asset count or rollout breadth. Incentive objectives refer to realized value, but the supporting results paragraph uses estimated benefits."], "assessment": "Provides a stage breakdown and an important estimate-versus-realization distinction for the previously broad company program.", "sources": [{"title": "2026 Oxy Proxy Statement", "url": "https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf", "date": "2026-03-19", "type": "Issuer proxy statement", "locator": "Compensation Discussion and Analysis, footnote (4), printed page 43 (PDF index 42)", "review_status": "Official 74-page PDF reviewed; project counts and estimated benefit wording independently checked.", "publisher": "Occidental", "id": "SW2-004", "accessed": "2026-09-29"}], "source_ids": ["SW2-004"], "field_sources": {"workflow": ["SW2-004"], "technology_data": ["SW2-004"], "deployment_scope": ["SW2-004"], "practical_relevance": ["SW2-004"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer proxy disclosure; estimated economic benefits, not independently established returns."}, {"key": "swarm2-ai-216", "packet_id": "swarm2-ai-216", "record_ids": ["AI-216"], "existing_ids": [], "related_existing_ids": [], "company": "ADNOC Gas", "title": "Real-time AI process optimization across gas plants", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Issuer says successfully deployed across multiple plants.", "practical_relevance": "A gas-processing workflow with a monetary result and a stated operating deployment, separate from ADNOC’s upstream ENERGYai program.", "business_problem": "", "workflow": "AI-driven optimizers allocate feed gas, reduce energy consumption and improve product recovery across multiple plants.", "technology_data": "The report confirms multi-plant deployment but gives no model architecture, plant list, control interface or operator-approval process.", "integration_human_role": "Not disclosed.", "deployment_scope": "Issuer says successfully deployed across multiple plants.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "More than US$20 million in value delivered to date; value composition and cumulative period unstated.", "period": "FY2025 report; underlying operating period not specified", "attribution": "Issuer-reported disclosure, not independently validated operating performance.", "source_ids": ["SW2-005"]}], "economics": "Reported economic value, not defined as net cash. Implementation cost and payback undisclosed.", "timeline": [{"date": "2026", "event": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "source_ids": ["SW2-005"]}], "limitations": ["No plant count, baseline, attribution method or value-component split. Future upside is excluded from reported delivered value."], "assessment": "A gas-processing workflow with a monetary result and a stated operating deployment, separate from ADNOC’s upstream ENERGYai program.", "sources": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report", "locator": "Printed pages 10–11, AI Highlights; PDF physical page 7", "review_status": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "publisher": "ADNOC Gas", "id": "SW2-005", "accessed": "2026-09-29"}], "source_ids": ["SW2-005"], "field_sources": {"workflow": ["SW2-005"], "technology_data": ["SW2-005"], "deployment_scope": ["SW2-005"], "practical_relevance": ["SW2-005"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported disclosure, not independently validated operating performance."}, {"key": "swarm2-ai-217", "packet_id": "swarm2-ai-217", "record_ids": ["AI-217"], "existing_ids": [], "related_existing_ids": ["AI-040"], "company": "ADNOC Gas", "title": "Neuron 5 asset-integrity and process-optimization platform", "department": "Maintenance & reliability", "secondary_departments": ["Refining, LNG & process operations"], "category": "Internal adoption", "stage": "Company describes an operating platform; deployment coverage and go-live dates undisclosed.", "practical_relevance": "Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.", "business_problem": "", "workflow": "A centralized AI platform combines predictive analytics and first-principles models to detect anomalies and support gas-plant reliability and process optimization.", "technology_data": "The report names Neuron 5 and the modeling approach; equipment coverage, interfaces and response controls are not described.", "integration_human_role": "Not disclosed.", "deployment_scope": "Company describes an operating platform; deployment coverage and go-live dates undisclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No Neuron 5-specific numerical result disclosed in the reviewed report section.", "period": "FY2025 report; underlying operating period not specified", "attribution": "Issuer-reported disclosure, not independently validated operating performance.", "source_ids": ["SW2-005"]}], "economics": "The adjacent RTO US$20 million claim is not assigned to Neuron 5.", "timeline": [{"date": "2026", "event": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "source_ids": ["SW2-005"]}], "limitations": ["No quantified reduction in shutdowns or project economics. Neuron 5 appeared in an earlier unreviewed ADNOC portfolio disclosure; this is a dedicated, operator-attributed profile."], "assessment": "Separates a named reliability platform from broad ADNOC portfolio references and unrelated subsurface agents.", "sources": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report", "locator": "Printed pages 10–11, AI Highlights; PDF physical page 7", "review_status": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "publisher": "ADNOC Gas", "id": "SW2-005", "accessed": "2026-09-29"}], "source_ids": ["SW2-005"], "field_sources": {"workflow": ["SW2-005"], "technology_data": ["SW2-005"], "deployment_scope": ["SW2-005"], "practical_relevance": ["SW2-005"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported disclosure, not independently validated operating performance."}, {"key": "swarm2-ai-218", "packet_id": "swarm2-ai-218", "record_ids": ["AI-218"], "existing_ids": [], "related_existing_ids": [], "company": "ADNOC Gas", "title": "Cantilever AI-enabled inspection management", "department": "Maintenance & reliability", "secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "stage": "Deployment stated; site count, equipment scope and start date undisclosed.", "practical_relevance": "Adds a named inspection workflow with explicit AI attribution and a company deployment statement.", "business_problem": "", "workflow": "An AI-enabled inspection platform supports non-intrusive inspections, forecasting and drone-enabled monitoring.", "technology_data": "ADNOC Gas names Cantilever as a deployed inspection-management platform. The report does not distinguish AI-model functions from drone hardware and other inspection methods.", "integration_human_role": "Not disclosed.", "deployment_scope": "Deployment stated; site count, equipment scope and start date undisclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Safety, accuracy and efficiency improvements described without numerical results.", "period": "FY2025 report; underlying operating period not specified", "attribution": "Issuer-reported disclosure, not independently validated operating performance.", "source_ids": ["SW2-005"]}], "economics": "No inspection cost saving, implementation cost or payback disclosed.", "timeline": [{"date": "2026", "event": "Reported in Integrated Report 2025, published in 2026; individual deployment dates undisclosed.", "source_ids": ["SW2-005"]}], "limitations": ["No detection accuracy, inspection-volume denominator, human sign-off process or measured safety effect."], "assessment": "Adds a named inspection workflow with explicit AI attribution and a company deployment statement.", "sources": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report", "locator": "Printed pages 10–11, AI Highlights; PDF physical page 7", "review_status": "Original 70-page issuer PDF downloaded, text read and relevant spread visually inspected.", "publisher": "ADNOC Gas", "id": "SW2-005", "accessed": "2026-09-29"}], "source_ids": ["SW2-005"], "field_sources": {"workflow": ["SW2-005"], "technology_data": ["SW2-005"], "deployment_scope": ["SW2-005"], "practical_relevance": ["SW2-005"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported disclosure, not independently validated operating performance."}, {"key": "swarm2-ai-219", "packet_id": "swarm2-ai-219", "record_ids": ["AI-219"], "existing_ids": [], "related_existing_ids": [], "company": "Petrobras", "title": "Machine learning for well configurations integrated with simulation", "department": "Reservoir engineering & field development", "secondary_departments": [], "category": "Internal adoption", "stage": "Presentation reports value already captured; deployment extent unspecified.", "practical_relevance": "Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.", "business_problem": "", "workflow": "Machine-learning well-configuration analysis is integrated with simulation systems.", "technology_data": "Slide identifies the decision workflow and its connection to simulation; model, well population and decision-approval controls are not disclosed.", "integration_human_role": "Not disclosed.", "deployment_scope": "Presentation reports value already captured; deployment extent unspecified.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "R$350 million captured; period and definition of captured value unstated.", "period": "November 2025 disclosure; result periods not specified except where annualized", "attribution": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed.", "source_ids": ["SW2-006"]}], "economics": "Reported captured value; no cash reconciliation, implementation cost or net return. Not an annual figure.", "timeline": [{"date": "2025-11", "event": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "source_ids": ["SW2-006"]}], "limitations": ["No baseline, project duration, well count or value-calculation method. Separate from Petrobras flare control and monitoring."], "assessment": "Adds a distinct subsurface decision workflow with a monetary disclosure whose economic basis remains open.", "sources": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation", "locator": "Slide 158, Accelerate value generation through digital technologies", "review_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "publisher": "Petrobras", "id": "SW2-006", "accessed": "2026-09-29"}], "source_ids": ["SW2-006"], "field_sources": {"workflow": ["SW2-006"], "technology_data": ["SW2-006"], "deployment_scope": ["SW2-006"], "practical_relevance": ["SW2-006"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed."}, {"key": "swarm2-ai-220", "packet_id": "swarm2-ai-220", "record_ids": ["AI-220"], "existing_ids": [], "related_existing_ids": [], "company": "Petrobras", "title": "AI predictive monitoring and integrated operations control rooms", "department": "Production & field operations", "secondary_departments": ["Maintenance & reliability"], "category": "Internal adoption", "stage": "Annualized avoided-loss result reported; asset coverage and persistence unspecified.", "practical_relevance": "Adds an operating-availability case while keeping physical production measures separate from financial returns.", "business_problem": "", "workflow": "AI enhances predictive monitoring and integrated operations control rooms (COI).", "technology_data": "Presentation connects AI-supported monitoring to avoided production losses; assets, event detection methods and intervention process are not described.", "integration_human_role": "Not disclosed.", "deployment_scope": "Annualized avoided-loss result reported; asset coverage and persistence unspecified.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "8.3 million BOE/year of losses avoided, described as approximately three days of production per year.", "period": "November 2025 disclosure; result periods not specified except where annualized", "attribution": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed.", "source_ids": ["SW2-006"]}], "economics": "Avoided production loss is an availability measure; it is not disclosed revenue, cash savings or incremental sales.", "timeline": [{"date": "2025-11", "event": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "source_ids": ["SW2-006"]}], "limitations": ["No counterfactual, event count, baseline or implementation cost. Do not monetize without entitlement, pricing and realization evidence."], "assessment": "Adds an operating-availability case while keeping physical production measures separate from financial returns.", "sources": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation", "locator": "Slide 158, Accelerate value generation through digital technologies", "review_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "publisher": "Petrobras", "id": "SW2-006", "accessed": "2026-09-29"}], "source_ids": ["SW2-006"], "field_sources": {"workflow": ["SW2-006"], "technology_data": ["SW2-006"], "deployment_scope": ["SW2-006"], "practical_relevance": ["SW2-006"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed."}, {"key": "swarm2-ai-221", "packet_id": "swarm2-ai-221", "record_ids": ["AI-221"], "existing_ids": [], "related_existing_ids": [], "company": "Petrobras", "title": "Machine-learning revenue forecasting for the domestic market", "department": "Trading & marketing", "secondary_departments": ["Finance, investment & strategy"], "category": "Internal adoption", "stage": "Improved model performance reported; operating scope undisclosed.", "practical_relevance": "Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.", "business_problem": "", "workflow": "Advanced machine-learning algorithms forecast domestic-market revenue.", "technology_data": "The presentation identifies the forecasting objective; features, model type, horizon and relationship to commercial decisions are unspecified.", "integration_human_role": "Not disclosed.", "deployment_scope": "Improved model performance reported; operating scope undisclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "51% greater forecast accuracy; accuracy measure and comparison baseline unspecified.", "period": "November 2025 disclosure; result periods not specified except where annualized", "attribution": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed.", "source_ids": ["SW2-006"]}], "economics": "Forecast accuracy is a model-performance measure, not measured revenue growth or savings.", "timeline": [{"date": "2025-11", "event": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "source_ids": ["SW2-006"]}], "limitations": ["The 51% is not an absolute accuracy rate or a percentage-point change. No test population, horizon, baseline or validation method."], "assessment": "Adds a commercial planning application with explicit ML attribution; useful for examining decisions beyond field operations.", "sources": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation", "locator": "Slide 158, Accelerate value generation through digital technologies", "review_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "publisher": "Petrobras", "id": "SW2-006", "accessed": "2026-09-29"}], "source_ids": ["SW2-006"], "field_sources": {"workflow": ["SW2-006"], "technology_data": ["SW2-006"], "deployment_scope": ["SW2-006"], "practical_relevance": ["SW2-006"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed."}, {"key": "swarm2-ai-146", "packet_id": "swarm2-ai-146", "record_ids": ["AI-146"], "existing_ids": ["AI-146"], "related_existing_ids": [], "company": "Petrobras", "title": "AI industrial-flare control and Smart Tocha monitoring", "department": "Refining, LNG & process operations", "secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "stage": "Annualized emissions result reported; Smart Tocha’s full rollout is not established by this slide.", "practical_relevance": "Adds quantified company evidence to the existing flare-control topic without claiming a confirmed product-level attribution.", "business_problem": "", "workflow": "Petrobras reports AI control of industrial flares in real time; earlier corporate material separately describes Smart Tocha automatic steam-flow adjustment.", "technology_data": "The plan slide adds an aggregate flare-control outcome but does not name Smart Tocha, sites or the rollout population. The earlier named product and this result are retained as related flare-workflow evidence.", "integration_human_role": "Not disclosed.", "deployment_scope": "Annualized emissions result reported; Smart Tocha’s full rollout is not established by this slide.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "10,000 tCO2/year reduction attributed to real-time AI industrial-flare control.", "period": "November 2025 disclosure; result periods not specified except where annualized", "attribution": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed.", "source_ids": ["SW2-006"]}], "economics": "Environmental result; no monetary saving or project payback disclosed.", "timeline": [{"date": "2025-11", "event": "November 2025 plan presentation reports operating results alongside forward strategy; benefits are not labeled FY2025-only.", "source_ids": ["SW2-006"]}], "limitations": ["Do not assign the full result specifically to Smart Tocha: the slide does not name it. No baseline, measurement period, plant count or emissions calculation method."], "assessment": "Adds quantified company evidence to the existing flare-control topic without claiming a confirmed product-level attribution.", "sources": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation", "locator": "Slide 158, Accelerate value generation through digital technologies", "review_status": "Original 163-page issuer PDF downloaded; slide 158 visually inspected to verify each metric-to-workflow mapping.", "publisher": "Petrobras", "id": "SW2-006", "accessed": "2026-09-29"}], "source_ids": ["SW2-006"], "field_sources": {"workflow": ["SW2-006"], "technology_data": ["SW2-006"], "deployment_scope": ["SW2-006"], "practical_relevance": ["SW2-006"]}, "verification": "verified primary", "cohort": "Second gap review · September 2026", "report_date": "2026-09-29", "evidence_note": "Issuer-reported presentation metric; calculation method and independent assurance undisclosed."}, {"key": "swarm3-ai-222", "packet_id": "swarm3-ai-222", "record_ids": ["AI-222"], "existing_ids": [], "related_existing_ids": [], "company": "Enbridge", "title": "Energy Optimizer for liquids-pipeline power decisions", "department": "Production & field operations", "secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "stage": "Company describes an operating tool; deployment footprint undisclosed.", "practical_relevance": "Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.", "business_problem": "", "workflow": "AI supplies operational insights to control-center staff choosing how to transport liquids with less power.", "technology_data": "The release identifies Azure Machine Learning and the Energy Optimizer tool; control-center operators make the decisions.", "integration_human_role": "Decision support for control-center operators; autonomous dispatch is not established.", "deployment_scope": "Company describes an operating tool; deployment footprint undisclosed.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No quantified energy, cost or emissions result.", "period": "October 2024 disclosure; operating measurement period unstated", "attribution": "Operator disclosure; performance not independently validated.", "source_ids": ["SW3-001"]}], "economics": "Potential cost and emissions benefits described; project cost and payback absent.", "timeline": [{"date": "2024-10-07", "event": "7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.", "source_ids": ["SW3-001"]}], "limitations": ["No asset count, baseline or measured outcome. Enterprise cloud-migration statistics do not measure this tool’s deployment."], "assessment": "Adds a specific midstream energy-use decision workflow distinct from pipeline-defect assessment.", "sources": [{"title": "Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization", "url": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "date": "2024-10-07", "type": "Operator press release", "locator": "AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "review_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "publisher": "Enbridge", "id": "SW3-001", "accessed": "2026-09-30"}], "source_ids": ["SW3-001"], "field_sources": {"workflow": ["SW3-001"], "technology_data": ["SW3-001"], "deployment_scope": ["SW3-001"], "practical_relevance": ["SW3-001"]}, "verification": "verified primary", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "Operator disclosure; performance not independently validated."}, {"key": "swarm3-ai-223", "packet_id": "swarm3-ai-223", "record_ids": ["AI-223"], "existing_ids": [], "related_existing_ids": [], "company": "Enbridge", "title": "AI-assisted aerial right-of-way threat monitoring", "department": "HSE & emissions", "secondary_departments": ["Maintenance & reliability"], "category": "Internal adoption", "stage": "Company says it uses the capability; monitored mileage and rollout dates undisclosed.", "practical_relevance": "Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.", "business_problem": "", "workflow": "AI helps review aerial pipeline-surveillance data for potentially damaging activity, hazardous conditions and loss of containment.", "technology_data": "The operator describes faster issue review and threat response; model, sensor mix and escalation controls are not specified.", "integration_human_role": "Alert triage and response responsibilities are not described.", "deployment_scope": "Company says it uses the capability; monitored mileage and rollout dates undisclosed.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Qualitative detection and response improvements only.", "period": "October 2024 disclosure; operating measurement period unstated", "attribution": "Operator disclosure; performance not independently validated.", "source_ids": ["SW3-001"]}], "economics": "No quantified damage avoidance, inspection saving or payback.", "timeline": [{"date": "2024-10-07", "event": "7 October 2024 company release describes tools in use; project-specific go-live dates undisclosed.", "source_ids": ["SW3-001"]}], "limitations": ["No accuracy, false-positive rate or response-time baseline. Separate from in-line inspection and deformation-feature assessment."], "assessment": "Adds an above-ground surveillance workflow that was absent from the existing pipeline-integrity record.", "sources": [{"title": "Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization", "url": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "date": "2024-10-07", "type": "Operator press release", "locator": "AI-Powered Initiatives: Energy Optimizer and Right of Way Monitoring", "review_status": "Original company release read in full; workflow-specific claims retained without assigning general platform statistics.", "publisher": "Enbridge", "id": "SW3-001", "accessed": "2026-09-30"}], "source_ids": ["SW3-001"], "field_sources": {"workflow": ["SW3-001"], "technology_data": ["SW3-001"], "deployment_scope": ["SW3-001"], "practical_relevance": ["SW3-001"]}, "verification": "verified primary", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "Operator disclosure; performance not independently validated."}, {"key": "swarm3-ai-224", "packet_id": "swarm3-ai-224", "record_ids": ["AI-224"], "existing_ids": [], "related_existing_ids": [], "company": "Williams", "title": "Companywide use of Copilot and approved AI tools", "department": "IT, data & enterprise transformation", "secondary_departments": ["Workforce & learning"], "category": "Internal adoption", "stage": "Company reports more than 90% of employees using approved AI tools.", "practical_relevance": "Separates internal workforce adoption from Williams’ AI-driven power investments and Sequent marketing test.", "business_problem": "", "workflow": "Employees use Copilot and other approved AI tools to reduce routine work and support decisions.", "technology_data": "Williams emphasizes approved-tool governance, employee experimentation and accountability for decisions; specific departmental processes are not identified.", "integration_human_role": "Employees retain accountability and judgment.", "deployment_scope": "Company reports more than 90% of employees using approved AI tools.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "More than 90% employee adoption; active-use threshold and observation period unspecified.", "period": "June 2026 disclosure; usage measurement window unspecified", "attribution": "Operator-reported adoption; not independently validated productivity.", "source_ids": ["SW3-002"]}], "economics": "No quantified cash savings, cost or payback. General days-to-hours language lacks a named process and baseline.", "timeline": [{"date": "2026-06-01", "event": "Article URL indicates 1 June 2026; no separate publication date is visible in the body. Launch and usage-measurement dates are unstated.", "source_ids": ["SW3-002"]}], "limitations": ["Adoption is not a productivity measure. No department-level usage, frequency, workflow result or paid-license denominator."], "assessment": "Separates internal workforce adoption from Williams’ AI-driven power investments and Sequent marketing test.", "sources": [{"title": "Williams tops 90% AI adoption as approved tools become part of everyday work", "url": "https://www.williams.com/2026/06/01/williams-tops-90-ai-adoption-as-approved-tools-become-part-of-everyday-work/", "date": "2026-06-01", "type": "Operator article", "locator": "Key takeaways and companywide adoption paragraphs", "review_status": "Original operator article read in full; date from article URL; adoption definition limitation retained.", "publisher": "Williams", "id": "SW3-002", "accessed": "2026-09-30"}], "source_ids": ["SW3-002"], "field_sources": {"workflow": ["SW3-002"], "technology_data": ["SW3-002"], "deployment_scope": ["SW3-002"], "practical_relevance": ["SW3-002"]}, "verification": "verified primary", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "Operator-reported adoption; not independently validated productivity."}, {"key": "swarm3-ai-225", "packet_id": "swarm3-ai-225", "record_ids": ["AI-225"], "existing_ids": [], "related_existing_ids": ["AI-152", "AI-106"], "company": "Oil and Natural Gas Corporation Limited (ONGC)", "title": "AI/ML data-physics water-injection optimization across Heera assets", "department": "Reservoir engineering & field development", "secondary_departments": ["Production & field operations"], "category": "Internal adoption", "stage": "Pilot began in June 2022; recommendations implemented in South Heera from March 2023 and subsequently in North and Mid Heera wells. ONGC reports continued field-performance monitoring in August 2025.", "practical_relevance": "Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.", "business_problem": "", "workflow": "Use AI/ML data-physics study recommendations to optimize water-injection rates: pilot in Heera, implement recommendations in South Heera and later in North and Mid Heera wells, adjusting rates for field performance and regularly monitoring production-injection performance.", "technology_data": "ONGC says Tachyus's AI/ML data-physics technique was adopted for water-injection optimization. Its FY2024-25 report describes recommendations implemented across the Heera assets, with ongoing production-injection monitoring. The source does not identify model architecture, data fields, or an automated control connection.", "integration_human_role": "The report describes changes for field performance and regular monitoring; it does not identify human approvers or automated execution authority.", "deployment_scope": "Pilot began in June 2022; recommendations implemented in South Heera from March 2023 and subsequently in North and Mid Heera wells. ONGC reports continued field-performance monitoring in August 2025.", "partners": ["Tachyus"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No quantified production gain, recovery uplift, injection reduction, accuracy, cost saving, or ROI disclosed.", "period": "August 7, 2025 report disclosure; pilot and implementation dates precede the research window.", "attribution": "Operator-reported implementation; no quantified AI outcome or independent validation disclosed.", "source_ids": ["SW3-003"]}], "economics": "No AI-specific economic result disclosed.", "timeline": [{"date": "2025-08-07", "event": "The report dates the Heera pilot to June 2022 and South Heera implementation to March 2023; later implementation in North and Mid Heera is undated. The first located in-window disclosure covering this implementation is ONGC's FY2024-25 annual report filed August 7, 2025.", "source_ids": ["SW3-003"]}], "limitations": ["Implementation is described, but no quantified result or validation is reported. The pilot and dated South Heera start precede the window; only the August 2025 report disclosure is in-window. The source does not state well count, input data, decision rights, or model design."], "assessment": "Distinct operator deployment in India, separate from ONGC's AI gas-lift tender and SLB's Tachyus/Aqueon product acquisition.", "sources": [{"title": "ONGC Integrated Annual Report 2024-25", "url": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "date": "2025-08-07", "type": "integrated annual report / statutory filing", "locator": "Annexures to the Board's Report, printed p. 131, item 30; PDF p. 157", "review_status": "Full primary PDF opened and reviewed; relevant disclosure read at PDF page 157.", "publisher": "Oil and Natural Gas Corporation Limited; filed with NSE", "id": "SW3-003", "accessed": "2026-09-30"}], "source_ids": ["SW3-003"], "field_sources": {"workflow": ["SW3-003"], "technology_data": ["SW3-003"], "deployment_scope": ["SW3-003"], "practical_relevance": ["SW3-003"]}, "verification": "verified primary", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "Operator-reported implementation; no quantified AI outcome or independent validation disclosed."}, {"key": "swarm3-ai-226", "packet_id": "swarm3-ai-226", "record_ids": ["AI-226"], "existing_ids": [], "related_existing_ids": [], "company": "ADNOC", "title": "Real-Time Operations Center AI drilling oversight across ADNOC's 120+ rig fleet", "department": "Drilling & completions", "secondary_departments": ["IT, data & enterprise transformation", "Production & field operations"], "category": "Internal adoption", "stage": "Operator reports deployed across more than 120 rigs as of August 4, 2026.", "practical_relevance": "Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.", "business_problem": "", "workflow": "Live rig data is consolidated into dashboards and AI insights for drilling teams to identify risks and coordinate decisions.", "technology_data": "SLB DrillOps enables the RTOC within ADNOC’s sovereign cloud. Engineers retain oversight; the release describes progression toward autonomy, not autonomous rig control.", "integration_human_role": "Engineers retain oversight and make operational decisions.", "deployment_scope": "Operator reports deployed across more than 120 rigs as of August 4, 2026.", "partners": ["SLB"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.", "period": "August 4, 2026 deployment announcement; outcome measurement period and baselines unspecified.", "attribution": "ADNOC operator press release, with SLB as technology partner; reported performance claims are not independently validated and their measurement basis is not given.", "source_ids": ["SW3-004"]}], "economics": "ADNOC describes reduced engineering effort and avoided rig downtime but reports no monetary savings, implementation cost, or payback.", "timeline": [{"date": "2026-08-04", "event": "ADNOC announced on August 4, 2026 that the AI-enabled Real-Time Operations Center platform had been deployed across more than 120 rigs. The release provides no initial go-live date or rollout milestones.", "source_ids": ["SW3-004"]}], "limitations": ["All performance figures are company-reported and lack measurement definitions, baselines, or independent validation. More than 120 rigs indicates platform deployment scope, not autonomous control. No model architecture, sensor list, or specific intervention examples are disclosed."], "assessment": "Distinct named ADNOC customer deployment and fleet-wide operating workflow. Existing ADNOC records cover subsurface ENERGYai, AiPSO, and AI-agent partnerships; the dataset's ONGC DrillOps record is a different customer and three-year contract, not ADNOC's deployed RTOC.", "sources": [{"title": "ADNOC and SLB Deploy AI Platform Across Over 120 Drilling Rigs to Strengthen Upstream Performance", "url": "https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/", "date": "2026-08-04", "type": "Operator press release", "locator": "Release body: paragraphs on RTOC deployment, workflow, operating scope and reported metrics", "review_status": "Full primary release opened and read; release body reviewed in full.", "publisher": "ADNOC", "id": "SW3-004", "accessed": "2026-09-30"}], "source_ids": ["SW3-004"], "field_sources": {"workflow": ["SW3-004"], "technology_data": ["SW3-004"], "deployment_scope": ["SW3-004"], "practical_relevance": ["SW3-004"]}, "verification": "verified primary", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "ADNOC operator press release, with SLB as technology partner; reported performance claims are not independently validated and their measurement basis is not given."}, {"key": "swarm3-ai-227", "packet_id": "swarm3-ai-227", "record_ids": ["AI-227"], "existing_ids": [], "related_existing_ids": ["AI-006", "AI-007", "AI-086"], "company": "Devon Energy", "title": "Harvey legal AI selection for land, transactional, litigation and compliance work", "department": "Land & legal", "secondary_departments": [], "category": "Internal adoption", "stage": "Harvey announced Devon’s selection/adoption on 14 August 2025. Go-live, production use, rollout completion and adoption breadth are not established.", "practical_relevance": "Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.", "business_problem": "", "workflow": "Devon announced adoption of Harvey's legal AI platform to support its in-house team's litigation, land, transactional and compliance functions. The announcement does not specify individual tasks or workflows within those functions.", "technology_data": "Harvey is named as the selected platform. No implementation architecture, integration, model configuration, access controls, rollout schedule or deployment scope is disclosed.", "integration_human_role": "Devon's Senior Counsel describes an aim to devote more time to higher-value work. Review steps, approval authority and controls around AI outputs are not described.", "deployment_scope": "Harvey announced Devon’s selection/adoption on 14 August 2025. Go-live, production use, rollout completion and adoption breadth are not established.", "partners": ["Harvey"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No workflow performance, usage or adoption metric reported.", "period": "14 August 2025 adoption announcement; no production measurement period disclosed", "attribution": "Vendor announcement with a named Devon Senior Counsel quote; not independently validated.", "source_ids": ["SW3-005"]}], "economics": "No realized savings, labor-cost reduction, implementation cost or payback disclosed. Efficiency and scale are stated objectives, not measured outcomes.", "timeline": [{"date": "2025-08-14", "event": "Harvey byline explicitly dates the announcement 14 August 2025. The announcement says Devon 'is adopting' the platform but gives no go-live date.", "source_ids": ["SW3-005"]}], "limitations": ["Vendor announcement with no task-level examples, rollout date, production evidence, controls or quantified results."], "assessment": "Adds a named, dated upstream legal-AI adoption announcement spanning land and transactional work, while preserving the distinction between announced selection and confirmed deployment.", "sources": [{"title": "Devon Energy Selects Harvey's Legal AI Platform", "url": "https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform", "date": "2025-08-14", "type": "Vendor announcement with named customer executive quote", "locator": "Byline and announcement text; Devon Senior Counsel Matt Tilly quote", "review_status": "Original full page opened and read. Exact byline date and stated legal-function scope verified; no production deployment or outcome claims inferred.", "publisher": "Harvey", "id": "SW3-005", "accessed": "2026-09-30"}], "source_ids": ["SW3-005"], "field_sources": {"workflow": ["SW3-005"], "technology_data": ["SW3-005"], "deployment_scope": ["SW3-005"], "practical_relevance": ["SW3-005"]}, "verification": "verified partner", "cohort": "Third gap review · September 2026", "report_date": "2026-09-30", "evidence_note": "Vendor announcement with a named Devon Senior Counsel quote; not independently validated."}, {"key": "swarm4-ai-228", "packet_id": "swarm4-ai-228", "record_ids": ["AI-228"], "existing_ids": [], "related_existing_ids": [], "company": "Moeve (formerly Cepsa)", "title": "Visbreaker stability and sulphur prediction at La Rábida", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Model development reported for named process units; production rollout extent unspecified.", "practical_relevance": "Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.", "business_problem": "", "workflow": "AI models predict outlet stability and sulphur content in the La Rábida Visbreaker unit.", "technology_data": "Process-property prediction; model design, input data and system integration are unspecified.", "integration_human_role": "Approval roles and control-system write authority are not disclosed.", "deployment_scope": "Model development reported for named process units; production rollout extent unspecified.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No accuracy, yield or financial result quantified.", "period": "FY2025; precise deployment and measurement dates undisclosed", "attribution": "Operator disclosure, not independent performance validation.", "source_ids": ["SW4-001"]}], "economics": "No project cost or payback disclosed.", "timeline": [{"date": "2026-02-26", "event": "FY2025 report adopted by the board 26 February 2026; online publication day unverified.", "source_ids": ["SW4-001"]}], "limitations": ["No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models."], "assessment": "Adds a process-specific AI application at an operator absent from the baseline, with an evidenced private-capital ownership link.", "sources": [{"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report", "locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "review_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "publisher": "Moeve", "id": "SW4-001", "accessed": "2026-10-01"}], "source_ids": ["SW4-001"], "field_sources": {"workflow": ["SW4-001"], "technology_data": ["SW4-001"], "deployment_scope": ["SW4-001"], "practical_relevance": ["SW4-001"]}, "verification": "verified primary", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator disclosure, not independent performance validation."}, {"key": "swarm4-ai-229", "packet_id": "swarm4-ai-229", "record_ids": ["AI-229"], "existing_ids": [], "related_existing_ids": [], "company": "Moeve (formerly Cepsa)", "title": "Real-time diesel and kerosene blending-property estimation at San Roque", "department": "Refining, LNG & process operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Model development reported for named process units; production rollout extent unspecified.", "practical_relevance": "Distinct from Visbreaker prediction: separate site, process and operational decision.", "business_problem": "", "workflow": "AI estimates diesel and kerosene blending properties in real time at San Roque.", "technology_data": "The report identifies the blending process but does not specify estimator inputs, property list or optimization constraints.", "integration_human_role": "Approval roles and control-system write authority are not disclosed.", "deployment_scope": "Model development reported for named process units; production rollout extent unspecified.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No accuracy, yield or financial result quantified.", "period": "FY2025; precise deployment and measurement dates undisclosed", "attribution": "Operator disclosure, not independent performance validation.", "source_ids": ["SW4-001"]}], "economics": "No project cost or payback disclosed.", "timeline": [{"date": "2026-02-26", "event": "FY2025 report adopted by the board 26 February 2026; online publication day unverified.", "source_ids": ["SW4-001"]}], "limitations": ["No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models."], "assessment": "Distinct from Visbreaker prediction: separate site, process and operational decision.", "sources": [{"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report", "locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "review_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "publisher": "Moeve", "id": "SW4-001", "accessed": "2026-10-01"}], "source_ids": ["SW4-001"], "field_sources": {"workflow": ["SW4-001"], "technology_data": ["SW4-001"], "deployment_scope": ["SW4-001"], "practical_relevance": ["SW4-001"]}, "verification": "verified primary", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator disclosure, not independent performance validation."}, {"key": "swarm4-ai-230", "packet_id": "swarm4-ai-230", "record_ids": ["AI-230"], "existing_ids": [], "related_existing_ids": [], "company": "Moeve (formerly Cepsa)", "title": "Mistral document-understanding assistant pilot for industrial centers", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Partnership and planned pilot announced; operating deployment not established.", "practical_relevance": "Separates generative document assistance from Moeve’s process-prediction models.", "business_problem": "", "workflow": "Planned industrial digital assistant to help staff understand documents and carry out business tasks.", "technology_data": "Mistral Applied AI and Moeve teams plan a tailored language-model assistant; document corpus, integrations and user count are not stated.", "integration_human_role": "Intended staff assistance; review and escalation controls not described.", "deployment_scope": "Partnership and planned pilot announced; operating deployment not established.", "partners": ["Mistral AI"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No measured adoption or productivity result.", "period": "December 2025 announcement; no operating results", "attribution": "Operator disclosure, not independent performance validation.", "source_ids": ["SW4-002", "SW4-001"]}], "economics": "No project cost or payback disclosed.", "timeline": [{"date": "2025-12-04", "event": "4 December 2025 agreement announces a first pilot to be developed; go-live unreported.", "source_ids": ["SW4-002", "SW4-001"]}], "limitations": ["Announcement is prospective. Neither rollout across all sites nor energy/cost savings is evidenced."], "assessment": "Separates generative document assistance from Moeve’s process-prediction models.", "sources": [{"title": "Moeve and Mistral AI join forces to spearhead generative AI solutions in the energy sector", "url": "https://www.moeveglobal.com/en/press/moeve-mistral-ai-alliance-ai-digital-assistant-industrial-efficiency", "date": "2025-12-04", "type": "Operator press release", "locator": "Pilot digital assistant paragraphs; press index spells out 4 December 2025", "review_status": "Full operator release and dated issuer press index read; planned pilot kept separate from deployment.", "publisher": "Moeve", "id": "SW4-002", "accessed": "2026-10-01"}, {"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report", "locator": "p. 13 Digitalisation and advanced analytics; p. 34 ownership; signed board certification PDF p. 174", "review_status": "Full issuer PDF downloaded and relevant pages read; signed date visually checked. Date is board adoption, not independently established web publication.", "publisher": "Moeve", "id": "SW4-001", "accessed": "2026-10-01"}], "source_ids": ["SW4-002", "SW4-001"], "field_sources": {"workflow": ["SW4-002", "SW4-001"], "technology_data": ["SW4-002", "SW4-001"], "deployment_scope": ["SW4-002", "SW4-001"], "practical_relevance": ["SW4-002", "SW4-001"]}, "verification": "verified primary", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator disclosure, not independent performance validation."}, {"key": "swarm4-ai-231", "packet_id": "swarm4-ai-231", "record_ids": ["AI-231"], "existing_ids": [], "related_existing_ids": [], "company": "Presidio Production Company (NYSE: FTW)", "title": "FTW Technologies live AI well surveillance and production intelligence at Presidio", "department": "Production & field operations", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "Company describes well surveillance and production intelligence as live in June 2026, with an AI surveillance tool deployed to operators and engineers. Full-fleet agentic orchestration, integrated OPEX/CAPEX agents, and a 3–5% production uplift are presented as targets; company reports approximately 1% YTD toward that target, with measurement basis unspecified.", "practical_relevance": "New named operator record in the reviewed baseline. Adds a concrete live surveillance / production-intelligence workflow at a sponsor-origin operator, while keeping its scale-up roadmap and value claims separate from deployment evidence.", "business_problem": "", "workflow": "Presidio says a 24/7 AI surveillance tool is live in well surveillance and production intelligence and available to all operators and engineers. Users can query live production data in natural language. The underlying well-intelligence layer combines field pumper voice/text observations with well files, drilling and completion reports, production volumes, and geology. Operations staff remain central; an asset-intelligence group uses analytics to drive field efficiencies. The deck also describes machine-learning pump-by-exception routing: high-cash-flow wells receive daily visits and other wells are visited on exceptions; field crews remain responsible for action.", "technology_data": "FTW Technologies, a Presidio business unit, provides a proprietary well-intelligence layer and surveillance/production-intelligence tools. The source does not specify model types, field-data integration architecture, alert workflows, or the number of wells whose data are actively covered by the live surveillance product. It describes continuous automatic ESP, rod-pump, and gas-lift adjustment, but in the context of the 2026 production-growth target; those controls should not be stated as a verified live deployment across all wells.", "integration_human_role": "Company says operating personnel stay at the center and that the live surveillance tool is deployed to all operators and engineers. It does not describe individual review or approval controls for each recommendation. The slide's future-state automation must not be mistaken for established autonomous operation at full fleet scale.", "deployment_scope": "Company describes well surveillance and production intelligence as live in June 2026, with an AI surveillance tool deployed to operators and engineers. Full-fleet agentic orchestration, integrated OPEX/CAPEX agents, and a 3–5% production uplift are presented as targets; company reports approximately 1% YTD toward that target, with measurement basis unspecified.", "partners": ["FTW Technologies (Presidio business unit)"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Company reports approximately 1% YTD production uplift toward a 3–5% 2026 target. No measurement cutoff, baseline, attribution method or independent validation is supplied. Separately, its pump-by-exception operating case reports 50% fewer well visits; scope is selected acquisition case studies, not necessarily the entire fleet.", "period": "June 2026 issuer presentation. AI described as live in well surveillance and production intelligence; the all-well orchestration and 3–5% uplift are stated as 2026 targets, with approximately 1% YTD reported.", "attribution": "Operator / issuer investor presentation filed with the SEC; company-reported, not independently audited. The presentation contains forward-looking-statement caveats.", "source_ids": ["SW4-003"]}], "economics": "The deck illustrates $2.5 million annual value for each 1% uplift at $30/BOE across the stated fleet; this is a scenario/sensitivity, not disclosed realized cash savings. No implementation cost, actual AI-attributed revenue, or payback is reported.", "timeline": [{"date": "2026-06", "event": "The deck is dated June 2026 and says well surveillance and production intelligence are live today; a 24/7 AI surveillance tool is deployed to operators and engineers. Its broader agentic orchestration across the approximately 2,300-well fleet is labeled a 2026 target, not a completed rollout. The deck says a 3–5% production-uplift target has achieved approximately 1% YTD, without defining the measurement cutoff.", "source_ids": ["SW4-003"]}], "limitations": ["Live surveillance is disclosed; fleet-wide orchestration remains a target. Use the explicit target/progress framing on slide 15. The column placement of the flattened slide 16 bullet was not independently verified. Visit reductions have a different scope from production uplift and must not be added to it. The $2.5 million sensitivity is illustrative. Presidio listed in March 2026; historical sponsor backing does not make it currently private."], "assessment": "New named operator record in the reviewed baseline. Adds a concrete live surveillance / production-intelligence workflow at a sponsor-origin operator, while keeping its scale-up roadmap and value claims separate from deployment evidence.", "sources": [{"title": "Presidio Production Company Investor Presentation, dated June 2026", "url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026069875/ea029480701_ex99-1.htm", "date": "2026-06", "type": "issuer investor presentation filed with SEC", "locator": "Slides 15–18 (SEC document lines 122–146); AI deployment, input data and live-versus-target distinctions; slides 13 and 26 for selected acquisition case studies and pump-by-exception routing", "review_status": "Full SEC-hosted HTML presentation opened and reviewed, including legal disclaimer and AI slides.", "publisher": "Presidio Production Company", "id": "SW4-003", "accessed": "2026-10-01"}], "source_ids": ["SW4-003"], "field_sources": {"workflow": ["SW4-003"], "technology_data": ["SW4-003"], "deployment_scope": ["SW4-003"], "practical_relevance": ["SW4-003"]}, "verification": "verified primary", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator / issuer investor presentation filed with the SEC; company-reported, not independently audited. The presentation contains forward-looking-statement caveats."}, {"key": "swarm4-ai-232", "packet_id": "swarm4-ai-232", "record_ids": ["AI-232"], "existing_ids": [], "related_existing_ids": [], "company": "SLB", "title": "AI Builder flight-invoice extraction for airline GST processing", "department": "Finance, investment & strategy", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "SLB describes the bot as implemented and processing invoices; the source gives no go-live date or rollout scope beyond the stated volume.", "practical_relevance": "A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.", "business_problem": "", "workflow": "SLB's Flight Airline Automation Bot retrieves flight invoices from email, uses AI Builder to extract 22 key fields and tax details, validates the data, and compiles an Airline GST report for the tax team.", "technology_data": "Power Automate retrieves and filters vendor emails, AI Builder extracts invoice fields, business validations check extracted details, and the result is compiled for tax processing. The page identifies Dataverse and Azure Key Vault in the back end. It does not describe a posting to the general ledger or payment execution.", "integration_human_role": "The tax team receives the compiled GST report and uses it for tax processing. Validation is described, but the page does not specify whether checks are automated, who resolves exceptions, or an approval gate before posting.", "deployment_scope": "SLB describes the bot as implemented and processing invoices; the source gives no go-live date or rollout scope beyond the stated volume.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "SLB reports processing more than 1,000 emails and 200 vendor invoices per month and effort equivalent to one full-time employee saved. The case separately says 22 invoice fields are extracted.", "period": "Microsoft Learn page last updated August 8, 2025; underlying bot go-live and metric measurement window not disclosed", "attribution": "Microsoft customer case study quoting named SLB Lead Architect Rushi Patel; results are customer-reported through a technology-provider case study, not independently validated.", "source_ids": ["SW4-004"]}], "economics": "One-FTE equivalent is reported workload capacity, not a disclosed cash payroll saving. No labor cost realization, software/deployment cost, or payback is reported.", "timeline": [{"date": "2025-08-08", "event": "Microsoft Learn displays 'Last updated on 2025-08-08'. The page confirms the workflow is implemented but does not date the implementation.", "source_ids": ["SW4-004"]}], "limitations": ["The August 8, 2025 date is the Microsoft Learn page's explicit last-updated date, not the deployment date. Vendor-published case; customer executive quote and detailed workflow are disclosed, but the source does not give accuracy, exception rate, baseline minutes, governance details for ambiguous records, or audited savings. Invoice scope is airline travel invoices and tax processing, not supplier AP/JIB/revenue accounting."], "assessment": "A specific, named oilfield-services finance workflow using document AI in production, with reported monthly throughput and a carefully bounded process-capacity metric.", "sources": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study with named SLB architect", "locator": "Flight Airline Automation Bot and tax process: page sections 148–174; page last-updated stamp at line 287; date is page last update, not go-live", "review_status": "Full official Microsoft Learn page opened and reviewed. Byline/customer quote, 22-field extraction, validation, throughput and one-FTE-equivalent wording checked against the full source.", "publisher": "Microsoft", "id": "SW4-004", "accessed": "2026-10-01"}], "source_ids": ["SW4-004"], "field_sources": {"workflow": ["SW4-004"], "technology_data": ["SW4-004"], "deployment_scope": ["SW4-004"], "practical_relevance": ["SW4-004"]}, "verification": "verified partner", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Microsoft customer case study quoting named SLB Lead Architect Rushi Patel; results are customer-reported through a technology-provider case study, not independently validated."}, {"key": "swarm4-ai-233", "packet_id": "swarm4-ai-233", "record_ids": ["AI-233"], "existing_ids": [], "related_existing_ids": [], "company": "SLB", "title": "AI-powered project discovery and recommendations across SLB facilities", "department": "IT, data & enterprise transformation", "secondary_departments": ["Finance, investment & strategy"], "category": "Internal adoption", "stage": "Implemented and in continued use; Microsoft page says SLB continues to benefit from the application. Launch date and global rollout coverage are not stated.", "practical_relevance": "Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.", "business_problem": "", "workflow": "SLB employees use an AI-powered project recommendation tool to find similar initiatives across global facilities, identify opportunities to consolidate work, and discover projects to replicate or join. Financial controllers validate project records; management reviews the resulting projects.", "technology_data": "Power Apps supports project submission/review; Dataverse records synchronize to Azure Blob Storage; a Power Automate flow calls Azure Functions/Python to retrieve similar projects using text similarity. The system excludes the current user’s own projects from recommendations. Financial controllers validate records and management reviews projects.", "integration_human_role": "Employees review and act on recommendations; controllers validate project records and management reviews projects. Source does not detail model governance or human adjudication of similarity matches.", "deployment_scope": "Implemented and in continued use; Microsoft page says SLB continues to benefit from the application. Launch date and global rollout coverage are not stated.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Within first 28 days of launch, SLB reports 800+ AI-powered matches and 150 similar projects inspiring facility initiatives. These are activity/claimed follow-on counts, not quantified savings or verified project consolidation.", "period": "Microsoft Learn page last updated 2025-08-08; workflow chronology/launch dates not disclosed", "attribution": "Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.", "source_ids": ["SW4-004"]}], "economics": "No financial savings, implementation costs, or payback disclosed. Benefits are described as collaboration and innovation.", "timeline": [{"date": "2025-08-08", "event": "Page states the initiative is part of SLB’s 2024 internal strategy. It reports outcomes within the first 28 days after launch but does not state the launch calendar date. The page itself was last updated 2025-08-08.", "source_ids": ["SW4-004"]}], "limitations": ["Source is a Microsoft-hosted customer story; no publication date beyond last-updated 2025-08-08, no deployment launch date, precision/recall, project-value attribution, costs, or independent verification. This is an internal project-discovery workflow, not a field production deployment."], "assessment": "Distinct corporate workflow absent from baseline: AI similarity recommendations across project records with a short-window usage metric.", "sources": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study", "locator": "AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live", "review_status": "Full official Microsoft Learn page opened and reviewed.", "publisher": "Microsoft", "id": "SW4-004", "accessed": "2026-10-01"}], "source_ids": ["SW4-004"], "field_sources": {"workflow": ["SW4-004"], "technology_data": ["SW4-004"], "deployment_scope": ["SW4-004"], "practical_relevance": ["SW4-004"]}, "verification": "verified partner", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated."}, {"key": "swarm4-ai-234", "packet_id": "swarm4-ai-234", "record_ids": ["AI-234"], "existing_ids": [], "related_existing_ids": [], "company": "SLB", "title": "AI extraction of shipping documents for SLB export-duty recovery", "department": "Finance, investment & strategy", "secondary_departments": ["Supply chain & logistics", "IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "Automated document workflow is described as implemented and processing received documents. Exact go-live date, total volume, and country coverage are not disclosed.", "practical_relevance": "Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.", "business_problem": "", "workflow": "SLB processes partner shipping documents for tools and parts across multiple countries to associate shipment details with SAP orders and entities and support export-duty recovery.", "technology_data": "Users upload shipment documents or invoices to a SharePoint library. A weekly Power Automate flow invokes AI Builder template models to extract shipment details, updates a SharePoint list and logs, and displays extracted data/documents in Power BI. Azure App Insights monitors completion. Failed/unrecognized documents go to the process-owner team for manual processing or assignment to a new template; once enough examples are accumulated, templates are trained for future recognition. Extracted shipment data are linked to ERP orders for tax claims; source names SAP orders/entities.", "integration_human_role": "Process owners receive failure reports and manually process failures or assign a new document template; the team retrains template models when sufficient examples are accumulated.", "deployment_scope": "Automated document workflow is described as implemented and processing received documents. Exact go-live date, total volume, and country coverage are not disclosed.", "partners": ["Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Microsoft page reports $10 million recovered in export duties; workflow description says it processes received documents. The source does not disclose the measurement period, baseline, or direct attribution method.", "period": "Microsoft Learn page last updated 2025-08-08; workflow chronology/launch dates not disclosed", "attribution": "Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated.", "source_ids": ["SW4-004"]}], "economics": "Reported recovery is $10 million of export duties, not necessarily recurring annual savings. No implementation/software cost or payback disclosed.", "timeline": [{"date": "2025-08-08", "event": "Page last updated 2025-08-08. Deployment/go-live and $10 million measurement dates are not disclosed.", "source_ids": ["SW4-004"]}], "limitations": ["Microsoft-hosted customer story, not independently validated. The $10 million figure has no stated period, reconciliation, or counterfactual. The exact publication date is not explicit beyond page last-updated 2025-08-08; no launch date or throughput/accuracy is disclosed. This is a distinct shipping-document/export-duty workflow, not the airline invoice GST bot also described on the same page."], "assessment": "Distinct named SLB finance/trade-compliance back-office deployment absent from baseline; links document extraction to ERP orders for export-duty recovery.", "sources": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study", "locator": "AI-powered project discovery sections (lines 53–96); automated document processing (lines 97–147); page date line 287; date is page last update, not go-live", "review_status": "Full official Microsoft Learn page opened and reviewed.", "publisher": "Microsoft", "id": "SW4-004", "accessed": "2026-10-01"}], "source_ids": ["SW4-004"], "field_sources": {"workflow": ["SW4-004"], "technology_data": ["SW4-004"], "deployment_scope": ["SW4-004"], "practical_relevance": ["SW4-004"]}, "verification": "verified partner", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Microsoft customer case study quoting named SLB employees; operational impact is provider-published/customer-reported and not independently validated."}, {"key": "swarm4-ai-235", "packet_id": "swarm4-ai-235", "record_ids": ["AI-235"], "existing_ids": [], "related_existing_ids": ["AI-022"], "company": "ExxonMobil / XTO Energy", "title": "Closed-loop machine-learning gas-lift optimization across 1,300+ Permian wells", "department": "Production & field operations", "secondary_departments": ["Reservoir engineering & field development"], "category": "Internal adoption", "stage": "Reported deployed across >1,300 operator wells by the October 2025 JPT highlight. The specific test-stage dates, eligibility criteria, and fraction of active wells are not disclosed.", "practical_relevance": "This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.", "business_problem": "", "workflow": "The well-by-well workflow remotely changes gas-lift injection-rate setpoints, performs multirate tests, automatically acquires well data, and iteratively uses the test response to select a production-optimizing setting. The abstract says a machine-learning model supplies recommendations on wells without downhole pressure gauges, while production data are quality-checked and uplift is model-tracked using productivity and steady-state tests.", "technology_data": "SPE/JPT describes an operating closed loop deployed on more than 1,300 Permian wells. JPT says this required no surface or downhole equipment changes and no additional personnel work for the repeated monthly optimization cycle. No model vendor or project-level software/integration cost is identified. The SPE abstract notes minimal disturbance to surface facilities.", "integration_human_role": "The first three-month stage retained field-person approval or rejection of recommendations. In the following three-month stage, most recommendations were automatically approved and implemented. The sources do not specify who retains override authority, safety interlocks, exception handling, or the fraction still requiring manual review.", "deployment_scope": "Reported deployed across >1,300 operator wells by the October 2025 JPT highlight. The specific test-stage dates, eligibility criteria, and fraction of active wells are not disclosed.", "partners": ["XTO Energy", "ExxonMobil Global Services Company"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Operator-authored SPE abstract reports about 2% average oil uplift. JPT’s October 2024 interview reports 1.87% in a three-month human-approval stage and about 2.2% in the next three-month, mostly automatic stage. These summaries are not additive; periods and cohorts are incompletely disclosed.", "period": "The SPE case was presented in August 2024 and JPT reported two three-month stages on 2024-10-01; exact operating dates and the production-result measurement window are not specified. Treat the stage results as historical by the 2024 conference disclosure. JPT reaffirmed deployment to >1,300 Permian wells on 2025-10-01; that later publication does not establish a new measurement period.", "attribution": "ExxonMobil/XTO coauthored SPE 219553; the operator's production optimization engineer described the staged test to JPT. All performance claims are operator-reported and not independently validated.", "source_ids": ["SW4-005", "SW4-006", "SW3-006"]}], "economics": "No license, implementation, integration, operating cost, cash benefit, or payback is disclosed. JPT describes incremental production without major expenditure and no equipment changes or added personnel work, but gives no amount; this qualitative statement is not a project cost or ROI estimate.", "timeline": [{"date": "2025-10-01", "event": "SPE 219553 was presented in August 2024. JPT reported the staged trial in October 2024 and published a deployment highlight in October 2025. Historical results are not newly measured 2025 outcomes.", "source_ids": ["SW4-005", "SW4-006", "SW3-006"]}], "limitations": ["The primary paper's accessible abstract states the full paper is unavailable, so full technical validation is not possible. The operator-reported uplift lacks a disclosed counterfactual or confidence interval. Gas-lift performance can drift when water cut or productivity changes; JPT says an existing setting may remain suboptimal for years without frequent monitoring. The control-boundary details and automatic-versus-human exception rates are undisclosed."], "assessment": "This is a distinct ExxonMobil operational workflow, not merely a company-level AI mention. Exact dedup review found AI-022 (ExxonMobil Permian/operating-site AI) to be broad executive commentary without gas-lift workflow, named SPE paper, staged human approvals, setpoint control, 1,300-well deployment, or these results. Related existing record: AI-022. The common company and upstream setting do not establish that AI-022 and SPE 219553 refer to the same initiative; retain as a new, specific case unless a later source crosswalk proves identity.", "sources": [{"title": "Gas Lift Optimization in the Permian Uses Machine Learning, Artificial Intelligence", "url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "date": "2025-10-01", "type": "SPE Journal of Petroleum Technology paper highlight", "locator": "Accessible abstract and first paragraph; lines 592–620", "review_status": "Accessible text read in full. The JPT page is marked member-exclusive after the opening summary; this supports 1,300+ deployment and workflow framing, not details beyond the accessible text.", "publisher": "Society of Petroleum Engineers", "id": "SW4-005", "accessed": "2026-10-01"}, {"title": "Gas Lift Optimization in the Permian Using Machine Learning and Artificial Intelligence (SPE 219553-MS)", "url": "https://www.researchgate.net/publication/383230634_Gas_Lift_Optimization_in_the_Permian_Using_Machine_Learning_and_Artificial_Intelligence", "date": "2024-08", "type": "Operator-authored SPE conference paper abstract", "locator": "Complete abstract, paragraphs 1–4", "review_status": "Full accessible abstract read. Page explicitly says the full text is unavailable and can be requested from the authors; do not characterize the complete paper as reviewed.", "publisher": "ResearchGate-hosted abstract of an SPE paper", "id": "SW4-006", "accessed": "2026-10-01"}, {"title": "Unsung Hero: Artificial Lift Boosts Production Without Breaking the Bank", "url": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted", "date": "2024-10-01", "type": "SPE Journal of Petroleum Technology feature with operator-engineer interview", "locator": "Article date at lines 592–599; ExxonMobil SPE 219553 workflow and claims at lines 625–646", "review_status": "Full relevant article section opened and read, including the two 3-month control stages, operator quotations, workflow description, and caveats. Reporting summarizes a paper and interview; it is not independent validation.", "publisher": "Society of Petroleum Engineers", "id": "SW3-006", "accessed": "2026-10-01"}], "source_ids": ["SW4-005", "SW4-006", "SW3-006"], "field_sources": {"workflow": ["SW4-005", "SW4-006", "SW3-006"], "technology_data": ["SW4-005", "SW4-006", "SW3-006"], "deployment_scope": ["SW4-005", "SW4-006", "SW3-006"], "practical_relevance": ["SW4-005", "SW4-006", "SW3-006"]}, "verification": "verified primary abstract", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "ExxonMobil/XTO coauthored SPE 219553; the operator's production optimization engineer described the staged test to JPT. All performance claims are operator-reported and not independently validated."}, {"key": "swarm4-ai-236", "packet_id": "swarm4-ai-236", "record_ids": ["AI-236"], "existing_ids": [], "related_existing_ids": [], "company": "Dagang Oilfield (PetroChina/CNPC)", "title": "Machine-learning production forecasting across Dagang Oilfield blocks", "department": "Reservoir engineering & field development", "secondary_departments": ["Production & field operations"], "category": "Internal adoption", "stage": "Trialed in multiple blocks in 2025; production-wide deployment is not established.", "practical_relevance": "Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.", "business_problem": "", "workflow": "A machine-learning production-forecast model was trialed in multiple Dagang Oilfield blocks to predict oilfield production.", "technology_data": "Article identifies machine learning and production forecasting; input data, forecast horizon, model design, integration, and how forecasts informed operating decisions are undisclosed.", "integration_human_role": "User, review, and approval roles are not disclosed; no automated control or production-setting authority is stated.", "deployment_scope": "Trialed in multiple blocks in 2025; production-wide deployment is not established.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "CNPC article reports that forecast accuracy “improved by 15%” (预测准确率提升15%). The comparison baseline, calculation, validation set, forecast horizon, and whether this is a relative improvement or percentage-point change are unspecified; it is not an absolute 15% accuracy score.", "period": "Article published 4 February 2026; the article states the model was trialed in multiple blocks during 2025.", "attribution": "CNPC news-center report of Dagang Oilfield activity; operator-group reported, not independently validated.", "source_ids": ["SW4-007"]}], "economics": "No production uplift, avoided loss, cost, or payback is attributed to this forecast model.", "timeline": [{"date": "2026-02-04", "event": "The article’s 2025 retrospective says the machine-learning production-forecast model was trialed in multiple blocks. It does not provide individual trial dates or say the model was rolled out beyond the trials.", "source_ids": ["SW4-007"]}], "limitations": ["The article says Dagang created 11 AI applications in 2025; this is a portfolio count, not 11 separately evidenced profiles. The 15% accuracy statement is an operator-group-reported model-performance claim without a disclosed denominator or validation method. Adjacent figures in the article—response speed up more than 60% and plan-preparation time down an average 30%—describe broader digital transformation/platform work and are not attributed to this ML forecast model."], "assessment": "Adds a named upstream ML workflow at Dagang Oilfield, with a dated multi-block trial and a scoped but methodologically underspecified model-performance claim.", "sources": [{"title": "Machine-learning production forecasting across Dagang Oilfield blocks", "url": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml", "date": "2026-02-04", "type": "CNPC News Center / China Petroleum News Center operator-group news article", "locator": "Article section “让科技生产力‘看得见’” (“Make technological productivity visible”), paragraph beginning “人工智能的应用体现出务实的导向” (“AI applications reflect a practical orientation”); published 2026-02-04 10:17.", "review_status": "Original CNPC public HTML fetched and decoded using GB18030; full article reviewed. Operator-group reporting, not independently audited performance.", "publisher": "CNPC News Center / China Petroleum News Center", "id": "SW4-007", "accessed": "2026-10-01"}], "source_ids": ["SW4-007"], "field_sources": {"workflow": ["SW4-007"], "technology_data": ["SW4-007"], "deployment_scope": ["SW4-007"], "practical_relevance": ["SW4-007"]}, "verification": "verified primary", "cohort": "Fourth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "CNPC news-center report of Dagang Oilfield activity; operator-group reported, not independently validated."}, {"key": "swarm5-ai-237", "packet_id": "swarm5-ai-237", "record_ids": ["AI-237"], "existing_ids": [], "related_existing_ids": [], "company": "Repsol", "title": "Repsol legal team Harvey rollout and adoption update", "department": "Land & legal", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "A dated 2026 vendor interview confirms that Repsol’s legal team evaluated, piloted and rolled out Harvey across a stated 150-person team. The sources do not establish exact go-live date, percentage adoption in March 2026, or task-level production volume.", "practical_relevance": "Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.", "business_problem": "", "workflow": "Repsol’s in-house legal team uses Harvey to assist legal work; the original Repsol announcement listed legal-document analysis, contract/regulation comparison, clause suggestions, translation, legal queries and document summaries. The March 2026 source establishes rollout scope and adoption-support practices, but does not add new task-level workflows.", "technology_data": "Harvey was evaluated against Microsoft Copilot and Repsol’s internal prototype before a pilot/rollout. Repsol’s dated interview describes a 150-person team, a Harvey Taskforce, champions and team-level use-case sharing/training. Technical integrations, model configuration and access controls are not disclosed.", "integration_human_role": "Repsol’s 2024 issuer announcement says its legal team is ultimately responsible for reviewing/validating Harvey proposals and deciding how to apply them. The dated 2026 interview describes a Repsol taskforce/champions, adoption support, training, beta testing and peer coaching. No automatic legal decision or filing authority is reported.", "deployment_scope": "A dated 2026 vendor interview confirms that Repsol’s legal team evaluated, piloted and rolled out Harvey across a stated 150-person team. The sources do not establish exact go-live date, percentage adoption in March 2026, or task-level production volume.", "partners": ["Harvey"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Harvey interview: Repsol tracks about 30 power users out of 150; approximately 6–7 meet both usage and qualitative-output definitions. These are adoption-support measures, not productivity outcomes. Repsol’s April 2024 issuer announcement estimated up to three hours per week of repetitive low-legal-content work saved per lawyer during implementation; this was an estimate, not an audited realized saving and predates the window.", "period": "Dated Harvey interview reports rollout across Repsol’s 150-person legal team; original agreement was announced in April 2024, outside this search window. The March 2026 source is a substantive in-window status update, not the original selection date.", "attribution": "Harvey vendor interview with Repsol Legal Counsel Javier Díez Tomillo; adoption and scope are named-customer statements, not independently verified.", "source_ids": ["SW5-001", "SW5-002"]}], "economics": "No implementation cost, operating cost, payback, or realized financial benefit disclosed. The up-to-three-hours-per-week estimate in the 2024 announcement is not an observed or audited cash saving.", "timeline": [{"date": "2026-03-30", "event": "Repsol announced its Harvey agreement on 2024-04-04, before the search window. A Harvey interview dated 2026-03-30 says Repsol’s 150-person legal team had completed evaluation, piloting and rollout; Javier Díez Tomillo says the team tracks about 30 power users, of whom about six or seven meet both quantitative and qualitative criteria. This is evidence of ongoing rollout/adoption by March 2026, not the date Harvey first entered production.", "source_ids": ["SW5-001", "SW5-002"]}], "limitations": ["The new dated source is a Harvey vendor interview, not a Repsol-issued performance report. It supports rollout across 150 people but not a completed adoption percentage or validated performance result. An undated Harvey Repsol customer page makes stronger claims (“96%” adoption; “4–6 hours” saved per lawyer/week, “two years later”) but has no verifiable publication date; treat those claims as an undated supplemental lead, not a dated in-window outcome."], "assessment": "Promotes the previously undated supplemental lead to its first dated initiative profile. Distinct from AI-031/032 enterprise platforms; this is legal-team adoption, not a new 2026 product launch.", "sources": [{"title": "Harvey Power Users: “People Call Me Harvey”", "url": "https://www.harvey.ai/blog/harvey-power-users-people-call-me-harvey", "date": "2026-03-30", "type": "Harvey customer interview with Repsol Legal Counsel", "locator": "Byline/date and rollout scope lines 62–77; legal-team size, rollout lead, training and power users.", "review_status": "Full page opened and read; date, attributed interview, 150-person legal team, 30 tracked power users and 6–7 meeting both definitions verified.", "publisher": "Harvey", "id": "SW5-001", "accessed": "2026-10-01"}, {"title": "Repsol becomes the first Spanish in-house legal team to use Harvey", "url": "https://www.repsol.com/en/press-room/press-releases/2024/repsol-becomes-the-first-spanish-in-house-legal-team-to-use-harvey/index.cshtml", "date": "2024-04-04 (pre-window context only)", "type": "Repsol issuer press release", "locator": "Press release date and scope lines 675–689: legal document analysis, contract/regulation comparison, clause suggestions, translation, legal queries, summaries; human validation; estimated up to three hours/week saved per lawyer.", "review_status": "Full issuer page opened and read. Pre-window agreement and estimated time savings used only as history; not counted as an in-window deployment or measured outcome.", "publisher": "Repsol", "id": "SW5-002", "accessed": "2026-10-01"}], "source_ids": ["SW5-001", "SW5-002"], "field_sources": {"workflow": ["SW5-001", "SW5-002"], "technology_data": ["SW5-001", "SW5-002"], "deployment_scope": ["SW5-001", "SW5-002"], "practical_relevance": ["SW5-001", "SW5-002"]}, "verification": "verified partner", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Harvey vendor interview with Repsol Legal Counsel Javier Díez Tomillo; adoption and scope are named-customer statements, not independently verified."}, {"key": "swarm5-ai-238", "packet_id": "swarm5-ai-238", "record_ids": ["AI-238"], "existing_ids": [], "related_existing_ids": [], "company": "Spearpoint Resources", "title": "AI-assisted non-op revenue statement extraction and WolfePak export at Spearpoint Resources", "department": "Finance, investment & strategy", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "In production since May 2026 per the vendor case study; output is exported in WolfePak-compatible format. The source does not say that Joltly posts transactions autonomously.", "practical_relevance": "Concrete in-window named operator deployment in oil-and-gas revenue accounting; addresses purchaser-format variation and extreme statement-size variability with attributed staffing-capacity metrics and explicit human exception review.", "business_problem": "", "workflow": "Spearpoint’s small accounting team processes monthly non-operated revenue statements arriving in 27 purchaser formats. Joltly extracts structured line items at well/lease level—including production dates, volumes, prices, interests, taxes and deductions—then users review exceptions and export the results for WolfePak posting.", "technology_data": "Upload or email-forward PDF statements without purchaser-specific templates. Joltly extracts each line item and retains coordinates linking figures to source-page locations. The accounting team reviews exceptions and exports in the format WolfePak expects. The case does not disclose models, training regimen, integration/API method, exception thresholds or posting automation.", "integration_human_role": "Revenue accounting staff review exceptions rather than every line, verify source-page coordinates when needed, then export to WolfePak and post. The case says each extracted value remains tied to its page location; it does not specify exception thresholds, accuracy, approval segregation, or automated posting.", "deployment_scope": "In production since May 2026 per the vendor case study; output is exported in WolfePak-compatible format. The source does not say that Joltly posts transactions autonomously.", "partners": ["Joltly"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Spearpoint Controller Elle Ross estimates Joltly gives the team 20–30 hours per week back across three entities. From May through August 2026 the case reports 82 statements processed, 5,441 pages and 56,358 line items across 27 purchaser formats. Time estimate is customer-reported; no baseline time study, quality/exception rate or independent validation is supplied.", "period": "Case study published 2026-09-05; says go-live May 2026 and reports statement activity from May through August 2026.", "attribution": "Joltly vendor case study quoting named Spearpoint Controller Elle Ross; throughput and workload benefit are customer-reported and not independently validated.", "source_ids": ["SW5-003"]}], "economics": "The 20–30 hours/week figure is reported staff capacity released, not payroll reduction or realized cash saving. No customer-specific implementation fee, operating cost or payback is disclosed.", "timeline": [{"date": "2026-09-05", "event": "Joltly says Spearpoint has been live since May 2026. The case study published 2026-09-05 reports data through end-August 2026: 82 statements, 5,441 pages, 56,358 line items across three entities and 27 purchaser formats.", "source_ids": ["SW5-003"]}], "limitations": ["Vendor-authored case study; no audited results, accuracy/error rate, control testing, implementation cost, customer-specific pricing, or independent attribution. The output is reviewed/exported by people; do not describe it as fully autonomous posting. It is a named but small non-operated asset owner described without basin or asset-level details."], "assessment": "Concrete in-window named operator deployment in oil-and-gas revenue accounting; addresses purchaser-format variation and extreme statement-size variability with attributed staffing-capacity metrics and explicit human exception review.", "sources": [{"title": "Three entities, 27 purchaser formats, and 20 to 30 hours a week back", "url": "https://www.joltly.io/blog/spear-point-revenue-accounting-case-study", "date": "2026-09-05", "type": "Joltly customer case study quoting Spearpoint controller", "locator": "Date lines 9–14; workflow/architecture lines 24–47 and 130–154; May–August 2026 volumes lines 49–78; outcome quote lines 156–170.", "review_status": "Full page opened/read. Named customer, controller quote, workflow, go-live date, period/throughput, review queue and export to WolfePak verified.", "publisher": "Joltly", "id": "SW5-003", "accessed": "2026-10-01"}], "source_ids": ["SW5-003"], "field_sources": {"workflow": ["SW5-003"], "technology_data": ["SW5-003"], "deployment_scope": ["SW5-003"], "practical_relevance": ["SW5-003"]}, "verification": "verified partner", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Joltly vendor case study quoting named Spearpoint Controller Elle Ross; throughput and workload benefit are customer-reported and not independently validated."}, {"key": "swarm5-ai-239", "packet_id": "swarm5-ai-239", "record_ids": ["AI-239"], "existing_ids": [], "related_existing_ids": [], "company": "INPEX Corporation", "title": "Machine-learning surrogate accelerates INPEX CCS reservoir simulation", "department": "Reservoir engineering & field development", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Research & development", "stage": "Model development stated. Project-level deployment or operational use is not established.", "practical_relevance": "Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.", "business_problem": "", "workflow": "A high-speed machine-learning surrogate model provides rapid predictions of pressure, gas saturation and other reservoir parameters from geological and operating conditions, supporting reservoir-efficiency and pressure-rise evaluation plus optimization and sensitivity analysis for CCS decision-making.", "technology_data": "INPEX says it developed the surrogate model to reduce the time required for CCS reservoir simulation. The report does not disclose model architecture, training data volume, simulator integration, validation set, or how technical teams review or approve its outputs.", "integration_human_role": "The report says the analyses support decision-making; specific users, review steps and approval authority are not identified. No automated control is described.", "deployment_scope": "Model development stated. Project-level deployment or operational use is not established.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "The stated benefit is reduced simulation time and the ability to evaluate parameters quickly; no measured speedup, accuracy figure, validation result or project result is quantified.", "period": "Disclosure in Integrated Report 2024, published 4 July 2025; development and operational dates not specified.", "attribution": "Operator report; described model capability is not independently validated.", "source_ids": ["SW5-004", "SW5-005", "SW5-006"]}], "economics": "No CCS project cost, saved engineering hours, capital avoided or payback disclosed.", "timeline": [{"date": "2025-07-04", "event": "The July 2025 report states INPEX had developed the surrogate model. The July 2026 report Growth Strategy/Digital Strategy section does not mention or update this CCS surrogate model. It gives no development date, CCS project deployment date, or evidence it was used in field operations.", "source_ids": ["SW5-004", "SW5-005", "SW5-006"]}], "limitations": ["This is an engineering decision-support model, not evidence of AI control of an operating CCS asset. No specific field, injection project, go-live date, user role, validation method, accuracy, quantified time reduction or economic result is reported."], "assessment": "Adds a dated, named machine-learning application for CCS reservoir evaluation at an operator absent from the current profile set, with practical simulation and decision-support tasks rather than a general AI strategy statement.", "sources": [{"title": "INPEX Integrated Report 2024", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf", "date": "2025-07-04", "type": "Operator integrated report", "locator": "Printed p. 41 (PDF physical page 42), “Thorough Utilization of Digital Technology,” Case example 2: “Accelerating CCS simulation using machine learning”; official integrated-report index dates publication 4 July 2025.", "review_status": "Full 86-page original issuer PDF opened; digital technology section and case example on printed p. 41 / PDF p. 42 reviewed. Issuer index confirms 4 July 2025 publication. Report describes development and decision-support functions, not a quantified deployed-project outcome.", "publisher": "INPEX Corporation", "id": "SW5-004", "accessed": "2026-10-01"}, {"title": "INPEX Integrated Report 2025 — Growth Strategy section", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "date": "2026-07-07", "type": "Operator integrated report (separate Growth Strategy PDF)", "locator": "Printed pp. 39–42 (PDF pp. 40–43), Technical Strategies and Digital Strategy sections; CCS surrogate was not discussed in the section reviewed.", "review_status": "Original 21-page issuer Growth Strategy PDF opened and reviewed, including full Digital Strategy section. It reports a separate deployed Ichthys O&M AI case and does not update or quantify the CCS surrogate model.", "publisher": "INPEX Corporation", "id": "SW5-005", "accessed": "2026-10-01"}, {"title": "INPEX Integrated Report index — publication date confirmation", "url": "https://www.inpex.com/english/ir/library/integrated_report/", "date": "Undated live index; accessed 2026-10-01", "type": "Issuer report index", "locator": "Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.", "review_status": "Full issuer index page opened and publication dates read.", "publisher": "INPEX Corporation", "id": "SW5-006", "accessed": "2026-10-01"}], "source_ids": ["SW5-004", "SW5-005", "SW5-006"], "field_sources": {"workflow": ["SW5-004", "SW5-005", "SW5-006"], "technology_data": ["SW5-004", "SW5-005", "SW5-006"], "deployment_scope": ["SW5-004", "SW5-005", "SW5-006"], "practical_relevance": ["SW5-004", "SW5-005", "SW5-006"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator report; described model capability is not independently validated."}, {"key": "swarm5-ai-240", "packet_id": "swarm5-ai-240", "record_ids": ["AI-240"], "existing_ids": [], "related_existing_ids": [], "company": "INPEX Corporation", "title": "Ichthys AI tool searches operational reports and structures maintenance history", "department": "Maintenance & reliability", "secondary_departments": ["IT, data & enterprise transformation"], "category": "Internal adoption", "stage": "Developed and deployed at Ichthys; specific go-live date and adoption scale are not disclosed.", "practical_relevance": "A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.", "business_problem": "", "workflow": "An Ichthys-specific generative-AI tool converts daily reports into a searchable database and answers natural-language questions with past operating performance and troubleshooting cases. A connected function structures maintenance history by categorizing work descriptions, failure details and affected components.", "technology_data": "INPEX says the Technical Headquarters tailored and deployed the tool to site needs. It draws on Ichthys operation reports and maintenance records, including approximately 110,000 previously unstructured maintenance records. Model/provider, architecture, integration details, active user count and records inclusion dates are undisclosed.", "integration_human_role": "The report says the search is “approved by engineers” and describes tool use for onsite needs; it does not specify whether every output receives human review or who authorizes follow-up maintenance actions.", "deployment_scope": "Developed and deployed at Ichthys; specific go-live date and adoption scale are not disclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "INPEX reports that searches formerly taking several hours now complete in seconds to minutes; the report graphic gives “seconds to minutes.” It says AI structured approximately 110,000 maintenance records and achieved over 90% classification accuracy. No accuracy metric definition, validation sample, baseline method, error rate or time-study method is provided.", "period": "2025 Integrated Report published 7 July 2026; Ichthys AI tool deployment date is not specified.", "attribution": "Operator report; throughput/quality descriptions are operator-reported, not independently validated.", "source_ids": ["SW5-005", "SW5-006"]}], "economics": "No cost reduction, avoided downtime, maintenance savings or payback is quantified or attributed to this tool.", "timeline": [{"date": "2026-07-07", "event": "The July 2026 report says the Technical Headquarters developed and deployed an Ichthys-dedicated AI tool; the deployment/go-live date is not stated. The underlying operation reports cover the period since Ichthys began operating in 2018.", "source_ids": ["SW5-005", "SW5-006"]}], "limitations": ["Operator-reported implementation and process metrics. The reported >90% concerns classification accuracy, not overall answer accuracy. Faster information retrieval and record structuring are not quantified as downtime or production improvements. Engineer approval of search quality does not establish mandatory human review of every response or maintenance action."], "assessment": "A concrete, site-specific O&M use case at INPEX’s operating Ichthys LNG project, with disclosed data scale and workflow measures; separate from the CCS reservoir-simulation surrogate candidate.", "sources": [{"title": "INPEX Integrated Report 2025 — Growth Strategy section", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "date": "2026-07-07", "type": "Operator integrated report (separate Growth Strategy PDF)", "locator": "Printed p. 40 (PDF physical page 17 of the 21-page Growth Strategy extract), “Technical Strategies,” Ichthys LNG O&M AI case: development and deployment, report retrieval and maintenance-record structuring; official report index confirms report publication 7 July 2026.", "review_status": "Original 21-page issuer Growth Strategy PDF opened; full relevant Ichthys case and digital strategy pages reviewed. Issuer index separately confirms publication date.", "publisher": "INPEX Corporation", "id": "SW5-005", "accessed": "2026-10-01"}, {"title": "INPEX Integrated Report index — publication date confirmation", "url": "https://www.inpex.com/english/ir/library/integrated_report/", "date": "Undated live index; accessed 2026-10-01", "type": "Issuer report index", "locator": "Latest Report section states Integrated Report 2025 published July 7, 2026; Back Numbers section dates Integrated Report 2024 July 4, 2025.", "review_status": "Full issuer index page opened and publication dates read.", "publisher": "INPEX Corporation", "id": "SW5-006", "accessed": "2026-10-01"}], "source_ids": ["SW5-005", "SW5-006"], "field_sources": {"workflow": ["SW5-005", "SW5-006"], "technology_data": ["SW5-005", "SW5-006"], "deployment_scope": ["SW5-005", "SW5-006"], "practical_relevance": ["SW5-005", "SW5-006"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator report; throughput/quality descriptions are operator-reported, not independently validated."}, {"key": "swarm5-ai-241", "packet_id": "swarm5-ai-241", "record_ids": ["AI-241"], "existing_ids": [], "related_existing_ids": [], "company": "Saipem", "title": "AI video safety alerts with human verification", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Company reports video analytics at 11 operational sites in 2025, after additions on two vessels and two yards. A December 2025 Livorno installation agreement is separate from completed installation evidence.", "practical_relevance": "Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.", "business_problem": "", "workflow": "Cameras identify selected unsafe conditions and notify trained safety personnel. Karimun examples include missing PPE and unsafe proximity to machinery.", "technology_data": "The January 2026 notice describes local software, a manufacturer-provided model and retraining if false alerts warrant it. Routine video is retained for 24 hours; incident and training retention follow separate rules. The March 2025 vessel notice separately specifies validation on unused data.", "integration_human_role": "A person checks each alert and decides the response. The system supports existing supervision; it does not assume decision authority.", "deployment_scope": "Company reports video analytics at 11 operational sites in 2025, after additions on two vessels and two yards. A December 2025 Livorno installation agreement is separate from completed installation evidence.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "11 operational sites reported for 2025. No attributable incident reduction or false-positive rate disclosed.", "period": "Disclosure status; exact operating dates not stated.", "attribution": "Company-reported scope; no independent performance validation.", "source_ids": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"]}], "economics": "No project cost, monetized benefit or payback disclosed.", "timeline": [{"date": "2026-05-20", "event": "Vessel notice: 31 March 2025; corporate notice: 28 January 2026; sustainability release: 20 May 2026.", "source_ids": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"]}], "limitations": ["Notices cover different entities/dates. The March 2025 vessel notice has inconsistent subsidiary naming; use it for the named vessel and stated controls only. No deployment-wide accuracy or causal safety result."], "assessment": "Adds source-backed operating controls and rollout boundaries to an underrepresented HSE workflow.", "sources": [{"title": "Saipem publishes 2025 sustainability results", "url": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "date": "2026-05-20", "locator": "Dated announcement and safety technology discussion", "type": "Issuer disclosure", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-007", "accessed": "2026-10-01"}, {"title": "Saipem Smart Cameras privacy notice", "url": "https://www.saipem.com/sites/default/files/2026-01/75_Privacy_Notice_Smart_Cameras.pdf", "date": "2026-01-28", "locator": "pp. 1–2: operating model, human verification, retraining and retention; date printed on all pages", "type": "Issuer operational notice", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-008", "accessed": "2026-10-01"}, {"title": "Saipem MISR Smart Cameras notice — Saipem 10000", "url": "https://www.saipem.com/sites/default/files/2025-04/Information%20notice%20Smart%20Cameras_Saipem%20MISR%20for%20Petroleum%20Services%20%28S.A.E.%29_0.pdf", "date": "2025-03-31", "locator": "pp. 1–3: named vessel, validation and responsible safety personnel", "type": "Issuer operational notice", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-009", "accessed": "2026-10-01"}, {"title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf", "date": "2026", "locator": "Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.", "type": "Issuer sustainability report", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-010", "accessed": "2026-10-01"}, {"title": "Saipem Consolidated Sustainability Statement 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf", "date": "2026", "locator": "Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests", "type": "Issuer annual-report extract", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-011", "accessed": "2026-10-01"}], "source_ids": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"], "field_sources": {"workflow": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"], "technology_data": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"], "deployment_scope": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"], "practical_relevance": ["SW5-007", "SW5-008", "SW5-009", "SW5-010", "SW5-011"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported scope; no independent performance validation."}, {"key": "swarm5-ai-242", "packet_id": "swarm5-ai-242", "record_ids": ["AI-242"], "existing_ids": [], "related_existing_ids": [], "company": "Saipem", "title": "AI-assisted drone surveys for dropped-object hazards", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Operational tests reported during 2025; routine fleet-wide deployment is not established.", "practical_relevance": "Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.", "business_problem": "", "workflow": "Drone imagery of elevated structures is processed by a trained AI algorithm to support dropped-object hazard surveys.", "technology_data": "High-resolution cameras feed image analysis; the report places this technology in the Karimun innovation programme.", "integration_human_role": "Not disclosed.", "deployment_scope": "Operational tests reported during 2025; routine fleet-wide deployment is not established.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No measured detection accuracy, inspection productivity or avoided-loss result.", "period": "Disclosure status; exact operating dates not stated.", "attribution": "Company-reported scope; no independent performance validation.", "source_ids": ["SW5-007", "SW5-010", "SW5-011"]}], "economics": "No project cost, monetized benefit or payback disclosed.", "timeline": [{"date": "2026-05-20", "event": "2025 activity reported in a document whose publication was announced on 20 May 2026.", "source_ids": ["SW5-007", "SW5-010", "SW5-011"]}], "limitations": ["The readiness graphic is not used as an outcome or deployment percentage. Asset count and operator approval procedures remain unspecified."], "assessment": "Distinct inspection workflow; kept separate from fixed-camera personnel safety monitoring.", "sources": [{"title": "Saipem publishes 2025 sustainability results", "url": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "date": "2026-05-20", "locator": "Dated announcement and safety technology discussion", "type": "Issuer disclosure", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-007", "accessed": "2026-10-01"}, {"title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf", "date": "2026", "locator": "Printed p. 65 / PDF page 67: Karimun safety technology. Publication announced 20 May 2026; report file path is not publication evidence.", "type": "Issuer sustainability report", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-010", "accessed": "2026-10-01"}, {"title": "Saipem Consolidated Sustainability Statement 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf", "date": "2026", "locator": "Printed pp. 264 and 275 / PDF pp. 148 and 159: union agreement, 11-site video analytics rollout and drone operational tests", "type": "Issuer annual-report extract", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-011", "accessed": "2026-10-01"}], "source_ids": ["SW5-007", "SW5-010", "SW5-011"], "field_sources": {"workflow": ["SW5-007", "SW5-010", "SW5-011"], "technology_data": ["SW5-007", "SW5-010", "SW5-011"], "deployment_scope": ["SW5-007", "SW5-010", "SW5-011"], "practical_relevance": ["SW5-007", "SW5-010", "SW5-011"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported scope; no independent performance validation."}, {"key": "swarm5-ai-243", "packet_id": "swarm5-ai-243", "record_ids": ["AI-243"], "existing_ids": [], "related_existing_ids": [], "company": "MODEC", "title": "Shape Reef Phase II safety-barrier and gas-leak modelling", "department": "HSE & emissions", "secondary_departments": [], "category": "Research & development", "stage": "A 36-month development and testing partnership, not completed Phase II deployment.", "practical_relevance": "Adds an absent FPSO operator and explicit process-safety R&D workflow.", "business_problem": "", "workflow": "Develop risk models for degraded safety barriers and gas-leak scenarios on offshore units.", "technology_data": "The project combines computational-fluid-dynamics and field data, building on earlier Fuzzy BowTie and neural-network leak-detection work. MODEC contributes FPSO operating expertise.", "integration_human_role": "Not disclosed.", "deployment_scope": "A 36-month development and testing partnership, not completed Phase II deployment.", "partners": ["Shell Brasil", "Shape Digital", "Unicamp"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No validated accuracy or incident-reduction result.", "period": "Disclosure status; exact operating dates not stated.", "attribution": "Company-reported scope; no independent performance validation.", "source_ids": ["SW5-012"]}], "economics": "Financed through ANP RD&I-clause funds; amount and commercial payback are undisclosed.", "timeline": [{"date": "2025-08-13", "event": "Partnership announced 13 August 2025; no exact project start/end supplied.", "source_ids": ["SW5-012"]}], "limitations": ["Existing Shape Reef use in Brazil, Mexico and Africa does not establish deployment of the new Phase II models."], "assessment": "Adds an absent FPSO operator and explicit process-safety R&D workflow.", "sources": [{"title": "MODEC, Shell Brasil, Shape Digital and Unicamp safety-barrier AI project", "url": "https://www.modec.com/news/2025/20250813_pr_shape-digital.html", "date": "2025-08-13", "locator": "Announcement paragraphs on 36-month project, Phase II and existing Shape Reef", "type": "Issuer project announcement", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-012", "accessed": "2026-10-01"}], "source_ids": ["SW5-012"], "field_sources": {"workflow": ["SW5-012"], "technology_data": ["SW5-012"], "deployment_scope": ["SW5-012"], "practical_relevance": ["SW5-012"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported scope; no independent performance validation."}, {"key": "swarm5-ai-244", "packet_id": "swarm5-ai-244", "record_ids": ["AI-244"], "existing_ids": [], "related_existing_ids": [], "company": "SBM Offshore", "title": "FPSO maintenance data foundation and planned full-field AI integration", "department": "Maintenance & reliability", "secondary_departments": ["IT, data & enterprise transformation", "Production & field operations"], "category": "Internal adoption", "stage": "Cognite platform expansion announced across the Brazilian fleet; Q3 update describes deployment to part of the overall fleet. Full-field optimization and advanced agents remain prospective.", "practical_relevance": "Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.", "business_problem": "", "workflow": "Unify engineering and operating data for maintenance prioritization and reliability analysis. A linked SLB alliance aims to extend insights from subsea systems through FPSO topsides.", "technology_data": "Cognite Data Fusion supplies the data foundation; Atlas AI is a prospective extension. The SLB alliance names OptiSite and Lumi. The October release describes 18 months of foundation work before scaling.", "integration_human_role": "Maintenance, operations and engineering teams use contextualized information. Autonomous control and approval rules are not disclosed.", "deployment_scope": "Cognite platform expansion announced across the Brazilian fleet; Q3 update describes deployment to part of the overall fleet. Full-field optimization and advanced agents remain prospective.", "partners": ["Cognite", "SLB"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No AI-attributed uptime or maintenance saving. Reported overall fleet uptime is not evidence of AI impact.", "period": "Disclosure status; exact operating dates not stated.", "attribution": "Company-reported scope; no independent performance validation.", "source_ids": ["SW5-013", "SW5-014", "SW5-015"]}], "economics": "No project cost, monetized benefit or payback disclosed.", "timeline": [{"date": "2025-11-13", "event": "Two linked announcements on 2 October 2025; status reiterated 13 November 2025. One programme profile prevents counting the shared data foundation twice.", "source_ids": ["SW5-013", "SW5-014", "SW5-015"]}], "limitations": ["Fleet expansion is not proof that every asset or AI use case is live. No vessel-by-vessel rollout, acceptance tests or quantified benefit."], "assessment": "Adds an absent offshore operator while distinguishing data infrastructure from implemented AI decisions.", "sources": [{"title": "SBM Offshore Q3 2025 trading update", "url": "https://www.sbmoffshore.com/newsroom/sbm-offshore-third-quarter-2025-trading-update/", "date": "2025-11-13", "locator": "Digitalization and AI section; separates platform rollout from broader alliance", "type": "Issuer disclosure", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-013", "accessed": "2026-10-01"}, {"title": "SBM Offshore and Cognite accelerate AI and data collaboration", "url": "https://www.sbmoffshore.com/newsroom/sbm-offshore-and-cognite-partner-to-accelerate-ai-and-data-collaboration-to-optimize-floating-production-operations/", "date": "2025-10-02", "locator": "Dated release: Brazilian fleet scope, 18-month foundation work and prospective maintenance uses", "type": "Issuer disclosure", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-014", "accessed": "2026-10-01"}, {"title": "SLB and SBM Offshore digital alliance", "url": "https://www.sbmoffshore.com/newsroom/slb-and-sbm-offshore-agree-digital-alliance-to-transform-fpso-production-performance/", "date": "2025-10-02", "locator": "OptiSite, Cognite Data Fusion and Lumi; planned full-field integration", "type": "Issuer disclosure", "review_status": "Original document opened and relevant full sections read; issuer claims are not independently validated.", "id": "SW5-015", "accessed": "2026-10-01"}], "source_ids": ["SW5-013", "SW5-014", "SW5-015"], "field_sources": {"workflow": ["SW5-013", "SW5-014", "SW5-015"], "technology_data": ["SW5-013", "SW5-014", "SW5-015"], "deployment_scope": ["SW5-013", "SW5-014", "SW5-015"], "practical_relevance": ["SW5-013", "SW5-014", "SW5-015"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported scope; no independent performance validation."}, {"key": "swarm5-ai-214", "packet_id": "swarm5-ai-214", "record_ids": ["AI-214"], "existing_ids": ["AI-214"], "related_existing_ids": [], "company": "Sentinel Peak Resources / Quantum Energy Partners", "title": "Undated operator corroboration of Intelligent Oilfield steam optimization", "department": "Production & field operations", "secondary_departments": ["HSE & emissions"], "category": "Internal adoption", "stage": "Operator describes use; undated webpage does not establish a new rollout.", "practical_relevance": "Adds operator corroboration while preserving the earlier dated sponsor disclosure and its historical outcome limits.", "business_problem": "", "workflow": "Sentinel Peak identifies machine-learning and AI analysis of reservoir data for real-time steam management.", "technology_data": "Operator corroboration of the existing Quantum/Sentinel Peak case; no new control architecture described.", "integration_human_role": "Approval, override and fallback controls undisclosed.", "deployment_scope": "Operator describes use; undated webpage does not establish a new rollout.", "partners": ["Quantum Energy Partners (sponsor and digital-team co-developer)"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "The operator estimates 200,000 metric tons CO2e per year enabled through intelligent-oilfield technologies. Period, baseline and AI-only contribution are unspecified.", "period": "Undated operator estimate; not a new measured 2026 result.", "attribution": "Operator estimate, not independently validated. Sponsor co-development comes from the existing dated dossier, not this page.", "source_ids": ["SW5-016"]}], "economics": "No project spend, monetized gas savings or payback disclosed.", "timeline": [{"date": "Undated", "event": "Current page accessed 1 October 2026; publication and measurement dates undisclosed.", "source_ids": ["SW5-016"]}], "limitations": ["Do not add this estimate to the baseline’s historical 2023 figures or treat either as AI-only net savings."], "assessment": "Adds operator corroboration while preserving the earlier dated sponsor disclosure and its historical outcome limits.", "sources": [{"title": "Environment — Sentinel Peak Resources", "url": "https://sentinelpeakresources.com/sustainability/environment/", "date": "undated; accessed 2026-10-01", "type": "Operator environment disclosure", "locator": "Step 1: Reduced Steam, lines 15–26", "review_status": "Full original HTML page opened and read. It names Sentinel Peak's Intelligent Oilfield as an ML/AI-powered data system for real-time steam-management optimization and estimates 200,000 MT CO2e/year enabled through intelligent oilfield technologies. The page gives no publication date, calculation, baseline, or AI-only attribution split.", "publisher": "Sentinel Peak Resources", "id": "SW5-016", "accessed": "2026-10-01"}], "source_ids": ["SW5-016"], "field_sources": {"workflow": ["SW5-016"], "technology_data": ["SW5-016"], "deployment_scope": ["SW5-016"], "practical_relevance": ["SW5-016"]}, "verification": "verified primary", "cohort": "Fifth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Operator estimate, not independently validated. Sponsor co-development comes from the existing dated dossier, not this page."}, {"key": "swarm6-ai-245", "packet_id": "swarm6-ai-245", "record_ids": ["AI-245"], "existing_ids": [], "related_existing_ids": ["AI-067"], "company": "PTTEP", "title": "AI-assisted internal audit data review", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Internal adoption", "stage": "Project initiated in 2025; production deployment not established.", "practical_relevance": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "business_problem": "", "workflow": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "technology_data": "Internal Audit Division initiated an AI project; model and integration details are undisclosed. Adjacent TeamMate+ and dashboards are separate systems, not identified as the AI implementation.", "integration_human_role": "Internal Audit Division owns the project; output validation and final decision rights are not described.", "deployment_scope": "Project initiated in 2025; production deployment not established.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No measured outcome disclosed. Faster review is an intended benefit.", "period": "2025 operating period unless stated otherwise", "attribution": "Company-reported; no initiative-specific independent assurance established.", "source_ids": ["SW6-001", "SW6-002"]}], "economics": "No initiative-level cost, benefit or payback disclosed.", "timeline": [{"date": "2026", "event": "2026 report on 2025 activities; precise publication day not established.", "source_ids": ["SW6-001", "SW6-002"]}], "limitations": ["No model, deployment scale, accuracy, approval controls or measured productivity. Do not relabel the separately described audit software as AI."], "assessment": "Combine multiple databases and auditee submissions for audit analysis; automate repeated data formats.", "sources": [{"title": "PTTEP 56-1 One Report 2025 — internal audit", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "2026; referenced in invitation dated 27 February 2026", "locator": "Printed p. 223; physical PDF page 212 of downloaded 557-page file, Oversight of Internal Audit Activities.", "type": "Operator annual report", "review_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "id": "SW6-001", "accessed": "2026-10-01"}, {"title": "PTTEP 2026 AGM invitation", "url": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-invitation-en.pdf", "date": "2026-02-27", "locator": "Physical PDF page 2: dated invitation lists the 2025 One Report as attachment.", "type": "Issuer meeting invitation", "review_status": "Original PDF opened; date and report attachment read.", "id": "SW6-002", "accessed": "2026-10-01"}], "source_ids": ["SW6-001", "SW6-002"], "field_sources": {"workflow": ["SW6-001", "SW6-002"], "technology_data": ["SW6-001", "SW6-002"], "deployment_scope": ["SW6-001", "SW6-002"], "practical_relevance": ["SW6-001", "SW6-002"]}, "verification": "verified primary", "cohort": "Sixth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported; no initiative-specific independent assurance established."}, {"key": "swarm6-ai-246", "packet_id": "swarm6-ai-246", "record_ids": ["AI-246"], "existing_ids": [], "related_existing_ids": [], "company": "MOL Group", "title": "MOL MOVE AI customer-offer personalization", "department": "Trading & marketing", "secondary_departments": [], "category": "Internal adoption", "stage": "Successfully tested and being scaled in 2025 AGM material covering FY2024.", "practical_relevance": "Personalize loyalty-program experiences and offers.", "business_problem": "", "workflow": "Personalize loyalty-program experiences and offers.", "technology_data": "MOL reports testing AI and scaling the capability within Consumer Services; no model or vendor disclosed.", "integration_human_role": "Not specified.", "deployment_scope": "Successfully tested and being scaled in 2025 AGM material covering FY2024.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No AI-attributed outcome. Nearly five million app downloads describe the broader platform.", "period": "FY2024 performance review in 2025 AGM material.", "attribution": "Company-reported; no initiative-specific independent assurance established.", "source_ids": ["SW6-003"]}], "economics": "No initiative-level cost, benefit or payback disclosed.", "timeline": [{"date": "2025", "event": "Disclosure in 2025 AGM material for meeting on 24 April; underlying FY2024 review. Exact publication and go-live dates not established.", "source_ids": ["SW6-003"]}], "limitations": ["Eight-market program reach is not eight-market AI deployment. No test population, conversion uplift, cost or model controls disclosed."], "assessment": "Personalize loyalty-program experiences and offers.", "sources": [{"title": "MOL 2025 AGM documents — Consumer Services", "url": "https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf", "date": "2025 AGM material; meeting 24 April 2025", "locator": "Printed p. 27; physical PDF page 29, MOL MOVE paragraph.", "type": "Issuer AGM management report", "review_status": "Original PDF opened; full relevant paragraph and surrounding pages read.", "id": "SW6-003", "accessed": "2026-10-01"}], "source_ids": ["SW6-003"], "field_sources": {"workflow": ["SW6-003"], "technology_data": ["SW6-003"], "deployment_scope": ["SW6-003"], "practical_relevance": ["SW6-003"]}, "verification": "verified primary", "cohort": "Sixth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported; no initiative-specific independent assurance established."}, {"key": "swarm6-ai-247", "packet_id": "swarm6-ai-247", "record_ids": ["AI-247"], "existing_ids": [], "related_existing_ids": ["AI-147"], "company": "Ecopetrol", "title": "EVO talent-profile and role matching", "department": "Workforce & learning", "secondary_departments": [], "category": "Internal adoption", "stage": "Described among three consolidated HR technology enablers for 2025; user scale not disclosed.", "practical_relevance": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "business_problem": "", "workflow": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "technology_data": "In-house analytics and AI tool; technical architecture undisclosed.", "integration_human_role": "Supports people-management decisions; automated final decisions are not established.", "deployment_scope": "Described among three consolidated HR technology enablers for 2025; user scale not disclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No measured outcome disclosed.", "period": "2025 operating period unless stated otherwise", "attribution": "Company-reported; no initiative-specific independent assurance established.", "source_ids": ["SW6-004"]}], "economics": "No initiative-level cost, benefit or payback disclosed.", "timeline": [{"date": "2026", "event": "2026 report on 2025 activities; precise publication day not established.", "source_ids": ["SW6-004"]}], "limitations": ["Do not attach adjacent SAP onboarding or Signavio results to EVO. Bias testing, privacy controls and human override are unspecified."], "assessment": "Match employee profiles to roles, identify skill gaps and inform succession, selection and training.", "sources": [{"title": "Ecopetrol 2025 Integrated Management Report — EVO talent map", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "locator": "Printed p. 229; physical PDF page 230, TEC para el talento humano / EVO.", "type": "Operator annual report", "review_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "id": "SW6-004", "accessed": "2026-10-01"}], "source_ids": ["SW6-004"], "field_sources": {"workflow": ["SW6-004"], "technology_data": ["SW6-004"], "deployment_scope": ["SW6-004"], "practical_relevance": ["SW6-004"]}, "verification": "verified primary", "cohort": "Sixth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported; no initiative-specific independent assurance established."}, {"key": "swarm6-ai-147", "packet_id": "swarm6-ai-147", "record_ids": ["AI-147"], "existing_ids": ["AI-147"], "related_existing_ids": [], "company": "Ecopetrol", "title": "AI portfolio: 2025 investment and benefits disclosure", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "2025 portfolio disclosure; initiative count does not prove all tools are live.", "practical_relevance": "Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.", "business_problem": "", "workflow": "Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.", "technology_data": "More than 35 initiatives generated during 2025; eight analytics hubs. Costs and outcomes are aggregated, not allocated to individual tools.", "integration_human_role": "Not specified.", "deployment_scope": "2025 portfolio disclosure; initiative count does not prove all tools are live.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "2025: US$1.88m reported AI investment, US$2.06m revenue increase and US$1.44m cost reduction (US$3.50m total benefits). Reported ROI: 7%, cumulative since 2021.", "period": "2025 annual metrics; ROI cumulative from 2021 through 2025.", "attribution": "Company-reported; no initiative-specific independent assurance established.", "source_ids": ["SW6-004"]}], "economics": "Some portfolio economics are disclosed, but the cost perimeter is incomplete. Do not compute a one-year ROI from US$3.50m / US$1.88m; spending, benefit timing and the stated ROI have different scopes.", "timeline": [{"date": "2026", "event": "2026 report on 2025 activities; precise publication day not established.", "source_ids": ["SW6-004"]}], "limitations": ["The 7% uses five-year consolidated, EBITDA-aligned figures with lagged benefits. The investment measure excludes other AI spending without direct financial impact. No project-level payback or independently verified causal return."], "assessment": "Portfolio reporting across process optimization, failure prediction, maintenance and energy efficiency.", "sources": [{"title": "Ecopetrol 2025 Integrated Management Report — AI KPIs", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "locator": "Printed pp. 318 and 447; physical PDF pp. 319 and 448, table 7.115 and explanatory note.", "type": "Operator annual report", "review_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "id": "SW6-004", "accessed": "2026-10-01"}], "source_ids": ["SW6-004"], "field_sources": {"workflow": ["SW6-004"], "technology_data": ["SW6-004"], "deployment_scope": ["SW6-004"], "practical_relevance": ["SW6-004"]}, "verification": "verified primary", "cohort": "Sixth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported; no initiative-specific independent assurance established."}, {"key": "swarm6-ai-148", "packet_id": "swarm6-ai-148", "record_ids": ["AI-148"], "existing_ids": ["AI-148"], "related_existing_ids": [], "company": "Ecopetrol", "title": "CADE energy dispatch: Rubiales model development update", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Energy-dispatch center operational; Rubiales dispatch model under MVP construction in H2 2025.", "practical_relevance": "Consolidate electrical operating signals and develop dispatch recommendations for Rubiales operations.", "business_problem": "", "workflow": "Consolidate electrical operating signals and develop dispatch recommendations for Rubiales operations.", "technology_data": "CADE uses advanced analytics and generative AI for decision support. The dispatch model passed through proof-of-concept work in H1 2025 and MVP construction in H2.", "integration_human_role": "Decision support and proposed operating setpoints; final approval and override procedure not specified.", "deployment_scope": "Energy-dispatch center operational; Rubiales dispatch model under MVP construction in H2 2025.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No quantified AI-attributed result disclosed.", "period": "2025 operating period unless stated otherwise", "attribution": "Company-reported; no initiative-specific independent assurance established.", "source_ids": ["SW6-004"]}], "economics": "No initiative-level cost, benefit or payback disclosed.", "timeline": [{"date": "2026", "event": "2026 report on 2025 activities; precise publication day not established.", "source_ids": ["SW6-004"]}], "limitations": ["An operating center does not establish deployed autonomous dispatch. No final model acceptance, production rollout, control bounds, cost saving or payback disclosed."], "assessment": "Consolidate electrical operating signals and develop dispatch recommendations for Rubiales operations.", "sources": [{"title": "Ecopetrol 2025 Integrated Management Report — CADE", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "locator": "Printed p. 89; physical PDF page 90, Centro Autónomo de Despacho Energético.", "type": "Operator annual report", "review_status": "Original PDF downloaded and relevant full pages read; source text and page locators independently checked.", "id": "SW6-004", "accessed": "2026-10-01"}], "source_ids": ["SW6-004"], "field_sources": {"workflow": ["SW6-004"], "technology_data": ["SW6-004"], "deployment_scope": ["SW6-004"], "practical_relevance": ["SW6-004"]}, "verification": "verified primary", "cohort": "Sixth gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Company-reported; no initiative-specific independent assurance established."}, {"key": "swarm7-ai-248", "packet_id": "swarm7-ai-248", "record_ids": ["AI-248"], "existing_ids": [], "related_existing_ids": [], "company": "Continental Resources", "title": "Agent-assisted drilling and completion data validation", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Named 2025 award recipient; current undated vendor case describes an operating workflow.", "practical_relevance": "Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.", "business_problem": "", "workflow": "UiPath interprets inconsistent well-data emails; a Databricks agent validates context and scores accuracy. Robots return errors to suppliers and deliver validated data to engineers. Award material also names WellView drilling-report integration.", "technology_data": "UiPath Agent Builder, Robots and Databricks; WellView named in award profile. Initial Action Center validation was later removed according to the current case; remaining exception controls are undisclosed.", "integration_human_role": "Case describes removal of initial manual validation; ongoing escalation and quality controls not disclosed.", "deployment_scope": "Named 2025 award recipient; current undated vendor case describes an operating workflow.", "partners": ["UiPath", "Databricks"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Undated vendor case: 200+ emails/day, approximately five minutes saved per previously manual email, one reusable robot supporting 40+ processes. Undated award page: accuracy rose from about 70% to the high 90% range.", "period": "Current metrics undated; award cohort 2025.", "attribution": "Source-reported claims; no independent validation.", "source_ids": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"]}], "economics": "Reported time capacity, not verified labor-cost reduction. Implementation and running costs absent.", "timeline": [{"date": "2025-09-29", "event": "Award announcement on 2025-09-29; detailed case and winners page undated. Do not date their outcomes to the award release.", "source_ids": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"]}], "limitations": ["2025-09-29 dates the award announcement only. Current workflow and metric dates are unknown; 40+ processes are not 40 AI initiatives. No test-set definition, error monitoring or independently verified cash savings."], "assessment": "Concrete upstream document workflow joins agent reasoning, validation and routing; distinct from autonomous physical drilling.", "sources": [{"title": "Continental Resources Streamlines Operations with Automation: From fragmented emails to orchestrated operations at Continental Resources", "url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "date": "No displayed publication/update date; accessed 2026-10-01", "locator": "Full page opened/read. Workflow and transition from human validation to automated handling at lines 241-255; quoted Sr. IT Manager Justin Cornell at 257-259; results at 260-265; 200+ emails/40+ processes/~5 minutes at 228-237.", "review_status": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed.", "type": "Vendor customer case", "id": "SW7-001", "accessed": "2026-10-01"}, {"title": "Honoring the 2025 UiPath AI25 Winners", "url": "https://www.uipath.com/ai/ai25-awards/winners", "date": "Undated page; references 2025 award cohort; accessed 2026-10-01", "locator": "Full page opened/read; Continental listing at lines 294-302, reports WellView and Databricks integration and high-90% accuracy vs ~70%, near-equal manual-intervention reduction.", "review_status": "Full page read. No publication/update or metric-measurement date displayed.", "type": "Vendor customer case", "id": "SW7-002", "accessed": "2026-10-01"}, {"title": "UiPath Announces 2025 AI25 Awards Winners", "url": "https://www.uipath.com/newsroom/uipath-announces-2025-ai25-awards-winners", "date": "2025-09-29", "locator": "Full press-release content reviewed from issuer's newsroom/search result. Names Continental among 2025 AI25 award winners. Does not disclose workflow details or performance metrics.", "review_status": "Dated primary UiPath press release reviewed. Used solely to date the award announcement, not the operational metrics.", "type": "Vendor customer case", "id": "SW7-003", "accessed": "2026-10-01"}, {"title": "UiPath + Databricks: Data Intelligence Meets Agentic Automation", "url": "https://www.uipath.com/platform/agentic-automation/ai-ecosystem/databricks", "date": "Undated page; accessed 2026-10-01", "type": "Vendor partner page", "locator": "Continental customer panel: Databricks Genie Agents and UiPath Agents for email processing.", "review_status": "Original page reviewed; no date or incremental result.", "id": "SW7-004", "accessed": "2026-10-01"}], "source_ids": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"], "field_sources": {"workflow": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"], "technology_data": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"], "deployment_scope": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"], "practical_relevance": ["SW7-001", "SW7-002", "SW7-003", "SW7-004"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Source-reported claims; no independent validation."}, {"key": "swarm7-ai-097", "packet_id": "swarm7-ai-097", "record_ids": ["AI-097"], "existing_ids": ["AI-097"], "related_existing_ids": [], "company": "Chord Energy", "title": "AI-driven rod-lift optimization", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Scaled during 2025 to approximately 99% of wells on rod lift.", "practical_relevance": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "business_problem": "", "workflow": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "technology_data": "Model inputs, vendor, architecture and control/approval mode are undisclosed.", "integration_human_role": "Not disclosed; do not infer autonomous control.", "deployment_scope": "Scaled during 2025 to approximately 99% of wells on rod lift.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Operator reports early approximately 25% improvement in rod-pump run times. The 99% denominator is rod-lift wells, not the entire well fleet.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer earnings releases and investor deck; performance not independently assured.", "source_ids": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"]}], "economics": "May 2026 deck: ~$50m Production/LOE, ~$25m corporate and $160m companywide savings refer to mixed improvement programs, not AI-only savings.", "timeline": [{"date": "2026-02-25", "event": "2025 deployment reported 2026-02-25; May deck supplies mixed-program economics; August release confirms continuing use.", "source_ids": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"]}], "limitations": ["Company-reported early result without a disclosed comparison method. Broader workover, ESP and savings results cannot be isolated to AI."], "assessment": "Chord applies AI-driven machine learning to rod-lift pumping operations; its August 2026 release confirms continued artificial-lift optimization.", "sources": [{"title": "Chord Energy FY2025 results release", "date": "2026-02-25", "type": "issuer earnings release", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "locator": "Operations Update, lines 210–216; specifically line 216 (AI deployment, run time metric, workover rig days).", "access": "Full source opened and read.", "review_status": "Full source opened and read.", "id": "SW7-005", "accessed": "2026-10-01"}, {"title": "Chord Energy 1Q26 Earnings Presentation", "date": "2026-05-05", "type": "issuer investor presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2026/05/CHRD-Q1-2026-Earnings-Presentation.pdf", "locator": "Slide 16; full extracted slide lines 457–505.", "access": "Full 28-page PDF opened; slide 16 read.", "review_status": "Full 28-page PDF opened; slide 16 read.", "id": "SW7-006", "accessed": "2026-10-01"}, {"title": "Chord Energy 2Q26 results release", "date": "2026-08-05", "type": "issuer earnings release", "url": "https://ir.chordenergy.com/2026-08-05-Chord-Energy-Reports-Second-Quarter-2026-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Updates-2026-Outlook", "locator": "Operations Update, lines 156–161; continued AI artificial-lift optimization, with no fresh AI-specific result.", "access": "Full source opened and read.", "review_status": "Full source opened and read.", "id": "SW7-007", "accessed": "2026-10-01"}, {"title": "Chord Energy 2Q25 Investor Presentation", "date": "2025-08-06", "type": "issuer investor presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2025/08/CHRD-2Q25-Earnings-Presentation.pdf", "locator": "Slide 11 / parsed PDF page 10, lines 273–306.", "access": "Full 24-page PDF opened; lists AI/ML scheduling and artificial-lift optimization, gas-lift algorithms, predictive maintenance, and remote monitoring as continuous-improvement initiatives; no distinct deployed AI result or attribution.", "review_status": "Full 24-page PDF opened; lists AI/ML scheduling and artificial-lift optimization, gas-lift algorithms, predictive maintenance, and remote monitoring as continuous-improvement initiatives; no distinct deployed AI result or attribution.", "id": "SW7-008", "accessed": "2026-10-01"}], "source_ids": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"], "field_sources": {"workflow": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"], "technology_data": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"], "deployment_scope": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"], "practical_relevance": ["SW7-005", "SW7-006", "SW7-007", "SW7-008"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer earnings releases and investor deck; performance not independently assured."}, {"key": "swarm7-ai-006", "packet_id": "swarm7-ai-006", "record_ids": ["AI-006"], "existing_ids": ["AI-006"], "related_existing_ids": [], "company": "Devon Energy", "title": "Autonomous artificial-lift optimization at operating wells", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Operating and scaled, as of May 6, 2026. >850 wells on fully autonomous artificial-lift optimization, per CEO on earnings call.", "practical_relevance": "AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.", "business_problem": "", "workflow": "AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.", "technology_data": "CEO describes AI translating drilling, completion and production data flows into actionable intelligence, and AI-created tools performing real-time artificial-lift optimization. No model architecture, lift types, control guardrails, or connectivity details disclosed.", "integration_human_role": "Engineering teams build/use tools; the reported 850-well subset is described as fully autonomous. The source does not specify exception handling or human approval steps.", "deployment_scope": "Operating and scaled, as of May 6, 2026. >850 wells on fully autonomous artificial-lift optimization, per CEO on earnings call.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "May 2026 CEO remarks report >850 fully autonomous lift wells without a new quantified productivity result. Earlier 2025 pilot uplift remains in the original profile and is not remeasured here.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer-reported; no independent validation.", "source_ids": ["FU001"]}], "economics": "CEO separately linked AI value to drilling/completion time and lower capital costs, and cited business optimization savings; no amount is attributed specifically to AI or the lift tool.", "timeline": [{"date": "2026-05-06", "event": "Operating and scaled, as of May 6, 2026.", "source_ids": ["FU001"]}], "limitations": ["Company-level CEO assertion; no baseline, period, control group, production uplift, uptime, savings, lift mix, well list or attribution. Do not allocate Devon's $1bn business-optimization target/achievement to AI."], "assessment": "AI-created tools use production and completion data to optimize artificial-lift settings; Devon reports fully autonomous control at more than 850 wells.", "sources": [{"title": "Devon Energy Q1 2026 Earnings Call Corrected Transcript", "date": "2026-05-06", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "locator": "pp. 3-4 (PDF pp. 3-4), CEO Clay Gaspar prepared remarks, lines 108-128; fully read 20-page transcript", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "FU001", "accessed": "2026-10-01"}], "source_ids": ["FU001"], "field_sources": {"workflow": ["FU001"], "technology_data": ["FU001"], "deployment_scope": ["FU001"], "practical_relevance": ["FU001"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer-reported; no independent validation."}, {"key": "swarm7-ai-249", "packet_id": "swarm7-ai-249", "record_ids": ["AI-249"], "existing_ids": [], "related_existing_ids": [], "company": "Devon Energy", "title": "AI alerting from SCADA and decision-support-center data for operational / methane issues", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024. Sensor-system coverage is stated at 100% of production operations collectively; AI deployment scale itself is not stated.", "practical_relevance": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "business_problem": "", "workflow": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "technology_data": "SCADA plus decision-support-center data. 2024 continuous monitoring devices covered 46% of production; integrated OGI, aerial flyovers and continuous monitoring together gave comprehensive monitoring coverage across 100% of production operations. The report does not say that AI analyzes all these sources or runs over the full covered footprint.", "integration_human_role": "Alert prompts staff to take action. Devon separately says detected-leak work orders are automated; source does not say AI autonomously creates or closes those work orders.", "deployment_scope": "AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024. Sensor-system coverage is stated at 100% of production operations collectively; AI deployment scale itself is not stated.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No AI-attributable emissions, repair-time, uptime or financial result disclosed. The 100% number is monitoring-technology coverage, not AI coverage or emissions reduction.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer-reported; no independent validation.", "source_ids": ["SW7-009"]}], "economics": "None attributed to AI.", "timeline": [{"date": "2025 (report; covers calendar 2024 unless noted)", "event": "AI program being deployed, according to the 2025 Sustainability Report; operational reporting is principally for 2024.", "source_ids": ["SW7-009"]}], "limitations": ["Deployment scope, model, alert precision/recall, response time and impact are undisclosed. Avoid conflating the continuous-monitoring work-order automation with AI decision support."], "assessment": "AI program analyzes SCADA and decision-support-center data and alerts teams to act on operational issues; described alongside continuous methane-monitoring data, dashboarding and a separate automated work-order process for detected leaks.", "sources": [{"title": "Devon Energy 2025 Sustainability Report", "date": "2025 (report; covers calendar 2024 unless noted)", "url": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "locator": "printed p. 15 / PDF p. 15, lines 1228-1261; full 90-page PDF read", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "SW7-009", "accessed": "2026-10-01"}], "source_ids": ["SW7-009"], "field_sources": {"workflow": ["SW7-009"], "technology_data": ["SW7-009"], "deployment_scope": ["SW7-009"], "practical_relevance": ["SW7-009"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer-reported; no independent validation."}, {"key": "swarm7-ai-250", "packet_id": "swarm7-ai-250", "record_ids": ["AI-250"], "existing_ids": [], "related_existing_ids": [], "company": "Devon Energy", "title": "Hourly computer-vision checks of flare equipment function", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated. Wellsite-camera workflow; quantitative scale not disclosed.", "practical_relevance": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "business_problem": "", "workflow": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "technology_data": "Company data-science analytics team developed custom AI models over repeated camera imagery. Specific vision architecture, camera counts, site count and deployment start are not reported.", "integration_human_role": "Human operator is alerted or dispatched to act; no autonomous repair or control claim.", "deployment_scope": "Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated. Wellsite-camera workflow; quantitative scale not disclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No performance or emissions metric attributed to this model. Devon separately reports flaring intensity but does not attribute that outcome to AI.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer-reported; no independent validation.", "source_ids": ["SW7-009"]}], "economics": "None reported.", "timeline": [{"date": "2025 (report; covers calendar 2024 unless noted)", "event": "Described as an operating capability in Devon's 2025 sustainability report; exact go-live date and current deployment breadth are not stated.", "source_ids": ["SW7-009"]}], "limitations": ["No model accuracy, false-alert rate, equipment coverage, avoided emissions, uptime or financial attribution. General statement about capabilities, not proof of quantified impact."], "assessment": "Custom AI models evaluate multiple wellsite camera images every hour for proper flare-equipment operation; a detected issue can trigger an operator alert or dispatch.", "sources": [{"title": "Devon Energy 2025 Sustainability Report", "date": "2025 (report; covers calendar 2024 unless noted)", "url": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "locator": "printed p. 9 / PDF p. 9, lines 713-738; full 90-page PDF read", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "SW7-009", "accessed": "2026-10-01"}], "source_ids": ["SW7-009"], "field_sources": {"workflow": ["SW7-009"], "technology_data": ["SW7-009"], "deployment_scope": ["SW7-009"], "practical_relevance": ["SW7-009"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer-reported; no independent validation."}, {"key": "swarm7-ai-251", "packet_id": "swarm7-ai-251", "record_ids": ["AI-251"], "existing_ids": [], "related_existing_ids": [], "company": "Diamondback Energy", "title": "Veriforce machine-learning contractor incident-risk scoring", "department": "HSE & emissions", "secondary_departments": [], "category": "Internal adoption", "stage": "Continuing operating partnership during 2024; described in Diamondback's 2025 CSR. Contractors performing services for production operations; count and coverage percentage not stated.", "practical_relevance": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "business_problem": "", "workflow": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "technology_data": "Vendor Veriforce ML and predictive modeling applied to contractor safety data. Diamondback does not disclose features, algorithm, score thresholds, volumes or validation results.", "integration_human_role": "Diamondback uses scores to manage contractor risk and eligibility. The precise review, appeal or human override process is not disclosed.", "deployment_scope": "Continuing operating partnership during 2024; described in Diamondback's 2025 CSR. Contractors performing services for production operations; count and coverage percentage not stated.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No Diamondback-specific incidents avoided or safety improvement attributed to the model. Its corporate safety statistics are not attributed to Veriforce scoring.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer-reported; no independent validation.", "source_ids": ["SW7-010"]}], "economics": "No cost, savings or return disclosed.", "timeline": [{"date": "2025 (2024 performance reporting)", "event": "Continuing operating partnership during 2024; described in Diamondback's 2025 CSR.", "source_ids": ["SW7-010"]}], "limitations": ["Vendor product description can establish generic mechanism only; no Diamondback-specific performance or exact decision threshold. Do not infer the model's accuracy or causal safety benefit."], "assessment": "Contractors serving production operations upload relevant safety information quarterly; Veriforce scores it using machine learning and predictive modeling to estimate whether a contractor will have a recordable incident during the next three months. Low scores can make contractors ineligible to work for Diamondback.", "sources": [{"title": "Diamondback Energy 2025 Corporate Sustainability Report", "date": "2025 (2024 performance reporting)", "url": "https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418", "locator": "printed p. 20 / PDF p. 20, lines 1422-1438; full 51-page PDF read", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "SW7-010", "accessed": "2026-10-01"}], "source_ids": ["SW7-010"], "field_sources": {"workflow": ["SW7-010"], "technology_data": ["SW7-010"], "deployment_scope": ["SW7-010"], "practical_relevance": ["SW7-010"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer-reported; no independent validation."}, {"key": "swarm7-ai-045", "packet_id": "swarm7-ai-045", "record_ids": ["AI-045"], "existing_ids": ["AI-045"], "related_existing_ids": [], "company": "Diamondback Energy", "title": "Machine learning / AI for artificial-lift choice and late-life well optimization", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives. Late-life producing wells and artificial-lift processes; scale not disclosed.", "practical_relevance": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "business_problem": "", "workflow": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "technology_data": "Issuer letter confirms ML and AI and the target workflow, but does not specify data inputs, models, automation/control mode, vendor, well count or roll-out boundaries.", "integration_human_role": "Production operations teams are described as driving the automation; no approval/override mechanism disclosed.", "deployment_scope": "Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives. Late-life producing wells and artificial-lift processes; scale not disclosed.", "partners": [], "disclosed_results": [{"metric": "Disclosed evidence", "value": "No result specifically attributed to this AI workflow. Nov 2025 letter states the intended gains; May 2026 letter says expected future uptime and LOE improvements. Reported corporate production and LOE are not causally linked to AI.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Issuer-reported; no independent validation.", "source_ids": ["SW7-011", "SW7-012"]}], "economics": "Intended to optimize lift cost and ultimately lower LOE; no AI-specific amount, achieved savings or payback reported.", "timeline": [{"date": "2025-11-03", "event": "Under development/being driven in November 2025; later May 2026 letter says field organization is currently scaling automation and data-driven predictive maintenance, including AI initiatives.", "source_ids": ["SW7-011", "SW7-012"]}], "limitations": ["Qualitative issuer disclosure only; implementation level and actual performance unquantified. May 2026 release explicitly frames uptime/LOE as expected in coming years."], "assessment": "Production-operations teams are driving process automation using ML and AI to select/optimize lift type and cost, reduce downtime, and improve performance of wells near the tail of their production curves.", "sources": [{"title": "Diamondback Energy Letter to Stockholders (Q3 2025)", "date": "2025-11-03", "url": "https://ir.diamondbackenergy.com/node/17156/pdf", "locator": "PDF p. 1, lines 61-64; full 3-page PDF read", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "SW7-011", "accessed": "2026-10-01"}, {"title": "Diamondback Energy Letter to Stockholders (Q1 2026)", "date": "2026-05-04", "url": "https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11", "locator": "lines 72-77, especially line 74; full release read", "type": "Issuer disclosure", "review_status": "Original source opened; relevant passage read.", "id": "SW7-012", "accessed": "2026-10-01"}], "source_ids": ["SW7-011", "SW7-012"], "field_sources": {"workflow": ["SW7-011", "SW7-012"], "technology_data": ["SW7-011", "SW7-012"], "deployment_scope": ["SW7-011", "SW7-012"], "practical_relevance": ["SW7-011", "SW7-012"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Issuer-reported; no independent validation."}, {"key": "swarm7-ai-166", "packet_id": "swarm7-ai-166", "record_ids": ["AI-166"], "existing_ids": ["AI-166"], "related_existing_ids": [], "company": "Encino Energy", "title": "Invoice AI and Snowflake-backed production-data agent rollout", "department": "IT, data & enterprise transformation", "secondary_departments": [], "category": "Internal adoption", "stage": "Invoice workflow in use; production-data agents rolling out in April 2025.", "practical_relevance": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "business_problem": "", "workflow": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "technology_data": "SCADA data enters Snowflake, then SQL virtual tables and Dataverse. A beta Snowflake connector was planned to remove the SQL step; AI uses Microsoft components.", "integration_human_role": "Invoice reviewer checks confidence; approvals remain human.", "deployment_scope": "Invoice workflow in use; production-data agents rolling out in April 2025.", "partners": ["Snowflake", "Microsoft"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Customer reports hundreds of administrative hours saved annually for one employee by the invoice application; not disaggregated to AI extraction alone. No production-agent outcome measured.", "period": "Annual customer estimate; measurement method undisclosed.", "attribution": "Source-reported claims; no independent validation.", "source_ids": ["RS085"]}], "economics": "Reported administrative time capacity; no AI project cost or cash payback.", "timeline": [{"date": "2025-04-29", "event": "Microsoft story published 2025-04-29; production-agent rollout incomplete then. Invoice app is described as operating.", "source_ids": ["RS085"]}], "limitations": ["Vendor/customer account. Maintenance-app savings are ordinary workflow automation, not AI-attributable. Snowflake supplies data; this is not proof of Snowflake-native inference."], "assessment": "AI Builder extracts invoice fields; confidence scores support reviewer checks before approval routing. Separately, Encino was building Copilot Studio agents for production-data questions.", "sources": [{"title": "Encino Energy streamlines operations with Microsoft Power Platform", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "date": "2025-04-29 (visible on page)", "locator": "Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83.", "review_status": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study.", "type": "Vendor customer case", "id": "RS085", "accessed": "2026-10-01"}], "source_ids": ["RS085"], "field_sources": {"workflow": ["RS085"], "technology_data": ["RS085"], "deployment_scope": ["RS085"], "practical_relevance": ["RS085"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Source-reported claims; no independent validation."}, {"key": "swarm7-ai-195", "packet_id": "swarm7-ai-195", "record_ids": ["AI-195"], "existing_ids": ["AI-195"], "related_existing_ids": [], "company": "Expand Energy", "title": "Snowflake forecasting ML, drilling analytics and site alerts", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Current undated case describes forecasting/monitoring in use; dashboard in development.", "practical_relevance": "Snowpark supports ML-based type-curve forecasts; Snowflake hosts drilling ML and Cortex Analyst questions. SCADA/site data produces alerts for operations-center investigation.", "business_problem": "", "workflow": "Snowpark supports ML-based type-curve forecasts; Snowflake hosts drilling ML and Cortex Analyst questions. SCADA/site data produces alerts for operations-center investigation.", "technology_data": "Snowflake, Snowpark, Cortex Analyst and site digital twins. A fuller AI drilling dashboard remains under development.", "integration_human_role": "Site alerts go to operations center for investigation.", "deployment_scope": "Current undated case describes forecasting/monitoring in use; dashboard in development.", "partners": ["Snowflake"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Estimated 1,000+ labor hours saved annually in one business unit; data monitoring covers 3,700 sites. Neither proves production uplift. Separately, 40% less data-related M&A time is a platform result.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Source-reported claims; no independent validation.", "source_ids": ["SW7-013"]}], "economics": "Estimated labor capacity; no model cost or cash ROI.", "timeline": [{"date": "Undated supplemental case", "event": "Undated supplemental vendor source; retain earlier dated profile evidence.", "source_ids": ["SW7-013"]}], "limitations": ["Supplemental case has no visible date. Its metrics cannot inherit the existing profile’s March 2026 disclosure date; model accuracy and AI-only financial return undisclosed."], "assessment": "Snowpark supports ML-based type-curve forecasts; Snowflake hosts drilling ML and Cortex Analyst questions. SCADA/site data produces alerts for operations-center investigation.", "sources": [{"title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "date": "No publication/update date displayed; accessed 2026-10-01", "locator": "Full page opened/read. Profile and 3,700-site scale at lines 16-21, 48-52; type-curve ML and labor estimate at lines 73-80; drilling ML, Cortex questions, monitoring/digital twins and alert escalation in full page lines 85-101; future AI dashboard/ML roadmap lines 110-112.", "review_status": "Full official Snowflake customer case read; customer executive quotes present. Undated, vendor-authored, estimates/claims not independently verified.", "type": "Vendor customer case", "id": "SW7-013", "accessed": "2026-10-01"}], "source_ids": ["SW7-013"], "field_sources": {"workflow": ["SW7-013"], "technology_data": ["SW7-013"], "deployment_scope": ["SW7-013"], "practical_relevance": ["SW7-013"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Source-reported claims; no independent validation."}, {"key": "swarm7-ai-202", "packet_id": "swarm7-ai-202", "record_ids": ["AI-202"], "existing_ids": ["AI-202"], "related_existing_ids": [], "company": "ConocoPhillips", "title": "Montney AI surveillance, artificial lift and flow assurance", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Multiwell-pad deployment evaluated after four months; further pads reported, second Canadian asset under consideration.", "practical_relevance": "IOCaaS monitors Montney wells, gathering equipment and compressors for gas-lift opportunities, hydrate risk, liquid loading and equipment anomalies. Operators review alerts and act.", "business_problem": "", "workflow": "IOCaaS monitors Montney wells, gathering equipment and compressors for gas-lift opportunities, hydrate risk, liquid loading and equipment anomalies. Operators review alerts and act.", "technology_data": "OPX Ai uses SCADA/historian streams, edge analytics and cloud dashboards. Existing standardized infrastructure and reusable templates supported approximately 120-day setup.", "integration_human_role": "Field operators and engineers act on exceptions.", "deployment_scope": "Multiwell-pad deployment evaluated after four months; further pads reported, second Canadian asset under consideration.", "partners": ["OPX Ai"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "Vendor authors report the asset team calculated 3–4% production above forecast and ~5% lower LOE after four months; no hydrate outages in the evaluation. Approximately 6% future uplift is a projection.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Source-reported claims; no independent validation.", "source_ids": ["FU003"]}], "economics": "No implementation price, recurring cost or AI-only cash ROI.", "timeline": [{"date": "2026-04-01", "event": "Article 2026-04-01; four-month operating evaluation, exact dates absent.", "source_ids": ["FU003"]}], "limitations": ["Forecast construction, well denominator, comparison group and evaluation dates undisclosed. Supplier executives authored the case; no independent validation. Monitoring does not establish closed-loop control."], "assessment": "IOCaaS monitors Montney wells, gathering equipment and compressors for gas-lift opportunities, hydrate risk, liquid loading and equipment anomalies. Operators review alerts and act.", "sources": [{"title": "Case Study: Field Deployments of AI-Based IOCaaS Advancing Artificial Lift and Flow Assurance", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "date": "2026-04-01", "type": "Society of Petroleum Engineers / Journal of Petroleum Technology article; vendor-contributed case study, coauthored by OPX Ai executives.", "locator": "Montney case; architecture, implementation, first-winter results and authorship.", "review_status": "Original full article opened/read, including the Montney deployment, architecture, timeline, outcomes, and authorship. The source gives operator/team attributions but does not provide underlying data or independent validation.", "id": "FU003", "accessed": "2026-10-01"}], "source_ids": ["FU003"], "field_sources": {"workflow": ["FU003"], "technology_data": ["FU003"], "deployment_scope": ["FU003"], "practical_relevance": ["FU003"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Source-reported claims; no independent validation."}, {"key": "swarm7-ai-072", "packet_id": "swarm7-ai-072", "record_ids": ["AI-072"], "existing_ids": ["AI-072"], "related_existing_ids": [], "company": "ExxonMobil", "title": "Guyana Neuro and DrillOps directional-drilling automation", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Field deployment: more than 48 km of 3D well paths, with >93% of the reported operations autonomous.", "practical_relevance": "Neuro controls trajectory and downlink requests; DrillOps automates surface equipment and drilling tasks. Shore teams monitor operations.", "business_problem": "", "workflow": "Neuro controls trajectory and downlink requests; DrillOps automates surface equipment and drilling tasks. Shore teams monitor operations.", "technology_data": "SLB Neuro and DrillOps; vendor labels product workflows AI-driven, while the case specifies closed-loop functions. Model, sensor and fallback design are undisclosed.", "integration_human_role": "Remote onshore monitoring; autonomy limits undisclosed.", "deployment_scope": "Field deployment: more than 48 km of 3D well paths, with >93% of the reported operations autonomous.", "partners": ["SLB"], "disclosed_results": [{"metric": "Disclosed evidence", "value": "SLB reports 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "period": "Period as described in source; metric measurement period not separately disclosed.", "attribution": "Source-reported claims; no independent validation.", "source_ids": ["SW7-014"]}], "economics": "No dollar savings, avoided rig days, software cost or return.", "timeline": [{"date": "2026 (source URL; exact date unknown)", "event": "Supplemental SLB case carries 2026 in URL; detailed operating period unverified.", "source_ids": ["SW7-014"]}], "limitations": ["2026 URL path, no exact publication or campaign date. Vendor-reported results; 93% is the cited operating scope, not all Exxon drilling. No sample or counterfactual. Separate Halliburton campaign metrics must not be merged."], "assessment": "Neuro controls trajectory and downlink requests; DrillOps automates surface equipment and drilling tasks. Shore teams monitor operations.", "sources": [{"title": "ExxonMobil Guyana Limited leverages drilling automation to set new performance benchmarks in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "date": "2026 (URL path); no exact date displayed in page text", "type": "SLB vendor case study", "locator": "Case study final sections: Neuro/DrillOps functions, 48 km, autonomous share and downlinks.", "review_status": "Original full page opened/read; deployment, named products, functions, 48-km path, autonomy share, and downlink metrics appear in final case-study section. SLB is the supplier and reports the performance claims.", "id": "SW7-014", "accessed": "2026-10-01"}], "source_ids": ["SW7-014"], "field_sources": {"workflow": ["SW7-014"], "technology_data": ["SW7-014"], "deployment_scope": ["SW7-014"], "practical_relevance": ["SW7-014"]}, "verification": "Original source reviewed", "cohort": "Seventh gap review · October 2026", "report_date": "2026-10-01", "evidence_note": "Source-reported claims; no independent validation."}, {"key": "oct4-ai-006", "packet_id": "oct4-ai-006", "record_ids": ["AI-006"], "existing_ids": ["AI-006"], "related_existing_ids": [], "company": "Devon Energy", "title": "Autonomous artificial-lift optimization at operating wells", "department": "Production & field operations", "secondary_departments": [], "category": "Internal adoption", "stage": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "practical_relevance": "Operator reports 1,000 live closed-loop AI lift wells in August 2026. Prior pilot benefits do not establish fleet-wide results.", "business_problem": "", "workflow": "Closed-loop AI adjusts lift settings in real time, 24 hours a day. Devon reports 1,000 operating wells, including first legacy Coterra wells; more than 2,000 additional Permian wells are prospective deployment opportunities. The Q2 disclosure does not provide a lift-type breakdown.", "technology_data": "Closed-loop AI artificial-lift optimization; detailed model architecture and the lift-type mix are not disclosed in the Q2 update.", "integration_human_role": "The Q2 call describes autonomous 24/7 optimization. Human exception handling, override rates and control guardrails are not specified.", "deployment_scope": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "partners": [], "disclosed_results": [{"metric": "Live deployment", "value": "1,000 wells; >2,000 additional wells prospective", "period": "August 2026", "attribution": "Devon management", "source_ids": ["SW8-001"]}], "economics": "CEO separately linked AI value to drilling/completion time and lower capital costs, and cited business optimization savings; no amount is attributed specifically to AI or the lift tool.", "timeline": [{"date": "2026-08-05", "event": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "source_ids": ["SW8-001"]}], "limitations": ["Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale."], "assessment": "Latest issuer disclosure establishes deployment reach, not fleet-wide production benefit.", "sources": [{"id": "SW8-001", "title": "Devon Q2 2026 earnings call transcript", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-08-05", "type": "Issuer-hosted earnings call", "locator": "Printed/PDF pp. 5 and 17; CEO prepared remarks and CTO Q&A", "review_status": "Relevant passages reviewed", "accessed": "2026-10-04"}], "source_ids": ["SW8-001"], "field_sources": {"workflow": ["SW8-001"], "technology_data": ["SW8-001"], "deployment_scope": ["SW8-001"]}, "verification": "Original source reviewed", "cohort": "October 4 evidence review", "report_date": "2026-10-04", "evidence_note": "Operator-reported deployment; no independent validation."}, {"key": "oct4-ai-007", "existing_ids": ["AI-007"], "company": "Devon Energy", "title": "AI agents for drilling and completions monitoring", "department": "Drilling & completions", "secondary_departments": [], "category": "Internal adoption", "stage": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "practical_relevance": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "business_problem": "Prior-register claim (not reverified): D&C monitoring models may identify unproductive time and repeat effective practices.", "workflow": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "technology_data": "AI analysis of drilling/completion performance; engineers review alerts and take action. Detailed model implementation not disclosed.", "integration_human_role": "Teams review performance alerts and act on identified gaps. Autonomous execution is not established.", "deployment_scope": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "partners": [], "disclosed_results": [{"metric": "Operational workflow", "value": "Live performance analysis and alerts; no AI-specific cash return established", "period": "August 2026", "attribution": "Devon management", "source_ids": ["SW8-001"]}], "economics": "No AI-specific economics verified; broader business optimization savings cannot be attributed to this workflow.", "timeline": [{"date": "2026-08-05", "event": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "source_ids": ["SW8-001"]}], "limitations": ["No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only."], "assessment": "The Q2 call supports a live analytics workflow; no discrete AI-only financial outcome is isolated.", "sources": [{"id": "SW8-001", "title": "Devon Q2 2026 earnings call transcript", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-08-05", "type": "Issuer-hosted earnings call", "locator": "Printed/PDF pp. 5 and 17; CEO prepared remarks and CTO Q&A", "review_status": "Relevant passages reviewed", "accessed": "2026-10-04"}], "field_sources": {"workflow": ["SW8-001"], "technology_data": ["SW8-001"], "deployment_scope": ["SW8-001"]}, "verification": "Original source reviewed", "corrections_to_existing": "AI-007 is a distinct D&C workflow and should not be merged into lift optimization. It remains unverified pending review of Devon’s presentation.", "related_existing_ids": [], "cohort": "October 4 evidence review", "packet_id": "oct4-ai-007", "source_ids": ["SW8-001"], "report_date": "2026-10-04", "evidence_note": "Operator-reported deployment; no independent validation.", "record_ids": ["AI-007"]}, {"key": "oct4-ai-210", "packet_id": "oct4-ai-210", "record_ids": ["AI-210"], "existing_ids": [], "related_existing_ids": [], "company": "Diversified Energy Company (DEC)", "title": "OneDEC: producer-statement normalization and governed finance/operations AI foundation", "department": "Finance, investment & strategy", "secondary_departments": [], "category": "Internal adoption", "stage": "Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.", "practical_relevance": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "business_problem": "", "workflow": "New company-authored detail for the existing OneDEC profile: statement automation digitizes and normalizes producer data across all sales points to reduce re-keying for Production Accounting, Marketing and Revenue Distribution; a centralized contract repository is described as the authoritative source for commercial terms used in pricing and settlements. The report also describes a single historian supporting manage-by-exception surveillance. This is an enrichment of the existing work, not a new duplicate record.", "technology_data": "OneDEC is described as a cloud-based data-centric operating environment. Microsoft Cloud provides identity/security/monitoring; operational and financial systems feed Snowflake through governed data contracts; ServiceNow enterprise playbooks provide approval/evidence workflows; APIs support portable model/agent use without data duplication. RAG agents use governed company data. Prompts, actions and outputs are versioned and monitored; logging through ServiceNow GRC supports SOX-aligned change control. The report does not specify document AI/OCR model vendors or whether producer-statement digitization itself uses generative AI.", "integration_human_role": "", "deployment_scope": "Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.", "partners": [], "disclosed_results": [{"metric": "Reported evidence", "value": "No workflow-level time, accuracy, volume, settlement, cash, or ROI metric disclosed.", "period": "2026-05-28", "attribution": "Source-reported; see evidence limits", "source_ids": ["M29-007"]}], "economics": "No quantified AI/statement-automation savings, costs or payback. ROI is an evaluation gate for each AI use case, not a reported result. Separate company-wide financial performance should not be attributed to this program.", "timeline": [{"date": "2026-05-28", "event": "Company-reported capabilities; workflow-specific go-live dates and adoption scale undisclosed.", "source_ids": ["M29-007"]}], "limitations": ["Source is a company-authored annual sustainability report rather than a technical deployment postmortem. It reports environmental/operational data for 2025 but publishes in 2026. The workflow list sits under 'AI Efficiency at Scale,' yet no specific model is linked to statement automation or the contract repository. Architecture is company-reported; no independent validation of automation benefits. Financial controls are unusually specific, but no actual JIB or AFE workflow is claimed."], "assessment": "Company and full research packets search did not identify a DEC profile; this is a primary operator disclosure of specifically named financial and commercial workstreams, beyond generic AI policy or product marketing.", "sources": [{"title": "Diversified Energy Company Demonstrates Proven Sustainability Leadership (2025 Sustainability Report release)", "url": "https://ir.div.energy/news-events/us-press-releases/detail/225/diversified-energy-company-demonstrates-proven-sustainability-leadership", "date": "2026-05-28 (issuer release announcing 2025 report); IR report index separately lists June 17, 2026", "locator": "Official issuer release confirms report publication; report content fully opened/read in linked PDF. OneDEC/Snowflake/ServiceNow architecture: report pp. 62-63; data and controls, HITL, audit, AI governance: p. 63; statement automation, historian surveillance and pricing/settlements: p. 63. Report data period is calendar 2025 (p. 67).", "review_status": "Full issuer-authored report PDF opened/read and cross-checked against official IR release and report index. Metric year (2025) and publication date (2026) kept separate.", "id": "SW8-002", "type": "Issuer report", "accessed": "2026-10-04"}, {"title": "Diversified 2025 Sustainability Report", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "date": "Announced 2026-05-28; report index lists 2026-06-17", "locator": "Full 85-page issuer PDF read; especially pp. 62-63 lines 4044-4217 and reporting-period notes p. 67 lines 4256-4274.", "review_status": "Full primary issuer document reviewed. No dated workflow-specific operational metric disclosed.", "id": "M29-007", "type": "Issuer report", "accessed": "2026-10-04"}], "source_ids": ["SW8-002", "M29-007"], "field_sources": {"workflow": ["SW8-002", "M29-007"], "technology_data": ["SW8-002", "M29-007"], "stage": ["SW8-002", "M29-007"], "economics": ["SW8-002", "M29-007"]}, "verification": "verified primary", "cohort": "October 4 evidence review", "report_date": "2026-10-04", "evidence_note": "Source review establishes the disclosure, not independently validated performance."}, {"title": "Presidio separates field recommendations from controller adjustments", "stage": "Existing operating profile; Q2 workflow detail clarified", "workflow": "Live 24/7 well surveillance and natural-language production queries support field operators and engineers. DOUG sends recommendations to field engineers; weekend routing provides pumpers a prioritized map. Twenty-four plunger boxes continuously adjust cycle timing; safeguards and overrides are undisclosed. Budget, acquisition-screening and fleet-orchestration agents remain 2026 deployment work, not established full-fleet autonomy.", "technology_data": "DOUG alerts field engineers; pumpers receive a prioritized map. Controllers on 24 wells continuously adjust cycle timing. Management says field judgment remains with staff; controller safeguards are unspecified.", "integration_human_role": "DOUG advises engineers; pumpers use a map; controller timing adjustments are automatic, with safeguards undisclosed.", "deployment_scope": "24 controller wells within Presidio’s operated base. Call: 12 August 2026; controller pre/post periods and weekend baseline unknown. Platform scope is a separate population.", "practical_relevance": "What happened on comparable wells, and what net benefit remained after controller and field-work costs?", "business_problem": "Surveillance and field-work priorities on mature producing assets.", "economics": "Operator reports 2.3% AI-linked production uplift through Q2: 1.4 points from DOUG and nearly 1 point from adjacent initiatives. DOUG averaged ~400 boe/d over the prior three months. Controllers on 24 wells: 2,680 to 2,946 Mcf/day, roughly 10% on that population. Weekend production reportedly rose 2.5%, with comparison baseline undisclosed. These overlap the program total; do not add them. $4.5m annualized revenue covers the initiatives, not net cash or DOUG alone. August 17 deck repeats 2.3% YTD, targets 3–5%, and cites ~2,300 platform wells on its GOAL slide. $2.5m per uplift point is illustrative revenue sensitivity.", "limitations": ["Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs. Controller pre/post periods, matched wells, decline adjustments, weekend baseline and controller safeguards remain unspecified. The 24 controller wells are a separate population from platform coverage."], "assessment": "The issuer connects field surveillance, staff knowledge and operating recommendations. The wider agent program and production target require follow-through; comparable net returns remain undisclosed.", "packet_id": "oct5-followup-ai-231", "key": "oct5-followup-ai-231", "record_ids": ["AI-231"], "company": "Presidio Production", "category": "Internal adoption", "department": "Production & field operations", "source_ids": ["OCT5-S024"], "report_date": "2026-10-18", "verification": "verified primary", "field_sources": {"business_problem": ["OCT5-S024"], "workflow": ["OCT5-S024"], "technology_data": ["OCT5-S024"], "deployment_scope": ["OCT5-S024"], "economics": ["OCT5-S024"], "integration_human_role": ["OCT5-S024"], "practical_relevance": ["OCT5-S024"]}, "disclosed_results": [{"metric": "Q2 2026 company report", "value": "Operator reports 2.3% AI-linked production uplift through Q2: 1.4 points from DOUG and nearly 1 point from adjacent initiatives. DOUG averaged ~400 boe/d incremental production over the prior three months. $4.5m annualized revenue spans the initiatives; it is not net cash or a DOUG-only result.", "period": "2026-08-12", "attribution": "Operator-reported, not independently evaluated", "source_ids": ["OCT5-S024"]}], "timeline": [{"date": "2026-08-12", "event": "Q2 update replaces the earlier progress figure; prior evidence remains in the following dossier and disclosure history.", "source_ids": ["OCT5-S024"]}, {"date": "2026-08-12", "review_on": "2026-10-18", "event": "Presidio’s Q2 call distinguishes DOUG recommendations, weekend visit routing and automatic plunger-cycle adjustments. Reported gas rates on the controller population: 2,680 to 2,946 Mcf/day, roughly 10%. Weekend production reportedly rose 2.5%. These overlap the existing 2.3% program-uplift claim and must not be added to it.", "source_ids": ["OCT5-S024"]}], "sources": [{"id": "OCT5-S024", "title": "Presidio Production Company Q2 2026 Earnings Call Transcript", "url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "date": "2026-08-12", "publisher": "Presidio Production", "locator": "Chris Hammack remarks, printed page 3: weekend coverage, 24 plunger boxes, human judgment; Will Ulrich page 2 for aggregate uplift/revenue", "excerpt": "None of these replace field judgment", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported.", "review_status": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "earlier_review": {"report_date": "2026-10-05", "assessment": "A specific path from data to field action is disclosed. Revenue attribution is useful, but missing costs prevent a project return calculation.", "limitations": ["Company-reported attribution; no independent counterfactual, measurement protocol or project costs. The $4.5m annualized revenue spans AI initiatives and is not realized net cash."], "note": "Superseded by targeted 18 October 2026 source check."}, "last_source_check": "2026-10-18"}, {"packet_id": "oct6-followup-ai-086", "review_date": "2026-10-06", "title": "6 October source review: Devon Energy", "workflow": "The interview describes AI-assisted simulfrac sequencing and human approval of recommendations. Further video-AI development was temporarily paused to prioritize other AI work, with a 2026 restart intended. Its compressor example uses physics-based optimization; the article does not explicitly identify that optimizer as AI.", "deployment_scope": "Internal tools in use as described in the July 2025 CTO interview; broader video-AI work was planned.", "economics": "Full implementation and recurring costs are not disclosed.", "result_label": "Reported result and attribution limits", "attribution": "Secondary interview", "verification": "secondary only", "limitations": ["Trade-press interview, not an issuer filing. The compressor result cannot establish AI uplift. The video-development pause reflects reprioritization; it does not establish failed deployment. Restart timing, measurement methods and costs are unconfirmed.", "Historical interview packet: does not establish issuer-verified AI-101 or 15–30% productivity. See watch21extra-ai-086 for the separate primary daily-use evidence."], "timeline_event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_review": "Full trade-press interview rechecked 6 October 2026", "timeline_date": "2026-10-06", "stage": "Internal tools in use as described in the July 2025 CTO interview; broader video-AI work was planned.", "source_locator": "Proprietary AI Platform section, productivity, compressor, simulfrac and video-development paragraphs", "key": "oct6-followup-ai-086", "record_ids": ["AI-086"], "company": "Devon Energy", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "source_ids": ["R039"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["R039"], "workflow": ["R039"], "technology_data": ["R039"], "deployment_scope": ["R039"], "practical_relevance": ["R039"], "economics": ["R039"], "integration_human_role": ["R039"]}, "disclosed_results": [{"metric": "Reported result and attribution limits", "value": "Reported 15–30% daily-task productivity gain and ~1,000 AI-101 trainees. A separate 5% compressor-site oil gain is not explicitly attributed to AI.", "period": "2025-07", "attribution": "Secondary interview", "source_ids": ["R039"]}], "timeline": [{"date": "2026-10-06", "event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_ids": ["R039"]}], "sources": [{"id": "R039", "title": "ChatDVN internal AI platform plus AI-101 training", "url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "date": "2025-07", "type": "American Oil & Gas Reporter", "publisher": "Devon Energy", "locator": "Proprietary AI Platform section, productivity, compressor, simulfrac and video-development paragraphs", "review_status": "Full trade-press interview rechecked 6 October 2026"}]}, {"packet_id": "oct6-followup-ai-201", "review_date": "2026-10-06", "title": "6 October source review: Chevron", "workflow": "IOCaaS connects SCADA and historian data to surveillance and bounded lift adjustments. The vendor describes roughly 12 months of phased integration, including tag mapping and historical-data cleaning; fully live by early 2022.", "deployment_scope": "Deployment began 2020; live by early 2022; retrospective disclosure April 2026", "economics": "Full implementation and recurring costs are not disclosed.", "result_label": "Reported result and attribution limits", "attribution": "Vendor-authored retrospective", "verification": "verified partner", "limitations": ["Vendor retrospective disclosed in April 2026 about a deployment begun in 2020. Facility changes and commodity prices also affected economics. Integration labor, service fees, recurring costs and a net-cash comparison are missing."], "timeline_event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_review": "Full vendor-authored JPT article rechecked 6 October 2026", "timeline_date": "2026-10-06", "stage": "Deployment began 2020; live by early 2022; retrospective disclosure April 2026", "source_locator": "Kaybob and Montney deployment sections; integration and measurement-period paragraphs", "key": "oct6-followup-ai-201", "record_ids": ["AI-201"], "company": "Chevron", "category": "Internal adoption", "department": "Production & field operations", "source_ids": ["FU003"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["FU003"], "workflow": ["FU003"], "technology_data": ["FU003"], "deployment_scope": ["FU003"], "practical_relevance": ["FU003"], "economics": ["FU003"], "integration_human_role": ["FU003"]}, "disclosed_results": [{"metric": "Reported result and attribution limits", "value": "5% lower LOE; BOE output about 6% above original projection after normalizing for new-drill wells, in first year.", "period": "2026-04-01", "attribution": "Vendor-authored retrospective", "source_ids": ["FU003"]}], "timeline": [{"date": "2026-10-06", "event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_ids": ["FU003"]}], "sources": [{"id": "FU003", "title": "Field deployments of AI-based IOCaaS", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "date": "2026-04-01", "type": "Vendor-authored technical case article", "publisher": "SPE / JPT; OPX Ai authors", "locator": "Kaybob and Montney deployment sections; integration and measurement-period paragraphs", "review_status": "Full vendor-authored JPT article rechecked 6 October 2026"}]}, {"packet_id": "oct6-followup-ai-202", "review_date": "2026-10-06", "title": "6 October source review: ConocoPhillips", "workflow": "IOCaaS monitors lift, hydrate risk and compressors. The vendor reports four months from kickoff to full deployment, helped by standardized SCADA/historians and reused integration templates; operating results cover a further four months.", "deployment_scope": "Pilot. The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "economics": "Full implementation and recurring costs are not disclosed.", "result_label": "Reported result and attribution limits", "attribution": "Vendor-authored retrospective", "verification": "verified partner", "limitations": ["Vendor-authored four-month operating comparison. 6% uplift was prospective, not observed. No new hardware does not mean zero cost; service fees, labor and net cash are undisclosed."], "timeline_event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_review": "Full vendor-authored JPT article rechecked 6 October 2026", "timeline_date": "2026-10-06", "stage": "Pilot. The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "source_locator": "Kaybob and Montney deployment sections; integration and measurement-period paragraphs", "key": "oct6-followup-ai-202", "record_ids": ["AI-202"], "company": "ConocoPhillips", "category": "Internal adoption", "department": "Production & field operations", "source_ids": ["FU003"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["FU003"], "workflow": ["FU003"], "technology_data": ["FU003"], "deployment_scope": ["FU003"], "practical_relevance": ["FU003"], "economics": ["FU003"], "integration_human_role": ["FU003"]}, "disclosed_results": [{"metric": "Reported result and attribution limits", "value": "3–4% production above forecast after four months of operation; about 5% lower LOE.", "period": "2026-04-01", "attribution": "Vendor-authored retrospective", "source_ids": ["FU003"]}], "timeline": [{"date": "2026-10-06", "event": "Source revisited on 6 October 2026; original disclosure date retained. Earlier dossier remains below.", "source_ids": ["FU003"]}], "sources": [{"id": "FU003", "title": "Field deployments of AI-based IOCaaS", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "date": "2026-04-01", "type": "Vendor-authored technical case article", "publisher": "SPE / JPT; OPX Ai authors", "locator": "Kaybob and Montney deployment sections; integration and measurement-period paragraphs", "review_status": "Full vendor-authored JPT article rechecked 6 October 2026"}]}, {"packet_id": "oct6-gap-ai-054", "review_date": "2026-10-06", "title": "QB Energy geospatial field planning with Quantum", "workflow": "QB Energy, identified in Quantum’s 2026 report, uses an engineering-reviewed planning tool. Azure Databricks with Apache Sedona processes terrain; an integer-linear-programming solver selects pads and wells against more than 25 constraints. A Copilot Studio agent explains results. Azure Blob Storage, Dataverse, App Service and Azure OpenAI are named in the stack.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Vendor/customer account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["Microsoft comparisons remain vendor-reported. Quantum names QB Energy and reports 12 areas over nine months. The ~$10bn development program informed is not savings or value created. Engineers review plans; no autonomous capital approval is established. Solver attribution: Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model."], "verification": "verified partner", "source_review": "Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation.", "source_locator": "Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph", "timeline_date": "2026-04-30", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-054", "record_ids": ["AI-054"], "company": "Quantum Capital Group", "category": "Internal adoption", "department": "Reservoir engineering & field development", "source_ids": ["OCT5-S032", "OCT9-QCG-2026"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S032"], "workflow": ["OCT5-S032", "OCT9-QCG-2026"], "technology_data": ["OCT5-S032"], "deployment_scope": ["OCT5-S032"], "practical_relevance": ["OCT5-S032"], "economics": ["OCT5-S032"], "integration_human_role": ["OCT5-S032"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "Microsoft reports a 20,000-acre scenario took 20 minutes versus three weeks; leasehold utilization rose from 50% to 85%. Production runs evaluated 2.3m pad candidates, 3.5m well possibilities and over 50m allocations; the accumulation period is unstated.", "period": "2026-04-30", "attribution": "Vendor/customer account", "source_ids": ["OCT5-S032"]}, {"metric": "Updated solver-version comparison", "value": "50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.", "period": "Updated version; evaluation period unknown", "attribution": "Microsoft customer story; algorithm, not LLM result", "source_ids": ["OCT5-S032"]}], "timeline": [{"date": "2026-04-30", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["OCT5-S032"]}], "sources": [{"id": "OCT5-S032", "title": "Field development planning at Quantum Capital Group", "url": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "date": "2026-04-30", "type": "Microsoft customer story", "publisher": "Microsoft", "locator": "Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph", "review_status": "Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation."}, {"id": "OCT9-QCG-2026", "title": "Quantum 2026 Stakeholder Report", "url": "https://www.quantumcap.com/wp-content/uploads/2026/08/QCG-2026-ESG-Stakeholder-Report.pdf", "date": "2026", "type": "Sponsor stakeholder report", "publisher": "Quantum Capital Group", "locator": "Pages 11–15; page 68 disclaimer", "review_status": "Relevant primary sections reviewed"}]}, {"packet_id": "oct6-gap-ai-088", "review_date": "2026-10-06", "title": "Proprietary generative AI for collaboration and operational data", "workflow": "EOG says it used proprietary generative-AI tools in 2025 to support collaboration, automate data capture and provide real-time operating insights. The annual report does not identify the model, users, controls or specific functional rollout.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Operator account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage."], "verification": "verified primary", "source_review": "Full original document reviewed on 6 October 2026", "source_locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "timeline_date": "2026 annual report covering 2025; exact publication day not confirmed", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-088", "record_ids": ["AI-088"], "company": "EOG Resources", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "source_ids": ["OCT5-S033"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S033"], "workflow": ["OCT5-S033"], "technology_data": ["OCT5-S033"], "deployment_scope": ["OCT5-S033"], "practical_relevance": ["OCT5-S033"], "economics": ["OCT5-S033"], "integration_human_role": ["OCT5-S033"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "No AI-specific outcome or rollout count disclosed.", "period": "2026 annual report covering 2025; exact publication day not confirmed", "attribution": "Operator account", "source_ids": ["OCT5-S033"]}], "timeline": [{"date": "2026 annual report covering 2025; exact publication day not confirmed", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["OCT5-S033"]}], "sources": [{"id": "OCT5-S033", "title": "Proprietary generative AI for collaboration and operational data", "url": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "date": "2026 annual report covering 2025; exact publication day not confirmed", "type": "Operator annual report", "publisher": "EOG Resources", "locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "review_status": "Full original document reviewed on 6 October 2026"}]}, {"packet_id": "oct6-gap-ai-089", "review_date": "2026-10-06", "title": "Proprietary machine-learning production optimizer", "workflow": "EOG says its proprietary production optimizer and machine-learning capabilities optimized base production and runtime in 2025. The annual report does not describe the model, number of wells or human-control boundary.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Operator account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["Full issuer-hosted annual report reviewed. Its 7% lower average well cost concerns wider operational achievements; it is not an AI-attributed result. The earlier transcript’s exact geology, safety, gas-lift and remote-Utica claims remain uncorroborated by this passage."], "verification": "verified primary", "source_review": "Full original document reviewed on 6 October 2026", "source_locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "timeline_date": "2026 annual report covering 2025; exact publication day not confirmed", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-089", "record_ids": ["AI-089"], "company": "EOG Resources", "category": "Internal adoption", "department": "Production & field operations", "source_ids": ["OCT5-S033"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S033"], "workflow": ["OCT5-S033"], "technology_data": ["OCT5-S033"], "deployment_scope": ["OCT5-S033"], "practical_relevance": ["OCT5-S033"], "economics": ["OCT5-S033"], "integration_human_role": ["OCT5-S033"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "No AI-specific outcome or rollout count disclosed.", "period": "2026 annual report covering 2025; exact publication day not confirmed", "attribution": "Operator account", "source_ids": ["OCT5-S033"]}], "timeline": [{"date": "2026 annual report covering 2025; exact publication day not confirmed", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["OCT5-S033"]}], "sources": [{"id": "OCT5-S033", "title": "Proprietary generative AI for collaboration and operational data", "url": "https://filecache.investorroom.com/mr5ir_eogresources2/462/2025_EOG_Annual_Report.pdf", "date": "2026 annual report covering 2025; exact publication day not confirmed", "type": "Operator annual report", "publisher": "EOG Resources", "locator": "PDF page 4, shareholder letter: Consistent Operational Excellence Compounds Over Time", "review_status": "Full original document reviewed on 6 October 2026"}]}, {"packet_id": "oct6-gap-ai-092", "review_date": "2026-10-06", "title": "Proprietary AI for drilling and centralized operations", "workflow": "Ovintiv’s filed 2025 report describes proprietary AI in drilling, fracturing and the centralized operations control center. Real-time field data supports drilling precision and mechanical-energy use; predictive modeling targets downtime. It qualitatively attributes lower costs and faster drilling cycles to this work.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Operator account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage."], "verification": "verified primary", "source_review": "Full original document reviewed on 6 October 2026", "source_locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "timeline_date": "2026-02-23", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-092", "record_ids": ["AI-092"], "company": "Ovintiv", "category": "Internal adoption", "department": "Drilling & completions", "source_ids": ["OCT5-S035"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S035"], "workflow": ["OCT5-S035"], "technology_data": ["OCT5-S035"], "deployment_scope": ["OCT5-S035"], "practical_relevance": ["OCT5-S035"], "economics": ["OCT5-S035"], "integration_human_role": ["OCT5-S035"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "No quantified AI-specific outcome disclosed.", "period": "2026-02-23", "attribution": "Operator account", "source_ids": ["OCT5-S035"]}], "timeline": [{"date": "2026-02-23", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["OCT5-S035"]}], "sources": [{"id": "OCT5-S035", "title": "Proprietary AI for drilling and centralized operations", "url": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "date": "2026-02-23", "type": "SEC-filed operator annual report", "publisher": "Ovintiv", "locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "review_status": "Full original document reviewed on 6 October 2026"}]}, {"packet_id": "oct6-gap-ai-094", "review_date": "2026-10-06", "title": "Real-time AI optimization of fracturing", "workflow": "Ovintiv’s filed 2025 report describes AI-assisted optimization of downhole hydraulics and sand ramps, avoiding wasted fracturing energy and limiting frac-hit magnitude and frequency on existing wells. It attributes lower costs and faster completions qualitatively.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Operator account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["The filed text confirms the workflow, not a measured return. No rollout denominator, test method, project cost or isolated benefit is provided. The earlier 70-input figure and acquired-rig DRIVE/artificial-lift specifics are secondary accounts, not confirmed by this passage."], "verification": "verified primary", "source_review": "Full original document reviewed on 6 October 2026", "source_locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "timeline_date": "2026-02-23", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-094", "record_ids": ["AI-094"], "company": "Ovintiv", "category": "Internal adoption", "department": "Drilling & completions", "source_ids": ["OCT5-S035"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S035"], "workflow": ["OCT5-S035"], "technology_data": ["OCT5-S035"], "deployment_scope": ["OCT5-S035"], "practical_relevance": ["OCT5-S035"], "economics": ["OCT5-S035"], "integration_human_role": ["OCT5-S035"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "No quantified AI-specific outcome disclosed.", "period": "2026-02-23", "attribution": "Operator account", "source_ids": ["OCT5-S035"]}], "timeline": [{"date": "2026-02-23", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["OCT5-S035"]}], "sources": [{"id": "OCT5-S035", "title": "Proprietary AI for drilling and centralized operations", "url": "https://www.sec.gov/Archives/edgar/data/1792580/000119312526064309/ovv-20251231.htm", "date": "2026-02-23", "type": "SEC-filed operator annual report", "publisher": "Ovintiv", "locator": "2025 Form 10-K, Business and Properties, Montney section, printed page 13", "review_status": "Full original document reviewed on 6 October 2026"}]}, {"packet_id": "oct6-gap-ai-171", "review_date": "2026-10-06", "title": "Pipeline IQ engineering-document verification agents", "workflow": "PIPER cross-checks pipeline material-test reports, weld maps and pressure-test records and gives source citations for an engineer’s final review. AWS names Bedrock, AgentCore, Textract, Step Functions, Claude and Strands Agents. SCOUT provides natural-language document search; equivalent rollout maturity is not established.", "deployment_scope": "Operating use described by the source; launch date and full rollout scope not established.", "result_label": "Reported result and measurement scope", "attribution": "Vendor/customer account", "economics": "Full implementation costs and realized net cash are not disclosed.", "limitations": ["Undated page: launch and measurement dates are unknown. The test sample and accuracy definition are unstated. Engineer review remains required. Expected capital avoidance is not achieved savings; full costs and realized cash are absent."], "verification": "verified partner", "source_review": "Full original document reviewed on 6 October 2026", "source_locator": "PIPER, SCOUT and outcome sections", "timeline_date": "Undated source; see chronology", "timeline_event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "stage": "Operating use described by the source; launch date and full rollout scope not established.", "key": "oct6-gap-ai-171", "record_ids": ["AI-171"], "company": "TC Energy", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "source_ids": ["RS095"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["RS095"], "workflow": ["RS095"], "technology_data": ["RS095"], "deployment_scope": ["RS095"], "practical_relevance": ["RS095"], "economics": ["RS095"], "integration_human_role": ["RS095"]}, "disclosed_results": [{"metric": "Reported result and measurement scope", "value": "AWS reports 85% less document-review time and more than 100 PIPER users. Early tests on documents up to 120 pages reported 95–100% accuracy. Over $18.7m in maintenance capital avoidance is expected over five years for PIPER and SCOUT combined; it is a forecast.", "period": "undated", "attribution": "Vendor/customer account", "source_ids": ["RS095"]}], "timeline": [{"date": "Undated source; see chronology", "event": "New source reviewed on 6 October 2026. Source date records the disclosure, not the system launch; earlier evidence remains in the history.", "source_ids": ["RS095"]}], "sources": [{"id": "RS095", "title": "TC Energy Pipeline IQ on AWS", "url": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "date": "undated", "type": "partner case study", "publisher": "Amazon Web Services", "locator": "PIPER, SCOUT and outcome sections", "review_status": "Full original document reviewed on 6 October 2026"}]}, {"packet_id": "oct6-three-rounds-ai-146", "review_date": "2026-10-06", "title": "P-52 flare-reduction application: platform scope added", "workflow": "Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.", "technology_data": "Petrobras reports applying AI to reduce flare burning at P-52 in 2025. It gives no platform-specific result or launch date. Roncador/Campos location is supported by a separate dated Petrobras field source; company-level flare metrics remain separate.", "deployment_scope": "AI applied to P-52 in 2025; no launch month or P-52-specific result. Separate Petrobras field source locates P-52 at Roncador in Campos Basin.", "result_label": "Supplemental source: scope and limits", "attribution": "Operator-reported", "economics": "No project cost and measured cash-benefit pair disclosed.", "limitations": ["P-52 is newly identified scope, not proof that the strategy-deck 10,000 tCO2/year result belongs to P-52 or Smart Tocha."], "verification": "verified primary", "source_review": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "source_locator": "2025 innovation highlights, physical/printed p. 133", "timeline_date": "2026-05", "timeline_event": "P-52 flare-reduction application: platform scope added", "stage": "Annualized emissions result reported; Smart Tocha’s full rollout is not established by this slide.", "key": "oct6-three-rounds-ai-146", "record_ids": ["AI-146"], "company": "Petrobras", "category": "Internal adoption", "department": "Refining, LNG & process operations", "source_ids": ["OCT5-S040"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S040"], "workflow": ["OCT5-S040"], "technology_data": ["OCT5-S040"], "deployment_scope": ["OCT5-S040"], "practical_relevance": ["OCT5-S040"], "economics": ["OCT5-S040"], "integration_human_role": ["OCT5-S040"]}, "disclosed_results": [{"metric": "Supplemental source: scope and limits", "value": "No P-52-specific outcome reported. The existing 10,000 tCO2/year deck claim remains at company/program scope.", "period": "2026-05", "attribution": "Operator-reported", "source_ids": ["OCT5-S040"]}], "timeline": [{"date": "2026-05", "event": "P-52 flare-reduction application: platform scope added", "source_ids": ["OCT5-S040"]}], "sources": [{"id": "OCT5-S040", "title": "Petrobras Relatório de Gestão 2025", "url": "https://transparencia.petrobras.com.br/documents/41711919/42223019/Relatorio%2Bde%2BGestao%2B-%2B2025.pdf/9ef35500-1f38-0cb6-539d-072b0569514f?download=true&t=1779977455000&version=1.0", "date": "2026-05", "type": "Issuer management report", "publisher": "Petrobras", "locator": "2025 innovation highlights, physical/printed p. 133", "review_status": "Full original source reviewed in Luna verification and editor checks on 6 October 2026."}]}, {"packet_id": "oct6-three-rounds-ai-195", "review_date": "2026-10-06", "title": "Snowflake architecture: operating drilling ML versus dashboard development", "workflow": "The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.", "technology_data": "The vendor describes drilling ML and centralized SCADA alerts in use, a Cortex drilling dashboard under development and an MLOps roadmap. Near-zero platform maintenance is qualitative. Its 40% M&A improvement is not AI-specific; over 1,000 hours per year is a forecast for one business unit.", "deployment_scope": "Operating drilling ML and SCADA alerts; a full-scale Cortex drilling dashboard under development and MLOps platform roadmap. Undated vendor page cannot establish launch dates.", "result_label": "Supplemental source: scope and limits", "attribution": "Vendor-published; not independently evaluated", "economics": "No project cost and measured cash-benefit pair disclosed.", "limitations": ["Current platform/ML use and future dashboard/MLOps work have different stages; platform efficiency does not establish AI-only financial benefit."], "verification": "verified partner", "source_review": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "source_locator": "Digging deeper into operational efficiency; Energizing the future", "timeline_date": "Undated; reviewed 2026-10-06", "timeline_event": "Snowflake architecture: operating drilling ML versus dashboard development", "stage": "Operating. Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.", "key": "oct6-three-rounds-ai-195", "record_ids": ["AI-195"], "company": "Expand Energy", "category": "Internal adoption", "department": "Drilling & completions", "source_ids": ["SW7-013"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["SW7-013"], "workflow": ["SW7-013"], "technology_data": ["SW7-013"], "deployment_scope": ["SW7-013"], "practical_relevance": ["SW7-013"], "economics": ["SW7-013"], "integration_human_role": ["SW7-013"]}, "disclosed_results": [{"metric": "Supplemental source: scope and limits", "value": "Near-zero maintenance is a qualitative platform description. The 40% M&A improvement and estimated >1,000 hours/year are not AI-specific realized results.", "period": "No publication/update date displayed; accessed 2026-10-01", "attribution": "Vendor-published; not independently evaluated", "source_ids": ["SW7-013"]}], "timeline": [{"date": "Undated; reviewed 2026-10-06", "event": "Snowflake architecture: operating drilling ML versus dashboard development", "source_ids": ["SW7-013"]}], "sources": [{"id": "SW7-013", "title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "date": "No publication/update date displayed; accessed 2026-10-01", "type": "Vendor customer case", "publisher": "Expand Energy", "locator": "Digging deeper into operational efficiency; Energizing the future", "review_status": "Full original source reviewed in Luna verification and editor checks on 6 October 2026."}]}, {"packet_id": "oct6-three-rounds-ai-217", "review_date": "2026-10-06", "title": "CPAD predecessor: historical deployment and benefit boundaries", "workflow": "The issuer reports US$29m saved through performance/reliability improvement and more than US$90m avoided costs through early prediction. GenAI integration and expansion to static equipment are roadmap work.", "technology_data": "The issuer reports US$29m saved through performance/reliability improvement and more than US$90m avoided costs through early prediction. GenAI integration and expansion to static equipment are roadmap work.", "deployment_scope": "CPAD historical rollout began in 2019. The 2024 report plans a transition to Neuron 5; the existing 2025 Neuron 5 disclosure remains the current evidence.", "result_label": "Historical CPAD predecessor claims; not Neuron 5 results", "attribution": "Operator-reported", "economics": "US$29m performance/reliability savings and >US$90m early-prediction cost avoidance have no common defined period or overlap reconciliation. Not summed, annualized or counted as Neuron 5 financial results.", "limitations": ["Historical CPAD lineage added to existing Neuron 5 dossier, not a new counted initiative. CPAD amounts are not Neuron 5 results, annual savings, net ROI or additive benefit."], "verification": "verified primary", "source_review": "Full original source reviewed in Luna verification and editor checks on 6 October 2026.", "source_locator": "Printed pp. 26–27, PDF physical p. 15; CPAD deployment and roadmap", "timeline_date": "2025 (2024 report; exact publication day not stated)", "timeline_event": "CPAD predecessor: historical deployment and benefit boundaries", "stage": "Company describes an operating platform; deployment coverage and go-live dates undisclosed.", "key": "oct6-three-rounds-ai-217", "record_ids": ["AI-217"], "company": "ADNOC Gas", "category": "Internal adoption", "department": "Maintenance & reliability", "source_ids": ["OCT5-S042"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S042"], "workflow": ["OCT5-S042"], "technology_data": ["OCT5-S042"], "deployment_scope": ["OCT5-S042"], "practical_relevance": ["OCT5-S042"], "economics": ["OCT5-S042"], "integration_human_role": ["OCT5-S042"]}, "disclosed_results": [{"metric": "Historical CPAD predecessor claims; not Neuron 5 results", "value": "US$29m reported savings and >US$90m avoided costs. These have no reconciled overlap or common measurement period.", "period": "2025 (2024 report; exact publication day not stated)", "attribution": "Operator-reported", "source_ids": ["OCT5-S042"]}], "timeline": [{"date": "2025 (2024 report; exact publication day not stated)", "event": "CPAD predecessor: historical deployment and benefit boundaries", "source_ids": ["OCT5-S042"]}], "sources": [{"id": "OCT5-S042", "title": "ADNOC Gas Integrated Report 2024", "url": "https://www.adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas-integrated-report-2024-english.ashx", "date": "2025 (2024 report; exact publication day not stated)", "type": "Issuer integrated annual report", "publisher": "ADNOC Gas", "locator": "Printed pp. 26–27, PDF physical p. 15; CPAD deployment and roadmap", "review_status": "Full original source reviewed in Luna verification and editor checks on 6 October 2026."}]}, {"packet_id": "oct6-executive-ai-037", "review_date": "2026-10-06", "title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "publisher": "Aramco", "source_title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "source_date": "2026-03-10", "source_type": "Operator annual report", "source_locator": "Printed pp. 64–65 / PDF pp. 37–38", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "Aramco reports the contribution of deployed AI solutions to its FY2025 technology portfolio. Individual projects are not separately valued here.", "technology_data": "TRV measures capital or operating expenditure avoided, or added revenue, versus a baseline. The annual report splits AI-driven and non-AI value; it does not disclose the calculation for each project.", "deployment_scope": "FY2025 aggregate; AI decomposition across value types and projects not disclosed.", "practical_relevance": "The AI subtotal gives a portfolio figure to investigate. Ask which projects produced it and how benefits compare with the full cost of implementation and ongoing operations.", "business_problem": "Understanding how much of a technology portfolio’s reported benefit comes from AI.", "economics": "TRV is not net cash ROI; no project-level bridge or verifier opinion reviewed.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Portfolio aggregate; issuer says third-party verified but verification scope not reviewed. No project-level denominator or AI-only benefit composition. Do not allocate total capex avoidance to AI or add distinct-year records."], "timeline_event": "FY2025 TRV and AI/non-AI splits reported in annual report; deployment dates not supplied.", "timeline_date": "2026-03-10", "source_ids": ["OCT5-S043", "RS109"], "supplementary_sources": [{"title": "Saudi Aramco FY 2025 Results presentation", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "date": "2026-03-10", "publisher": "Aramco", "locator": "Slide 10 / PDF p. 10: $3.9bn capex avoidance, $1.0bn opex reduction, $0.4bn revenue enhancement; mix is total TRV only."}], "stage": "FY2025 aggregate technology value reported for deployed technologies; AI subtotal is portfolio-level.", "key": "oct6-executive-ai-037", "record_ids": ["AI-037"], "company": "Aramco", "category": "Internal adoption", "department": "Finance, investment & strategy", "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S043"], "workflow": ["OCT5-S043"], "technology_data": ["OCT5-S043"], "deployment_scope": ["OCT5-S043"], "practical_relevance": ["OCT5-S043"], "economics": ["OCT5-S043"], "integration_human_role": ["OCT5-S043"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": "Issuer-reported $2.6bn AI-attributed TRV; $3.9bn capex avoidance is for total portfolio and must not be allocated to AI.", "period": "2026-03-10", "attribution": "Operator account", "source_ids": ["OCT5-S043"]}], "timeline": [{"date": "2026-03-10", "event": "FY2025 TRV and AI/non-AI splits reported in annual report; deployment dates not supplied.", "source_ids": ["OCT5-S043"]}], "sources": [{"id": "OCT5-S043", "title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "date": "2026-03-10", "type": "Operator annual report", "publisher": "Aramco", "locator": "Printed pp. 64–65 / PDF pp. 37–38", "review_status": "Relevant original primary source reviewed on 2026-10-06."}, {"id": "RS109", "title": "Saudi Aramco FY 2025 Results presentation", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-english.pdf", "date": "2026-03-10", "type": "Investor presentation", "publisher": "Aramco", "locator": "Slide 10 / PDF p. 10: Technology Realized Value, total-portfolio mix", "review_status": "Full primary text reviewed"}]}, {"packet_id": "oct6-executive-ai-077", "review_date": "2026-10-06", "title": "TotalEnergies and Mistral announce a three-year frontier subsurface AI program", "publisher": "TotalEnergies", "source_title": "TotalEnergies and Mistral announce a three-year frontier subsurface AI program", "source_date": "2026-09-15", "source_type": "Operator press release", "source_locator": "Release body, “Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data”", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "TotalEnergies and Mistral plan to build frontier models for exploration, reservoir characterization and development through a joint scientific laboratory.", "technology_data": "The program combines AI, the operator’s geoscience expertise and subsurface data. Model benchmarks, field deployment and partner budget allocations have not been disclosed.", "deployment_scope": "Announced joint development program; deployment and partner budget allocations not reported.", "practical_relevance": "This is a commitment to build subsurface models. Track validation against existing methods and field deployment before assuming an exploration advantage.", "business_problem": "Developing models that can help interpret subsurface data and inform reservoir decisions.", "economics": "Not cash already spent or realized benefit; investment allocation between partners is unstated.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Existing linked disclosure; not a new project/profile. Exclude dimensionally invalid “petaflops of data” metric."], "timeline_event": "Joint program announced 2026-09-15; model deployment date not established.", "timeline_date": "2026-09-15", "source_ids": ["R021"], "stage": "Three-year joint development program announced; no field deployment or return disclosed.", "key": "oct6-executive-ai-077", "record_ids": ["AI-077"], "company": "TotalEnergies", "category": "Internal adoption", "department": "Exploration & geoscience", "report_date": "2026-10-06", "field_sources": {"business_problem": ["R021"], "workflow": ["R021"], "technology_data": ["R021"], "deployment_scope": ["R021"], "practical_relevance": ["R021"], "economics": ["R021"], "integration_human_role": ["R021"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": ">€100m investment commitment over three years; no deployed-model result or return.", "period": "2026-09-15", "attribution": "Operator account", "source_ids": ["R021"]}], "timeline": [{"date": "2026-09-15", "event": "Joint program announced 2026-09-15; model deployment date not established.", "source_ids": ["R021"]}], "sources": [{"id": "R021", "title": "TotalEnergies and Mistral launch €100M frontier subsurface AI program", "url": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "date": "2026-09-15", "type": "TotalEnergies", "publisher": "TotalEnergies", "locator": "Release body, “Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data”", "review_status": "Relevant original primary source reviewed on 2026-10-06."}]}, {"packet_id": "oct6-executive-ai-078", "review_date": "2026-10-06", "title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "publisher": "TotalEnergies", "source_title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "source_date": "2026-02-11", "source_type": "Operator results-call transcript", "source_locator": "PDF pp. 9–10 / lines 358–385", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "TotalEnergies is rolling out real-time asset-data collection and a common data layer, with a subsequent AI layer intended to improve production and availability.", "technology_data": "AspenTech Inmation collects real-time data; Cognite transforms, enriches, stores and exposes it. Management targets ten times as many asset data points and full deployment by end-2027.", "deployment_scope": "Signed platform rollout underway; full-deployment target end-2027; not completed AI deployment.", "practical_relevance": "This shows the data work behind a large operator’s AI plans. Compare data capture, interoperability and asset rollout; the number of sites does not measure AI value.", "business_problem": "Making operating data accessible across 40 upstream and 16 refining and chemical sites.", "economics": "No cost or realized return disclosed.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Enabling data layer; figures are targets. The 500-engineer goal covers HSE, digital plant and Integrated Power modelling, not AI-only."], "timeline_event": "February 2026 call describes ongoing rollout and end-2027 target.", "timeline_date": "2026-02-11", "source_ids": ["OCT5-S046"], "stage": "Signed data-platform rollout in progress; end-2027 full-deployment objective.", "key": "oct6-executive-ai-078", "record_ids": ["AI-078"], "company": "TotalEnergies", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S046"], "workflow": ["OCT5-S046"], "technology_data": ["OCT5-S046"], "deployment_scope": ["OCT5-S046"], "practical_relevance": ["OCT5-S046"], "economics": ["OCT5-S046"], "integration_human_role": ["OCT5-S046"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": "No realized AI result reported; site counts/data expansion/headcount are management objectives.", "period": "2026-02-11", "attribution": "Operator account", "source_ids": ["OCT5-S046"]}], "timeline": [{"date": "2026-02-11", "event": "February 2026 call describes ongoing rollout and end-2027 target.", "source_ids": ["OCT5-S046"]}], "sources": [{"id": "OCT5-S046", "title": "TotalEnergies 2025 Results & 2026 Objectives presentation transcript", "url": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-2026-objectives-transcript_2026_en.pdf", "date": "2026-02-11", "type": "Operator results-call transcript", "publisher": "TotalEnergies", "locator": "PDF pp. 9–10 / lines 358–385", "review_status": "Relevant original primary source reviewed on 2026-10-06."}]}, {"packet_id": "oct6-executive-ai-216", "review_date": "2026-10-06", "title": "ADNOC Gas Q4/FY2025 Results Presentation", "publisher": "ADNOC Gas", "source_title": "ADNOC Gas Q4/FY2025 Results Presentation", "source_date": "2026-02-09", "source_type": "Operator results presentation", "source_locator": "Slide 7 / PDF p. 6; disclaimer PDF p. 2", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "ADNOC Gas describes AI process optimization across plants and closed-loop real-time optimization at Ruwais.", "technology_data": "The FY2025 slide names closed-loop RTO within the production-optimization pillar. It does not specify the model architecture, control limits or how much of the pillar’s savings came from RTO.", "deployment_scope": "Ruwais closed-loop RTO deployment listed; $50m is pillar-level, not expressly RTO-only.", "practical_relevance": "Production optimization has a reported financial result to investigate. Ask ADNOC Gas to reconcile the plant-program figure with the wider pillar before comparing savings.", "business_problem": "Improving plant operating decisions and production performance.", "economics": "No costs, baseline or net/gross value bridge. Presentation states information is unverified.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Do not sum separate savings claims or allocate the full $50m to RTO; row-level scope and period are unclear."], "timeline_event": "FY2025 achievements slide published 2026-02-09; exact savings period not stated.", "timeline_date": "2026-02-09", "source_ids": ["OCT5-S047"], "stage": "Multiple plant optimization described as deployed; closed-loop RTO at Ruwais listed among FY2025 achievements.", "key": "oct6-executive-ai-216", "record_ids": ["AI-216"], "company": "ADNOC Gas", "category": "Internal adoption", "department": "Refining, LNG & process operations", "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S047"], "workflow": ["OCT5-S047"], "technology_data": ["OCT5-S047"], "deployment_scope": ["OCT5-S047"], "practical_relevance": ["OCT5-S047"], "economics": ["OCT5-S047"], "integration_human_role": ["OCT5-S047"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": "Issuer reports $50m savings realized on row; period/baseline/method unknown. Preserve separate >$20m disclosure.", "period": "2026-02-09", "attribution": "Operator account", "source_ids": ["OCT5-S047"]}], "timeline": [{"date": "2026-02-09", "event": "FY2025 achievements slide published 2026-02-09; exact savings period not stated.", "source_ids": ["OCT5-S047"]}], "sources": [{"id": "OCT5-S047", "title": "ADNOC Gas Q4/FY2025 Results Presentation", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "date": "2026-02-09", "type": "Operator results presentation", "publisher": "ADNOC Gas", "locator": "Slide 7 / PDF p. 6; disclaimer PDF p. 2", "review_status": "Relevant original primary source reviewed on 2026-10-06."}]}, {"packet_id": "oct6-executive-ai-218", "review_date": "2026-10-06", "title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "publisher": "ADNOC Gas", "source_title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "source_date": "2025-11-02", "source_type": "Operator press release", "source_locator": "Paragraphs 107–125", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "Gecko inspection data and Cantilever are intended to connect inspection, analysis and maintenance decisions, with AIQ supplying predictive models.", "technology_data": "The three-year program phases deployment and data modeling before wider access and continuous optimization. The company’s later results describe Cantilever across assets without a completion denominator.", "deployment_scope": "Phased agreement; later slide says Cantilever/Gecko “in all assets,” denominator/completion undefined.", "practical_relevance": "This is a practical example of connecting inspection data to maintenance planning. Ask which assets and decisions are live, and how responsibilities are divided between the operator and providers.", "business_problem": "Turning inspection findings into maintenance decisions.", "economics": "No baseline, contract/system costs, denominator, realized results or net return disclosed.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Forward-looking cost-savings claim, separate from production-optimization pillar; not net ROI."], "timeline_event": "Agreement announced 2025-11-02; year-one deployment/modeling, then years 2–3 scaling.", "timeline_date": "2025-11-02", "source_ids": ["OCT5-S048", "OCT5-S047"], "supplementary_source_urls": ["https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx"], "field_source_urls": {"deployment_scope": ["https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx"], "technology_data": ["https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx"], "workflow": ["https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx"]}, "stage": "Cantilever deployment described in current operator reporting; the three-year AIQ/Gecko collaboration is phased, with five-year savings forecast.", "key": "oct6-executive-ai-218", "record_ids": ["AI-218"], "company": "ADNOC Gas", "category": "Internal adoption", "department": "Maintenance & reliability", "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S048"], "workflow": ["OCT5-S048", "OCT5-S047"], "technology_data": ["OCT5-S048", "OCT5-S047"], "deployment_scope": ["OCT5-S048", "OCT5-S047"], "practical_relevance": ["OCT5-S048"], "economics": ["OCT5-S048"], "integration_human_role": ["OCT5-S048"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": "Issuer forecasts >$300m maintenance/inspection cost savings over five years; no realized value reported.", "period": "2025-11-02", "attribution": "Operator account", "source_ids": ["OCT5-S048"]}], "timeline": [{"date": "2025-11-02", "event": "Agreement announced 2025-11-02; year-one deployment/modeling, then years 2–3 scaling.", "source_ids": ["OCT5-S048"]}], "sources": [{"id": "OCT5-S048", "title": "ADNOC Gas Partners with AIQ and Gecko Robotics to Launch Pioneering Program to Transform Industrial Maintenance", "url": "https://www.adnocgas.ae/en/news-and-media/press-releases/2025/gecko-aiq", "date": "2025-11-02", "type": "Operator press release", "publisher": "ADNOC Gas", "locator": "Paragraphs 107–125", "review_status": "Relevant original primary source reviewed on 2026-10-06."}, {"id": "OCT5-S047", "title": "ADNOC Gas Q4/FY2025 Results Presentation", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2025/adnoc-gas---q4-2025-results-presentation.ashx", "date": "2026-02-09", "type": "Operator results presentation", "publisher": "ADNOC Gas", "locator": "Slide 7 / PDF p. 6; disclaimer PDF p. 2", "review_status": "Relevant original primary source reviewed on 2026-10-06."}]}, {"packet_id": "oct6-executive-ai-022", "review_date": "2026-10-06", "title": "Introducing Vantage: A new lens on upstream operations", "publisher": "ExxonMobil", "source_title": "Introducing Vantage: A new lens on upstream operations", "source_date": "2025-10-10", "source_type": "Operator article", "source_locator": "“Big numbers, bigger impact,” lines 54–64", "source_review": "Relevant original primary source reviewed on 2026-10-06.", "workflow": "Vantage centralizes remote upstream operations. Its disclosed scale and field-optimization results provide operating context; the source does not attribute those results to AI.", "technology_data": "The issuer article describes remote operations, centralized monitoring and optimization. It does not identify an AI model or separate the effect of AI from other operational changes.", "deployment_scope": "Company reports centralized real-time upstream monitoring/management; this is context for AI-022.", "practical_relevance": "Vantage helps explain ExxonMobil’s operating setup. Its broader optimization metrics cannot be used as an AI return without a separate attribution bridge.", "business_problem": "Monitoring and coordinating a large upstream asset base.", "economics": "No AI-specific economics; full implementation/operating costs not disclosed.", "result_label": "Issuer-reported disclosure and evidence boundary", "attribution": "Operator account", "verification": "verified primary", "limitations": ["Do not count metrics as AI-caused benefits; source covers broad operations optimization."], "timeline_event": "Vantage article dated 2025-10-10; oil-uplift measurement period not stated.", "timeline_date": "2025-10-10", "source_ids": ["OCT5-S049"], "stage": "Deployed", "key": "oct6-executive-ai-022", "record_ids": ["AI-022"], "company": "ExxonMobil", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S049"], "workflow": ["OCT5-S049"], "technology_data": ["OCT5-S049"], "deployment_scope": ["OCT5-S049"], "practical_relevance": ["OCT5-S049"], "economics": ["OCT5-S049"], "integration_human_role": ["OCT5-S049"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and evidence boundary", "value": "Company-reported values are not AI-attributed: 2024 flaring/driving data specified; oil uplift period unstated.", "period": "2025-10-10", "attribution": "Operator account", "source_ids": ["OCT5-S049"]}], "timeline": [{"date": "2025-10-10", "event": "Vantage article dated 2025-10-10; oil-uplift measurement period not stated.", "source_ids": ["OCT5-S049"]}], "sources": [{"id": "OCT5-S049", "title": "Introducing Vantage: A new lens on upstream operations", "url": "https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations", "date": "2025-10-10", "type": "Operator article", "publisher": "ExxonMobil", "locator": "“Big numbers, bigger impact,” lines 54–64", "review_status": "Relevant original primary source reviewed on 2026-10-06."}]}, {"packet_id": "oct6-executive-ai238-joltly-context", "review_date": "2026-10-06", "source_title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "publisher": "Joltly", "publisher_note": "Source publisher is Joltly; do not substitute Spearpoint (the customer) as publisher.", "source_date": "2026-08-08", "source_type": "First-party vendor workflow article", "source_locator": "Sections describing JIB/revenue-statement processing and exceptions", "source_review": "Full original vendor pages reviewed on 2026-10-06.", "workflow": "Joltly describes line-item JIB and purchaser-statement extraction, well/account mapping, exception flagging, line-level accounting output, deposit matching and posting revenue by well/product. Pricing page adds monthly platform plus usage-based fee structure and published operator plan cards.", "technology_data": "Offer-side product descriptions only; no independently validated accuracy or customer-specific controls.", "deployment_scope": "Context for existing Spearpoint customer case AI-238 only; no new adoption record.", "practical_relevance": "Enables buy-versus-build and cost diligence with real document/payment volume, contract terms, exception effort and close controls.", "business_problem": "JIB and purchaser revenue-statement processing.", "economics": "Joltly public pricing cards show $750/$1,500 monthly plan prices and usage-based fees; these are not Spearpoint contract prices. Customer implementation and recurring costs remain undisclosed.", "result_label": "Vendor offer context, not customer contract or outcome", "attribution": "Vendor", "verification": "verified partner", "limitations": ["Do not attribute $750/$1,500 pricing, included volume or implementation timing to Spearpoint.", "Vendor workflow article is capability description, not observed Spearpoint controls or independent accuracy study.", "Existing profile’s outputs are reviewed/exported by people; do not call this autonomous posting."], "timeline_date": "2026-10-06", "timeline_event": "Offer-side context reviewed 2026-10-06; does not change the existing customer implementation timeline.", "supplementary_source_urls": ["https://www.joltly.io/pricing", "https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas"], "title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "stage": "In production since May 2026 per the vendor case study; output is exported in WolfePak-compatible format. The source does not say that Joltly posts transactions autonomously.", "key": "oct6-executive-ai238-joltly-context", "record_ids": ["AI-238"], "company": "Spearpoint Resources", "category": "Internal adoption", "department": "Finance, investment & strategy", "source_ids": ["OCT5-S054", "OCT5-S053"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S054"], "workflow": ["OCT5-S054"], "technology_data": ["OCT5-S054"], "deployment_scope": ["OCT5-S054"], "practical_relevance": ["OCT5-S054"], "economics": ["OCT5-S054"], "integration_human_role": ["OCT5-S054"]}, "disclosed_results": [{"metric": "Vendor offer context, not customer contract or outcome", "value": "No additional customer result from these offer-side pages. Existing AI-238 reports 20–30 hours/week returned by the customer; do not convert time into cash savings.", "period": "2026-08-08", "attribution": "Vendor", "source_ids": ["OCT5-S054"]}], "timeline": [{"date": "2026-10-06", "event": "Offer-side context reviewed 2026-10-06; does not change the existing customer implementation timeline.", "source_ids": ["OCT5-S054"]}], "sources": [{"id": "OCT5-S054", "title": "Automating JIBs and Revenue Statements: The Two Documents That Eat Your Month", "url": "https://www.joltly.io/blog/automating-jib-revenue-statements-oil-gas", "date": "2026-08-08", "type": "First-party vendor workflow article", "publisher": "Joltly", "locator": "Sections describing JIB/revenue-statement processing and exceptions", "review_status": "Full original vendor pages reviewed on 2026-10-06."}, {"id": "OCT5-S053", "title": "Joltly Pricing — Operators & Non-Op / Royalty Owners", "url": "https://www.joltly.io/pricing", "date": "Undated; accessed 2026-10-06", "type": "First-party vendor offer page", "publisher": "Joltly", "locator": "Pricing cards and FAQ: monthly platform fee, usage fees, integrations and implementation claims", "review_status": "full text reviewed"}]}, {"review_date": "2026-10-06", "verification": "verified primary", "source_review": "Relevant original issuer source independently reviewed on 2026-10-06.", "attribution": "Operator account", "result_label": "Issuer-reported disclosure and scope limits", "packet_id": "oct6-executive-ai281", "title": "Aramco Intelligent Live Earth Model for AI-assisted geosteering", "source_locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “Aramco Intelligent Live Earth Model”", "workflow": "Aramco describes its Intelligent Live Earth Model as using AI to optimize well placement and maximize reservoir contact in geosteering operations.", "technology_data": "Company report identifies the workflow and use but gives no model architecture, data feeds or control/approval design.", "deployment_scope": "Listed among the company’s FY2025 impactful AI solutions; first deployment date, well count, regions and rollout denominator are not stated.", "practical_relevance": "A named upstream geosteering workflow links subsurface interpretation to well placement and drilling time.", "economics": "No well cost, implementation/sustaining cost, monetized benefit or payback disclosed.", "limitations": ["Operator-reported annual-report metrics. Baseline, time period, sample size, measurement method and well count are not stated.", "Do not translate two days per well into savings without rig/day cost and incremental drilling scope.", "No independent verification or project costs reviewed."], "timeline_date": "2026-03-10", "timeline_event": "FY2025 Annual Report describes the workflow; launch date and measurement period are not disclosed.", "stage": "FY2025 report presents the model as an impactful AI solution used in geosteering; first deployment date, well count and rollout denominator are not stated.", "key": "oct6-executive-ai281", "record_ids": ["AI-281"], "company": "Aramco", "category": "Internal adoption", "department": "Drilling & completions", "source_ids": ["OCT5-S043"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S043"], "workflow": ["OCT5-S043"], "technology_data": ["OCT5-S043"], "deployment_scope": ["OCT5-S043"], "practical_relevance": ["OCT5-S043"], "economics": ["OCT5-S043"], "integration_human_role": ["OCT5-S043"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and scope limits", "value": "Aramco reports approximately 10% increased reservoir contact and an average two-day reduction in drilling time per well.", "period": "2026-03-10", "attribution": "Operator account", "source_ids": ["OCT5-S043"]}], "timeline": [{"date": "2026-03-10", "event": "FY2025 Annual Report describes the workflow; launch date and measurement period are not disclosed.", "source_ids": ["OCT5-S043"]}], "sources": [{"id": "OCT5-S043", "title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "date": "2026-03-10", "type": "Operator annual report", "publisher": "Aramco", "locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “Aramco Intelligent Live Earth Model”", "review_status": "Relevant original issuer source independently reviewed on 2026-10-06."}]}, {"review_date": "2026-10-06", "verification": "verified primary", "source_review": "Relevant original issuer source independently reviewed on 2026-10-06.", "attribution": "Operator account", "result_label": "Issuer-reported disclosure and scope limits", "packet_id": "oct6-executive-ai282", "title": "Aramco PE.AI advanced petrophysical prediction and synthetic well logs", "source_locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “PE.AI advanced petrophysical prediction”", "workflow": "Aramco says PE.AI generates synthetic well logs and reduces the need for traditional well logging and drilling additional pilot holes.", "technology_data": "Named AI-powered platform in company report; model family, evaluation set, accuracy, quality controls and human review are not disclosed.", "deployment_scope": "Named among FY2025 impactful AI solutions; deployment start date, assets, wells and workflow rollout status are not specified.", "practical_relevance": "Potentially changes petrophysical data acquisition and subsurface interpretation requirements.", "economics": "No quantified cost avoidance, operating result, implementation cost or return disclosed.", "limitations": ["No denominator, accuracy benchmark, validation protocol, avoided logging/pilot-hole count, cost, or realized value.", "Do not convert qualitative “reduces the need” wording into a measured savings result."], "timeline_date": "2026-03-10", "timeline_event": "FY2025 Annual Report names the solution; operating start and scope are not disclosed.", "stage": "Named as an impactful AI solution; launch date, operating asset count and volume are not disclosed.", "key": "oct6-executive-ai282", "record_ids": ["AI-282"], "company": "Aramco", "category": "Internal adoption", "department": "Exploration & geoscience", "source_ids": ["OCT5-S043"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S043"], "workflow": ["OCT5-S043"], "technology_data": ["OCT5-S043"], "deployment_scope": ["OCT5-S043"], "practical_relevance": ["OCT5-S043"], "economics": ["OCT5-S043"], "integration_human_role": ["OCT5-S043"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and scope limits", "value": "Qualitative issuer claim that synthetic logs reduce the need for traditional logging and additional pilot holes; no quantified outcome.", "period": "2026-03-10", "attribution": "Operator account", "source_ids": ["OCT5-S043"]}], "timeline": [{"date": "2026-03-10", "event": "FY2025 Annual Report names the solution; operating start and scope are not disclosed.", "source_ids": ["OCT5-S043"]}], "sources": [{"id": "OCT5-S043", "title": "Saudi Aramco Annual Report 2025 — FY2025 Technology Realized Value", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf", "date": "2026-03-10", "type": "Operator annual report", "publisher": "Aramco", "locator": "Annual Report 2025, printed p. 64 / PDF p. 37, “Impactful AI solutions” → “PE.AI advanced petrophysical prediction”", "review_status": "Relevant original issuer source independently reviewed on 2026-10-06."}]}, {"review_date": "2026-10-06", "verification": "verified primary", "source_review": "Relevant original issuer source independently reviewed on 2026-10-06.", "attribution": "Operator account", "result_label": "Issuer-reported disclosure and scope limits", "packet_id": "oct6-executive-ai283", "title": "Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D", "source_locator": "Release paragraphs 38–50: signed contract, investment, compute capacity, workloads and efficiency forecasts", "workflow": "Pangea 5 is a supercomputer contract supporting advanced seismic engineering and multiple scientific/digital workloads, including AI research and development.", "technology_data": "TotalEnergies names Dell Technologies and NVIDIA as collaborators; system architecture/utilization and workload allocation are not reported.", "deployment_scope": "Contract signed; system at CSTJF in Pau; first commissioning planned in 2027.", "practical_relevance": "A material compute enabler behind subsurface imaging and AI R&D, separately classified as infrastructure rather than operator adoption.", "economics": "Company reports >€100m whole-system investment. No AI-only share, operating cost, compute utilization, model value or return disclosed.", "limitations": ["Infrastructure contract supports several scientific/digital uses; amount is not AI-specific spend.", "Energy and cooling savings are forecasts, not realized values.", "Do not add >€100m to the separate TotalEnergies–Mistral joint program investment or count as AI ROI."], "timeline_date": "2026-05-06", "timeline_event": "Contract announced 2026-05-06; first commissioning targeted in 2027.", "stage": "Contract signed; first commissioning planned for 2027.", "key": "oct6-executive-ai283", "record_ids": ["AI-283"], "company": "TotalEnergies", "category": "Strategy / enabling technology", "department": "IT, data & enterprise transformation", "source_ids": ["OCT5-S052"], "report_date": "2026-10-06", "field_sources": {"business_problem": ["OCT5-S052"], "workflow": ["OCT5-S052"], "technology_data": ["OCT5-S052"], "deployment_scope": ["OCT5-S052"], "practical_relevance": ["OCT5-S052"], "economics": ["OCT5-S052"], "integration_human_role": ["OCT5-S052"]}, "disclosed_results": [{"metric": "Issuer-reported disclosure and scope limits", "value": "Sixfold computing-power increase is planned. Company forecasts approximately 40% lower energy use at equal performance and fivefold lower cooling-system consumption.", "period": "2026-05-06", "attribution": "Operator account", "source_ids": ["OCT5-S052"]}], "timeline": [{"date": "2026-05-06", "event": "Contract announced 2026-05-06; first commissioning targeted in 2027.", "source_ids": ["OCT5-S052"]}], "sources": [{"id": "OCT5-S052", "title": "Pangea 5 high-performance computing infrastructure supporting seismic and AI R&D", "url": "https://totalenergies.com/newsroom/totalenergies-develops-pangea-5-a-next-generation-supercomputer-469512/?lang=eng", "date": "2026-05-06", "type": "Operator press release", "publisher": "TotalEnergies", "locator": "Release paragraphs 38–50: signed contract, investment, compute capacity, workloads and efficiency forecasts", "review_status": "Relevant original issuer source independently reviewed on 2026-10-06."}]}, {"packet_id": "oct10-presidio-august-deck", "key": "oct10-presidio-august-deck", "record_ids": ["AI-231"], "company": "Presidio Production", "title": "August presentation: achieved uplift, platform scope and 2026 roadmap", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-10", "verification": "verified primary", "business_problem": "Turn field observations and production data into timely operating decisions.", "workflow": "Live 24/7 well surveillance and natural-language production queries support field operators and engineers. The August deck repeats 2.3% YTD uplift and the 3–5% target, and outlines budget, acquisition-screening and fleet-orchestration agents as 2026 deployment work. Its roughly 2,300-well platform figure appears on the goal slide; no completed full-fleet autonomous rollout is established.", "deployment_scope": "Live surveillance is described; full-fleet agentic orchestration is roadmap language.", "integration_human_role": "Operators and engineers use surveillance and queries. The source does not establish autonomous budget approval or acquisition decisions.", "economics": "Earlier Q2 revenue estimates are not net cash. The later revenue-per-uplift-point is illustrative; full costs remain undisclosed.", "limitations": ["Company-reported attribution; no independent counterfactual or full costs. The August 11 release says ~1%, August 12 remarks report 2.3%, and August 17 slides repeat 2.3% YTD; dates and populations need reconciliation. The Q2 call describes ~2,000 platform wells; the goal slide cites ~2,300. Platform coverage is not proof that every well is under autonomous control. Budget agents, acquisition screening and full-fleet orchestration remain roadmap items. $4.5m annualized revenue across initiatives and $2.5m per uplift point are different measures; neither is net cash ROI. No-new-drilling or no-capex language does not disclose software, controller, staffing or support costs."], "source_ids": ["OCT10-PRESIDIO-0817"], "disclosed_results": [{"metric": "August presentation: achieved claim, target and platform scope", "value": "August deck: 2.3% YTD achieved, company-reported; 3–5% 2026 production-uplift target. Roughly 2,300 wells appears on the intelligence-platform GOAL slide; coverage date and active intervention population are not established. $2.5m annual revenue per 1% uplift at $30/boe is illustrative, not a realized return.", "period": "2026-08-17", "attribution": "Issuer-reported; not independently evaluated", "source_ids": ["OCT10-PRESIDIO-0817"]}], "timeline": [{"date": "2026-08-17", "event": "The August deck repeats 2.3% YTD AI-linked production uplift and a 3–5% 2026 target. It describes live surveillance and natural-language production queries. Its goal slide cites roughly 2,300 platform wells; fleet-wide agentic orchestration, budget agents and automatic acquisition screening appear in the 2026 deployment column, not the NOW column.", "source_ids": ["OCT10-PRESIDIO-0817"]}], "sources": [{"id": "OCT10-PRESIDIO-0817", "title": "Presidio August 2026 investor presentation", "url": "https://ir.bypresidio.com/sec-filings/all-sec-filings/content/0001213900-26-090532/ea030229001ex99-1.htm", "date": "2026-08-17", "type": "SEC-filed investor presentation", "publisher": "Presidio Production", "locator": "Slides 20–23; current surveillance versus 2026 deployment column on slide 21; GOAL slide 23", "review_status": "Relevant issuer-filed presentation text reviewed on 10 October 2026; column interpretation also recorded in the October 9 review."}], "field_sources": {"business_problem": ["OCT10-PRESIDIO-0817"], "workflow": ["OCT10-PRESIDIO-0817"], "deployment_scope": ["OCT10-PRESIDIO-0817"], "integration_human_role": ["OCT10-PRESIDIO-0817"], "economics": ["OCT10-PRESIDIO-0817"]}}, {"packet_id": "oct10-field-ai-284", "key": "oct10-field-ai-284", "record_ids": ["AI-284"], "company": "BPX Energy", "title": "Daily lift recommendations across a 441-well Permian pilot", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-10", "verification": "verified partner", "stage": "Pilot on a 441-well asset; result population limited to wells adopting recommendations", "business_problem": "Lift settings and intervention choices must account for changing well conditions and shared gas-injection constraints.", "workflow": "Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.", "technology_data": "Xecta combines physics-based well-flow models with AI and historical and daily production data. It recommends gas-lift injection rates, ESP frequencies, lift-system transitions and acid-job candidates, while considering shared compressor and injection-pressure constraints.", "integration_human_role": "Operators receive daily recommendations. The paper does not establish autonomous actuation or disclose how many wells adopted the recommendations.", "deployment_scope": "Permian asset with 441 gas-lift and ESP wells. BPX is bp’s US onshore business; this is a distinct field case, not a count of bp-wide deployment.", "economics": "Lower gas-lift use could reduce resource demand, but no dollar saving, implementation cost or net return is disclosed.", "practical_relevance": "A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.", "assessment": "A useful example of recommendations that account for field constraints. Ask how many recommendations were accepted, what changed in the field and how incremental production was calculated.", "limitations": ["The 6% result applies to wells adopting recommendations, not all 441 wells. The gas-lift figure is a separate resource-use measure. Duration, adopter count, comparison method, costs and net return are not disclosed. The operator/vendor-coauthored abstract is not independent validation. The source is dated 2026; an exact publication or deployment day is not established."], "source_ids": ["OCT10-AI-284-SWPSC"], "partners": ["Xecta"], "disclosed_results": [{"metric": "Author-reported field results", "value": "Authors report average 6% incremental oil on wells that adopted recommendations, plus more than 20% lower average field gas-lift use. The number of adopting wells is not disclosed.", "period": "Duration not disclosed", "attribution": "Operator/vendor account; not independently evaluated", "source_ids": ["OCT10-AI-284-SWPSC"]}], "timeline": [{"date": "2026", "event": "Pilot on a 441-well asset; result population limited to wells adopting recommendations. Source date is not the deployment start.", "source_ids": ["OCT10-AI-284-SWPSC"]}], "sources": [{"id": "OCT10-AI-284-SWPSC", "title": "Daily lift recommendations across a 441-well Permian pilot", "url": "https://www.swpshortcourse.org/paper/2026051-pilot-test-continuous-production-optimization-using-digital-solution-permian-basin", "date": "2026", "type": "Operator/vendor-coauthored technical paper", "publisher": "Southwestern Petroleum Short Course", "locator": "SWPSC 2026051 abstract: method, 441-well asset and reported results", "review_status": "Relevant original source text reviewed; performance remains author-reported, not independently validated."}], "field_sources": {"business_problem": ["OCT10-AI-284-SWPSC"], "workflow": ["OCT10-AI-284-SWPSC"], "technology_data": ["OCT10-AI-284-SWPSC"], "integration_human_role": ["OCT10-AI-284-SWPSC"], "deployment_scope": ["OCT10-AI-284-SWPSC"], "economics": ["OCT10-AI-284-SWPSC"]}}, {"packet_id": "oct10-field-ai-285", "key": "oct10-field-ai-285", "record_ids": ["AI-285"], "company": "Burk Royalty", "title": "Pump-by-exception monitoring and control at a West Texas lease", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-10", "verification": "verified partner", "stage": "Operating small-site system; eight-month evaluation", "business_problem": "A remote satellite lease required routine visits and manual pumping checks, with risks from equipment problems and tank overfill.", "workflow": "Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.", "technology_data": "Amplified combines controllers, sensors, tank and facility monitoring, cameras, cloud AI/ML and mobile tools. Automated setpoint optimization and protective shutdowns respond to conditions such as parted rods, abnormal pressure and tank overfill. People handle the exceptions.", "integration_human_role": "Software adjusts setpoints and can shut equipment down for protection; staff review alerts and respond to exceptions. This is an integrated instrumentation, control and work-process change, not a stand-alone chatbot.", "deployment_scope": "West Texas satellite lease. The full paper describes five sucker-rod producers and one injector; its HTML abstract says six producers and one injector. The source population conflict remains unresolved.", "economics": "Some operating costs are disclosed: electricity expense falls from $660 to $376 per well-month, alongside hot-oiling and workload measures. Authors calculate ~$65,000 incremental profit over eight months at $65/bbl WTI and payback slightly over one month. The capital amount and full recurring-cost bridge are undisclosed, so this is partial cost evidence, not a validated all-in ROI. Potential avoided spill-cleanup cost is an event estimate, not recurring savings.", "practical_relevance": "An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.", "assessment": "An unusually concrete small-site account of changes in field workload and operating cost. Its next test is whether the calculation holds after all hardware, software, installation and support costs, and across more leases.", "limitations": ["The full paper says five producers plus one injector; the abstract says six producers plus one injector. Routine field time is measured in hours, not a count of interventions. Production aggregation and the counterfactual decline need clarification. Profit and payback are author-calculated, not independently validated; actual capital and an itemized total-cost bridge are absent. Benefits apply to the integrated system; the AI-only contribution is not isolated. Future profit and wider-fleet extrapolations are projections, not achieved results. Exact publication and deployment days are not established."], "source_ids": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "partners": ["Amplified Industries"], "disclosed_results": [{"metric": "Author-reported field results", "value": "Over eight months, authors report 43% lower electricity consumption and daily strokes; electricity expense of $660 to $376 per well-month; routine field time of 126 to about 18 hours per month (86% lower); and 26% lower hot-oiling cost. Production stayed around 19 BOPD versus an expected 4.25% annual decline. Authors calculate about $65,000 incremental profit at $65/bbl WTI and capital recovery in slightly over one month.", "period": "Eight months", "attribution": "Operator/vendor account; not independently evaluated", "source_ids": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"]}], "timeline": [{"date": "2026", "event": "Operating small-site system; eight-month evaluation. Source date is not the deployment start.", "source_ids": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"]}], "sources": [{"id": "OCT10-AI-285-SWPSC", "title": "Pump-by-exception monitoring and control at a West Texas lease", "url": "https://www.swpshortcourse.org/paper/2026052-field-optimization-reimagined-data-core-exceptions-center-people-control", "date": "2026", "type": "Operator/vendor-coauthored technical paper", "publisher": "Southwestern Petroleum Short Course", "locator": "SWPSC 2026052 abstract; linked full paper PDF pp. 3, 7–9 (site configuration, eight-month results, economics and projections)", "review_status": "Relevant original source text reviewed; performance remains author-reported, not independently validated."}, {"id": "OCT10-BURK-PDF", "title": "Field Optimization Reimagined: full paper", "url": "https://www.swpshortcourse.org/sites/default/files/papers/2026052%20-%20Field%20Optimization%20Reimagined%20Data%20At%20The%20Core%20Exceptions%20At%20The%20Center%2C%20People%20In%20Control.pdf", "date": "2026", "type": "Operator/vendor-coauthored technical paper", "publisher": "Southwestern Petroleum Short Course", "locator": "PDF pp. 3, 7–9; figure on p. 8", "review_status": "Relevant original source text reviewed; performance remains author-reported, not independently validated."}], "field_sources": {"business_problem": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "workflow": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "technology_data": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "integration_human_role": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "deployment_scope": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"], "economics": ["OCT10-AI-285-SWPSC", "OCT10-BURK-PDF"]}}, {"packet_id": "oct10-field-ai-286", "key": "oct10-field-ai-286", "record_ids": ["AI-286"], "company": "Expand Energy", "title": "Autonomous liquid unloading on nine Haynesville gas wells", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-10", "verification": "verified partner", "stage": "Operating field test on nine wells across eight pads; 63–83-day results", "business_problem": "Liquid loading in intermittent gas wells requires timely shut-ins and reopening to restore flow without unnecessary production loss.", "workflow": "An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.", "technology_data": "An edge IIoT gateway uses surface pressure, temperature and flow data with physics-based critical-velocity logic. It decides shut-in timing and actuates the choke locally. Cloud machine learning predicts shut-in duration; local control does not require continuous cloud connectivity or operator oversight.", "integration_human_role": "The system makes local shut-in and choke decisions. The sources do not describe a language model or a general-purpose agent; this is specialist production control.", "deployment_scope": "Nine Haynesville wells across eight pads. The related July 2026 JPT account reports results over 63–83 days; it describes the same deployment, not an independent replication.", "economics": "The reported short-period production gains and estimated annual gas volume are not net financial returns. Full implementation and running costs are not disclosed in the reviewed primary passages.", "practical_relevance": "A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.", "assessment": "A concrete example of local control with cloud learning. Ask how the baseline accounts for unloading cycles and decline, what safety permissions apply, and whether gains persist through a full operating year.", "limitations": ["Short field-test results do not establish a year-long or fleet-wide outcome. More than 80 MMscf per well-year is explicitly estimated. The conference abstract and SLB-authored JPT account describe the same deployment. The reviewed primary passages do not disclose full hardware, software, integration or ongoing support costs or net ROI. Annual financial figures from an author-uploaded manuscript are held pending publisher-source access. The conference source is dated 2026; its exact publication day is not established."], "source_ids": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "partners": ["SLB"], "disclosed_results": [{"metric": "Author-reported field results", "value": "SLB’s JPT account reports 70–139% cumulative gas-production gains over 63–83 days on nine wells across eight pads, with daily gains up to 350 Mscf/d. More than 80 MMscf additional gas per well-year is an estimate, not a year of observed production.", "period": "63–83 days", "attribution": "Operator/vendor account; not independently evaluated", "source_ids": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"]}], "timeline": [{"date": "2026-07-01", "event": "Operating field test on nine wells across eight pads; 63–83-day results. Source date is not the deployment start.", "source_ids": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"]}], "sources": [{"id": "OCT10-AI-286-SWPSC", "title": "Autonomous liquid unloading on nine Haynesville gas wells", "url": "https://www.swpshortcourse.org/paper/2026001-autonomous-edge-based-optimization-liquid-loading-intermittent-gas-wells", "date": "2026", "type": "Operator/vendor-coauthored technical paper", "publisher": "Southwestern Petroleum Short Course", "locator": "SWPSC 2026001 abstract; JPT 1 July 2026, Haynesville Field Results", "review_status": "Relevant original source text reviewed; performance remains author-reported, not independently validated."}, {"id": "OCT10-EXPAND-JPT", "title": "Edge Computing and IIoT Enable Autonomous Artificial Lift and Well Optimization", "url": "https://jpt.spe.org/case-study-edge-computing-and-iiot-enable-autonomous-artificial-lift-and-well-optimization-across-multi-basin-deployments", "date": "2026-07-01", "type": "Supplier-authored technical account", "publisher": "Journal of Petroleum Technology / SPE", "locator": "Field Results: Haynesville Basin", "review_status": "Relevant original source text reviewed; performance remains author-reported, not independently validated."}], "field_sources": {"business_problem": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "workflow": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "technology_data": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "integration_human_role": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "deployment_scope": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"], "economics": ["OCT10-AI-286-SWPSC", "OCT10-EXPAND-JPT"]}}, {"packet_id": "oct12-flywheel-app-platform", "key": "oct12-flywheel-app-platform", "record_ids": ["AI-287"], "company": "Flywheel Energy", "title": "Internal AI app platform and business-team automation", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-12", "verification": "verified primary", "stage": "Company-reported platform use; adoption unquantified", "business_problem": "Let business teams improve their own workflows without asking the central software team to build every application.", "workflow": "Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.", "technology_data": "Custom AI-enabled webapp platform; model, provider and implementation stack not disclosed.", "integration_human_role": "Employees build apps; engineering owns the platform, training, QA and transition of important apps to central support. This is not autonomous executive decision-making.", "deployment_scope": "Nontechnical business teammates and basin-team training; no basin, user or app population is quantified.", "economics": "Growing adoption is described qualitatively. No user/app count, quantified business result, complete cost or net return is disclosed.", "practical_relevance": "Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.", "assessment": "Flywheel is a useful example of employees building business applications with internal engineering support. The next questions are what they build, how important apps are governed and supported, and whether the improvement survives maintenance and review costs.", "limitations": ["The employer posting supports the existence of an internal AI-enabled app platform, but does not disclose its name, model/provider, user or application count, production workload, error rate, cost, savings or ROI. The phrase “this year” and relative posting dates do not establish an exact launch date. Hiring tasks and qualifications are separate from existing platform use. No named field-optimization deployment or documented AI failure was established in this bounded search; this does not prove absence. Quorum customer status does not establish use of QAI or its AI extraction. The intern source is indexed-only after a redirect."], "source_ids": ["OCT12-FLY-PLATFORM", "OCT12-FLY-ACCOUNTING", "OCT12-FLY-FEED", "OCT12-FLY-QUORUM", "OCT12-FLY-INTERN", "OCT12-FLY-COMPANY"], "partners": [], "disclosed_results": [], "timeline": [{"date": "2026", "event": "Flywheel says it developed a custom platform that lets nontechnical teammates use AI to build web applications for business use cases. The platform-engineer posting refers to growing adoption and hiring to maintain and expand the platform. Responsibilities cover the roadmap, database and portal integration, external data, demonstrations, training, quality assurance and moving important applications into central software-team support. Basin-team training is included.", "source_ids": ["OCT12-FLY-PLATFORM"]}], "sources": [{"id": "OCT12-FLY-PLATFORM", "title": "Flywheel AI Platform Engineer", "url": "https://www.linkedin.com/jobs/view/ai-platform-engineer-at-flywheel-energy-llc-4474350750", "date": "2026", "type": "Employer recruitment disclosure", "publisher": "Flywheel Energy / LinkedIn", "locator": "Job Summary and Key Responsibilities", "review_status": "Original employer-authored full description reviewed 12 October 2026. Existing custom platform and growing adoption are company claims; no outcome validation."}, {"id": "OCT12-FLY-ACCOUNTING", "title": "Flywheel AI Analyst — accounting automation", "url": "https://www.linkedin.com/jobs/view/ai-analyst-at-flywheel-energy-llc-4472009913", "date": "2026", "type": "Employer recruitment disclosure", "publisher": "Flywheel Energy / LinkedIn", "locator": "Role Summary; Automation; Data Management; Required Qualifications", "review_status": "Original employer-authored full description reviewed 12 October 2026. Responsibilities and candidate qualifications are not completed deployments."}, {"id": "OCT12-FLY-FEED", "title": "Flywheel recruitment update: land, accounting and AI platform", "url": "https://www.linkedin.com/company/flywheel-energy", "date": "2026", "type": "Employer recruitment disclosure", "publisher": "Flywheel Energy / LinkedIn", "locator": "AI recruitment update naming AI Analyst–Land, AI Analyst–Accounting and AI Platform Engineer", "review_status": "Company-authored LinkedIn feed update reviewed 12 October 2026; relative date only. Feed position may change."}, {"id": "OCT12-FLY-QUORUM", "title": "Quorum On Demand Land — Flywheel testimonial", "url": "https://www.quorumsoftware.com/solutions/upstream-on-demand/land/", "date": "Undated", "type": "Vendor customer testimonial", "publisher": "Quorum Software", "locator": "Proven for High-Growth Land Organizations; Flywheel customer testimonial", "review_status": "Original vendor page reviewed 12 October 2026. Named conventional software relationship; current product AI capabilities are not confirmed for Flywheel."}, {"id": "OCT12-FLY-INTERN", "title": "Flywheel AI Intern — Summer 2027", "url": "https://www.linkedin.com/jobs/view/ai-intern-summer-2027-at-flywheel-energy-llc-4461575049", "date": "2026", "type": "Employer recruitment disclosure", "publisher": "Flywheel Energy / LinkedIn", "locator": "Indexed employer description: About the AI Internship and What You Will Do", "review_status": "Full employer text available in search index on 12 October 2026; direct URL redirected to unrelated expired-job results. Prospective internship work, not a deployment source."}, {"id": "OCT12-FLY-COMPANY", "title": "Flywheel Energy — company and operating partner context", "url": "https://www.flywheelenergy.com/", "date": "Undated", "type": "Company website", "publisher": "Flywheel Energy", "locator": "Private E&P description; Stone Ridge footer", "review_status": "Official company page reviewed 12 October 2026. Identifies Flywheel as an operating partner of Stone Ridge Energy; does not establish a precise legal ownership schedule or internal AI adoption."}], "field_sources": {"business_problem": ["OCT12-FLY-PLATFORM"], "workflow": ["OCT12-FLY-PLATFORM"], "technology_data": ["OCT12-FLY-PLATFORM"], "integration_human_role": ["OCT12-FLY-PLATFORM"], "deployment_scope": ["OCT12-FLY-PLATFORM"], "economics": ["OCT12-FLY-PLATFORM"]}, "hiring_context": "Separately, Flywheel advertises land and accounting AI analyst roles. The accounting role specifies AI-assisted workflows, document extraction, API/data-pipeline integration, acquisition-data normalization and migration, retries, fallbacks and exception queues. It calls for human review of ownership or legal determinations. These are staffing responsibilities, not confirmation that every workflow is deployed. OpenAI/Azure OpenAI are examples of desired candidate experience, not confirmed vendors.", "adjacent_technology": "Quorum names Flywheel as an On Demand Land customer for more than five years. Its testimonial describes easier data movement and access for nontechnical users. Quorum’s newer AI features are general product capabilities; Flywheel-specific use is not established.", "prospective_research": "A search-indexed Summer 2027 internship description proposes LLM, agentic and MCP integration prototypes, plus ML work on production forecasting and predictive maintenance. The direct posting expired/redirected during review; these prospective tasks are not operating field AI."}, {"packet_id": "oct14-ai-288", "key": "oct14-ai-288", "record_ids": ["AI-288"], "company": "CNX Resources", "title": "Safety cameras that flag unsafe conditions", "category": "Internal adoption", "department": "HSE & emissions", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Company describes current use; start date unknown", "business_problem": "Identify unsafe conditions and improve safety decisions.", "workflow": "Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "technology_data": "Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "deployment_scope": "The current company page describes use, but gives no publication date or deployment start. Its 2025 safety milestones do not date the AI workflow.", "economics": "No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.", "practical_relevance": "Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.", "assessment": "Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.", "integration_human_role": "Which unsafe conditions trigger useful interventions, how many alerts are rejected, and did incident rates change on comparable sites?", "limitations": ["Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch."], "source_ids": ["OCT14-CNX-SAFETY"], "disclosed_results": [], "timeline": [{"date": "Undated", "event": "CNX says it uses AI video analytics to identify unsafe conditions, coach employees in real time and inform training. HSE++ also analyzes earlier incidents and safety observations to find trends.", "source_ids": ["OCT14-CNX-SAFETY"]}], "sources": [{"id": "OCT14-CNX-SAFETY", "title": "CNX Workplace Health and Safety", "url": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "date": "Undated", "type": "Company website", "publisher": "CNX Resources", "locator": "Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections", "review_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["OCT14-CNX-SAFETY"], "workflow": ["OCT14-CNX-SAFETY"], "technology_data": ["OCT14-CNX-SAFETY"], "deployment_scope": ["OCT14-CNX-SAFETY"], "economics": ["OCT14-CNX-SAFETY"], "integration_human_role": ["OCT14-CNX-SAFETY"], "practical_relevance": ["OCT14-CNX-SAFETY"]}}, {"packet_id": "oct14-ai-289", "key": "oct14-ai-289", "record_ids": ["AI-289"], "company": "Aker BP", "title": "Agents for risk assessment and alarm reviews", "category": "Internal adoption", "department": "HSE & emissions", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Implementation announced; operating results not established", "business_problem": "Coordinate operational risk assessment and alarm reviews across industrial systems.", "workflow": "The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "technology_data": "The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "deployment_scope": "Published 21 May 2026. Aker BP signed on to implement; completion date, site scope and sustained live use are not established.", "economics": "No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.", "practical_relevance": "This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.", "assessment": "This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.", "integration_human_role": "Which risk and alarm reviews are live, what actions can agents take, and where are human approval and safety-system boundaries?", "limitations": ["Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed."], "source_ids": ["OCT14-AKER-ABB"], "disclosed_results": [], "timeline": [{"date": "2026-05-21", "event": "Cognite and ABB name Aker BP as the first customer signed on to implement an agentic offering for risk assessment and alarm rationalization.", "source_ids": ["OCT14-AKER-ABB"]}], "sources": [{"id": "OCT14-AKER-ABB", "title": "Cognite and ABB collaborate on industrial agentic applications", "url": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "date": "2026-05-21", "type": "Vendor announcement", "publisher": "Cognite", "locator": "Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits", "review_status": "Relevant original source text reviewed 14 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["OCT14-AKER-ABB"], "workflow": ["OCT14-AKER-ABB"], "technology_data": ["OCT14-AKER-ABB"], "deployment_scope": ["OCT14-AKER-ABB"], "economics": ["OCT14-AKER-ABB"], "integration_human_role": ["OCT14-AKER-ABB"], "practical_relevance": ["OCT14-AKER-ABB"]}}, {"packet_id": "oct14b-ai-290", "key": "oct14b-ai-290", "record_ids": ["AI-290"], "company": "PetroChina", "title": "A model that diagnoses rod-pumped wells in Xinjiang", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Company reports scaled use on more than 3,800 wells", "business_problem": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.", "workflow": "A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "technology_data": "A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "deployment_scope": "15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.", "integration_human_role": "Staff review the outputs and act on exceptions; autonomous control is not established.", "economics": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "practical_relevance": "This connects a large industrial model to a specific engineering job: diagnose abnormal wells and explain what to inspect. The published results do not establish autonomous operation or a net return.", "assessment": "This connects a large industrial model to a specific engineering job: diagnose abnormal wells and explain what to inspect. The published results do not establish autonomous operation or a net return.", "limitations": ["Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain."], "source_ids": ["OCT14B-CNPC-XINJIANG"], "disclosed_results": [{"metric": "Source-reported workflow results", "value": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "period": "15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.", "attribution": "CNPC / China Petroleum News; not independently validated", "source_ids": ["OCT14B-CNPC-XINJIANG"]}], "timeline": [{"date": "2026-04-15", "event": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.", "source_ids": ["OCT14B-CNPC-XINJIANG"]}], "sources": [{"id": "OCT14B-CNPC-XINJIANG", "title": "CNPC: industrial large model for rod-pumped-well production optimization", "url": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "date": "2026-04-15", "type": "Company news article", "publisher": "CNPC / China Petroleum News", "locator": "Article dated 15 April 2026; first four body paragraphs", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["OCT14B-CNPC-XINJIANG"], "workflow": ["OCT14B-CNPC-XINJIANG"], "technology_data": ["OCT14B-CNPC-XINJIANG"], "deployment_scope": ["OCT14B-CNPC-XINJIANG"], "integration_human_role": ["OCT14B-CNPC-XINJIANG"], "economics": ["OCT14B-CNPC-XINJIANG"], "practical_relevance": ["OCT14B-CNPC-XINJIANG"]}}, {"packet_id": "oct14b-ai-291", "key": "oct14b-ai-291", "record_ids": ["AI-291"], "company": "CNX Resources", "title": "Finding methane-release precursors and prioritizing repairs", "category": "Internal adoption", "department": "HSE & emissions", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Company describes current use; exact start date unknown", "business_problem": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.", "workflow": "Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "technology_data": "Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "deployment_scope": "Undated current company page, with explicit 2026 program and Q1 2026 progress. No model launch date or deployment population is given.", "integration_human_role": "Staff review the outputs and act on exceptions; autonomous control is not established.", "economics": "CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.", "practical_relevance": "Methane response involves prediction, source identification and repair. The next useful evidence is whether this integration prevents releases or shortens the work needed to fix them.", "assessment": "Methane response involves prediction, source identification and repair. The next useful evidence is whether this integration prevents releases or shortens the work needed to fix them.", "limitations": ["This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority."], "source_ids": ["OCT14B-CNX-METHANE"], "disclosed_results": [], "timeline": [{"date": "Undated", "event": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.", "source_ids": ["OCT14B-CNX-METHANE"]}], "sources": [{"id": "OCT14B-CNX-METHANE", "title": "CNX Commitment to Methane Reduction", "url": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "date": "Undated", "type": "Company website", "publisher": "CNX Resources", "locator": "2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "field_sources": {"business_problem": ["OCT14B-CNX-METHANE"], "workflow": ["OCT14B-CNX-METHANE"], "technology_data": ["OCT14B-CNX-METHANE"], "deployment_scope": ["OCT14B-CNX-METHANE"], "integration_human_role": ["OCT14B-CNX-METHANE"], "economics": ["OCT14B-CNX-METHANE"], "practical_relevance": ["OCT14B-CNX-METHANE"]}}, {"packet_id": "oct14b-ai-292", "key": "oct14b-ai-292", "record_ids": ["AI-292"], "company": "HG Oil Holdings", "title": "An invoice assistant with human exception review", "category": "Internal adoption", "department": "Finance, investment & strategy", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Undated vendor-described use; benefits not independently verified", "business_problem": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.", "workflow": "Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "technology_data": "Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "deployment_scope": "Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.", "integration_human_role": "Staff review the outputs and act on exceptions; autonomous control is not established.", "economics": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "practical_relevance": "Invoice processing is a practical AI use case for smaller operators. Time released to other work is valuable capacity; it becomes a cash benefit only when the business avoids costs or improves an outcome.", "assessment": "Invoice processing is a practical AI use case for smaller operators. Time released to other work is valuable capacity; it becomes a cash benefit only when the business avoids costs or improves an outcome.", "limitations": ["Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata."], "source_ids": ["OCT14B-HG-INVOICES", "OCT14B-HG-IDENTITY", "OCT14B-HG-CONTROLS"], "disclosed_results": [{"metric": "Source-reported workflow results", "value": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "period": "Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.", "attribution": "Preisser Solutions; not independently validated", "source_ids": ["OCT14B-HG-INVOICES"]}], "timeline": [{"date": "Undated", "event": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.", "source_ids": ["OCT14B-HG-INVOICES"]}], "sources": [{"id": "OCT14B-HG-INVOICES", "title": "HG Oil Holdings: AI invoice processing", "url": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "date": "Undated", "type": "Vendor customer case", "publisher": "Preisser Solutions", "locator": "Before; What we built; Operational handoffs; Outcomes", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}, {"id": "OCT14B-HG-IDENTITY", "title": "HG Oil Holdings About", "url": "https://www.hgoilholdings.com/about", "date": "Undated", "type": "Company website", "publisher": "HG Oil Holdings", "locator": "Company description; current company-wide wells are not AI scope", "review_status": "Original page reviewed 14 October 2026; adjacent context, not independent outcome validation."}, {"id": "OCT14B-HG-CONTROLS", "title": "Preisser AI invoice processing for small businesses", "url": "https://preissersolutions.com/use-cases/ai-invoice-processing-small-business", "date": "Undated", "type": "Vendor product page", "publisher": "Preisser Solutions", "locator": "Human payment approval; names HG as reference; general product controls", "review_status": "Original page reviewed 14 October 2026; adjacent context, not independent outcome validation."}], "field_sources": {"business_problem": ["OCT14B-HG-INVOICES"], "workflow": ["OCT14B-HG-INVOICES"], "technology_data": ["OCT14B-HG-INVOICES"], "deployment_scope": ["OCT14B-HG-INVOICES"], "integration_human_role": ["OCT14B-HG-INVOICES", "OCT14B-HG-CONTROLS"], "economics": ["OCT14B-HG-INVOICES"], "practical_relevance": ["OCT14B-HG-INVOICES"]}}, {"packet_id": "oct14c-ai-293", "key": "oct14c-ai-293", "record_ids": ["AI-293"], "company": "Aramco", "title": "Connecting rock samples to a geological knowledge platform", "category": "Internal adoption", "department": "Exploration & geoscience", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Company describes use; additional ML work is underway", "business_problem": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.", "workflow": "Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "technology_data": "Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "deployment_scope": "Published 25 August 2026. The center’s 150+ employees and archive of more than one million feet of samples are facility context, not AI deployment or validation denominators.", "integration_human_role": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.", "economics": "No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.", "practical_relevance": "Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?", "assessment": "Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?", "limitations": ["Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established."], "source_ids": ["OCT14C-ARAMCO-CORE"], "sources": [{"id": "OCT14C-ARAMCO-CORE", "title": "Understanding the subsurface: inside Aramco’s Core Laboratories Center", "url": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "date": "2026-08-25", "type": "Company magazine article", "publisher": "Aramco", "locator": "25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-08-25", "event": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.", "source_ids": ["OCT14C-ARAMCO-CORE"]}], "field_sources": {"business_problem": ["OCT14C-ARAMCO-CORE"], "workflow": ["OCT14C-ARAMCO-CORE"], "technology_data": ["OCT14C-ARAMCO-CORE"], "deployment_scope": ["OCT14C-ARAMCO-CORE"], "economics": ["OCT14C-ARAMCO-CORE"]}}, {"packet_id": "oct14c-ai-072", "key": "oct14c-ai-072", "record_ids": ["AI-072"], "company": "ExxonMobil", "title": "Autonomous drilling in Guyana, with supplier-reported results", "category": "Internal adoption", "department": "Drilling & completions", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Existing operating profile; supplier-reported results now visible", "business_problem": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.", "workflow": "Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "technology_data": "Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "deployment_scope": "Undated case in the supplier’s 2026 URL directory. Exact start, observation period, well count and definition of the operations denominator are not supplied.", "integration_human_role": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.", "economics": "SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "practical_relevance": "How much well-delivery time changed after automation, and how often did shore teams intervene?", "assessment": "How much well-delivery time changed after automation, and how often did shore teams intervene?", "limitations": ["Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained."], "source_ids": ["SW7-014"], "sources": [{"id": "SW7-014", "title": "ExxonMobil Guyana: drilling automation in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "date": "Undated (2026 URL)", "type": "Vendor customer case", "publisher": "SLB", "locator": "Results: Autonomous drilling and rig automation drive record-setting performance in Guyana", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "Undated (2026 URL)", "event": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.", "source_ids": ["SW7-014"]}], "field_sources": {"business_problem": ["SW7-014"], "workflow": ["SW7-014"], "technology_data": ["SW7-014"], "deployment_scope": ["SW7-014"], "economics": ["SW7-014"]}}, {"packet_id": "oct14c-ai-148", "key": "oct14c-ai-148", "record_ids": ["AI-148"], "company": "Ecopetrol", "title": "An operating energy desk, with a dispatch model still being built", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-14", "verification": "verified primary", "stage": "CADE described in use; Rubiales dispatch-model MVP under construction in 2025", "business_problem": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI.", "workflow": "A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.", "technology_data": "A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.", "deployment_scope": "2025 operating and development activity, described in the 2026 annual-report edition. Exact publication day and MVP commissioning date are not established.", "integration_human_role": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given.", "economics": "No measured energy, emissions, reliability, cost or production benefit is disclosed for these AI components.", "practical_relevance": "Are Rubiales recommendations now used, who approves a setpoint, and how is the result compared with the previous dispatch method?", "assessment": "Are Rubiales recommendations now used, who approves a setpoint, and how is the result compared with the previous dispatch method?", "limitations": ["The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given."], "source_ids": ["OCT14C-ECOPETROL-CADE"], "sources": [{"id": "OCT14C-ECOPETROL-CADE", "title": "Ecopetrol Integrated Management Report 2025", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 (2025 annual report)", "type": "Operator annual report", "publisher": "Ecopetrol", "locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026 (2025 annual report)", "event": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI.", "source_ids": ["OCT14C-ECOPETROL-CADE"]}], "field_sources": {"business_problem": ["OCT14C-ECOPETROL-CADE"], "workflow": ["OCT14C-ECOPETROL-CADE"], "technology_data": ["OCT14C-ECOPETROL-CADE"], "deployment_scope": ["OCT14C-ECOPETROL-CADE"], "economics": ["OCT14C-ECOPETROL-CADE"]}}, {"packet_id": "oct14c-ai-143", "key": "oct14c-ai-143", "record_ids": ["AI-143"], "company": "PETRONAS", "title": "Deployed exploration tools and a separate equipment-development agreement", "category": "Internal adoption", "department": "Exploration & geoscience", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Earlier exploration tools deployed; new surface-equipment work remains development", "business_problem": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.", "workflow": "The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "technology_data": "The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "deployment_scope": "Company release dated 8 July 2026. Earlier exploration tools came from 2025 agreements; the third agreement was newly signed.", "integration_human_role": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.", "economics": "No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.", "practical_relevance": "What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?", "assessment": "What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?", "limitations": ["This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed."], "source_ids": ["OCT14C-PETRONAS-TRICIPTA"], "sources": [{"id": "OCT14C-PETRONAS-TRICIPTA", "title": "PETRONAS advances TriCipta AI through the third joint development agreement", "url": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "date": "2026-07-08", "type": "Company release", "publisher": "PETRONAS", "locator": "8 July 2026 release: third JDA scope and final two paragraphs on earlier tools", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-07-08", "event": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.", "source_ids": ["OCT14C-PETRONAS-TRICIPTA"]}], "field_sources": {"business_problem": ["OCT14C-PETRONAS-TRICIPTA"], "workflow": ["OCT14C-PETRONAS-TRICIPTA"], "technology_data": ["OCT14C-PETRONAS-TRICIPTA"], "deployment_scope": ["OCT14C-PETRONAS-TRICIPTA"], "economics": ["OCT14C-PETRONAS-TRICIPTA"]}}, {"packet_id": "oct14c-ai-097", "key": "oct14c-ai-097", "record_ids": ["AI-097"], "company": "Chord Energy", "title": "Rod-lift AI sits inside a broader artificial-lift improvement program", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-14", "verification": "verified primary", "stage": "Existing scaled AI workflow; broader operating results kept separate", "business_problem": "Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register.", "workflow": "The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.", "technology_data": "The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.", "deployment_scope": "2025 results, disclosed 25 February 2026. ESP comparisons start at the beginning of 2025; an exact matched AI evaluation period is not supplied.", "integration_human_role": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.", "economics": "The broader program reports over 50% improvement in ESP replacement cycle times, over 25% improvement in failure rates since early 2025, and about 1,200 fewer workover rig days in 2025. These are not separately isolated ML effects.", "practical_relevance": "What part of fewer failures and workovers comes from ML rather than equipment changes, maintenance and field practices?", "assessment": "What part of fewer failures and workovers comes from ML rather than equipment changes, maintenance and field practices?", "limitations": ["Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted."], "source_ids": ["OCT14C-CHORD-LIFT"], "sources": [{"id": "OCT14C-CHORD-LIFT", "title": "Chord fourth-quarter and full-year 2025 results", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "date": "2026-02-25", "type": "Operator earnings release", "publisher": "Chord Energy", "locator": "Operations Update / Production and LOE; two consecutive artificial-lift paragraphs", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-02-25", "event": "Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register.", "source_ids": ["OCT14C-CHORD-LIFT"]}], "field_sources": {"business_problem": ["OCT14C-CHORD-LIFT"], "workflow": ["OCT14C-CHORD-LIFT"], "technology_data": ["OCT14C-CHORD-LIFT"], "deployment_scope": ["OCT14C-CHORD-LIFT"], "economics": ["OCT14C-CHORD-LIFT"]}}, {"packet_id": "watch14pm-ai-294", "key": "watch14pm-ai-294", "record_ids": ["AI-294"], "company": "INPEX", "title": "Employees build agents for research, rules and management meetings", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Shared employee agents in use; active-agent population unknown", "business_problem": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.", "workflow": "SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "technology_data": "SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "deployment_scope": "Published 1 September 2026; Copilot became the core platform in 2025. Exact measurement windows, active-agent count, survey size and value calculation are not supplied.", "integration_human_role": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.", "economics": "Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.", "practical_relevance": "Which agents do people keep using, and does better research or meeting preparation change a decision?", "assessment": "Which agents do people keep using, and does better research or meeting preparation change a decision?", "limitations": ["Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method."], "source_ids": ["WATCH14PM-INPEX"], "sources": [{"id": "WATCH14PM-INPEX", "title": "INPEX transforms business processes with Microsoft 365 Copilot and AI agents", "url": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "date": "2026-09-01", "type": "Vendor customer story", "publisher": "Microsoft / named INPEX speakers", "locator": "Research tool; Building a foundation for change; Looking beyond estimated annual value", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-09-01", "event": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.", "source_ids": ["WATCH14PM-INPEX"]}], "field_sources": {"business_problem": ["WATCH14PM-INPEX"], "workflow": ["WATCH14PM-INPEX"], "technology_data": ["WATCH14PM-INPEX"], "deployment_scope": ["WATCH14PM-INPEX"], "economics": ["WATCH14PM-INPEX"]}}, {"packet_id": "watch14pm-ai-224", "key": "watch14pm-ai-224", "record_ids": ["AI-224"], "company": "Williams", "title": "A field technician turns equipment reports into turbine alerts", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Employee-built alert workflow described in use", "business_problem": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.", "workflow": "The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "technology_data": "The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "deployment_scope": "Published 9 September 2026; the existing profile retains Williams’s June company-wide adoption disclosure. The detector’s start date and asset count are unspecified.", "integration_human_role": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.", "economics": "More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.", "practical_relevance": "Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?", "assessment": "Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?", "limitations": ["Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share."], "source_ids": ["WATCH14PM-WILLIAMS"], "sources": [{"id": "WATCH14PM-WILLIAMS", "title": "Williams transforms frontline work with Copilot, gaining 90% adoption", "url": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "date": "2026-09-09", "type": "Vendor customer story", "publisher": "Microsoft / named Williams speakers", "locator": "Bringing AI to the frontline first; Accelerating operations with Copilot", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-09-09", "event": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.", "source_ids": ["WATCH14PM-WILLIAMS"]}], "field_sources": {"business_problem": ["WATCH14PM-WILLIAMS"], "workflow": ["WATCH14PM-WILLIAMS"], "technology_data": ["WATCH14PM-WILLIAMS"], "deployment_scope": ["WATCH14PM-WILLIAMS"], "economics": ["WATCH14PM-WILLIAMS"]}}, {"packet_id": "watch14pm-ai-158", "key": "watch14pm-ai-158", "record_ids": ["AI-158"], "company": "NOV", "title": "Egyptian drilling teams expand an edge-based warning system", "category": "Internal adoption", "department": "Drilling & completions", "report_date": "2026-10-14", "verification": "verified partner", "stage": "Vendor reports expansion from two rigs to a double-digit fleet", "business_problem": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.", "workflow": "Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "technology_data": "Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "deployment_scope": "Article published 11 March 2026. Egypt introduction dates to 2023; the source does not date each intervention or disclose exact rig/well counts.", "integration_human_role": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.", "economics": "NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.", "practical_relevance": "How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?", "assessment": "How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?", "limitations": ["Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established."], "source_ids": ["WATCH14PM-NOV"], "sources": [{"id": "WATCH14PM-NOV", "title": "NOV’s Drilling Beliefs & Analytics advances digital operations in Egypt", "url": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "date": "2026-03-11", "type": "Vendor deployment article", "publisher": "NOV", "locator": "11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}], "disclosed_results": [], "timeline": [{"date": "2026-03-11", "event": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.", "source_ids": ["WATCH14PM-NOV"]}], "field_sources": {"business_problem": ["WATCH14PM-NOV"], "workflow": ["WATCH14PM-NOV"], "technology_data": ["WATCH14PM-NOV"], "deployment_scope": ["WATCH14PM-NOV"], "economics": ["WATCH14PM-NOV"]}}, {"packet_id": "watch15am-ai-145", "key": "watch15am-ai-145", "record_ids": ["AI-145"], "company": "CNOOC", "title": "One oilfield connects reservoir, production and equipment data", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-15", "verification": "verified primary", "stage": "Operating integrated digital field; AI contribution not isolated", "business_problem": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.", "workflow": "More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.", "technology_data": "More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.", "deployment_scope": "2025 reporting period; report released 9 April 2026. Start date, observation periods and deployed user count are not specified. More than 26,000 points measures instrumentation, not people or wells.", "integration_human_role": "Automatic plan generation is reported; execution authority, review and safety limits are not disclosed.", "economics": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.", "practical_relevance": "The competitive advantage may come from connecting data and decisions across a field. Ask what changed in field work, how recommendations are checked, and which benefits remain after the cost of keeping the system reliable.", "assessment": "The competitive advantage may come from connecting data and decisions across a field. Ask what changed in field work, how recommendations are checked, and which benefits remain after the cost of keeping the system reliable.", "limitations": ["Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved."], "source_ids": ["WATCH15AM-CNOOC"], "sources": [{"id": "WATCH15AM-CNOOC", "title": "2025 Environmental, Social and Governance Report", "url": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "date": "2026-04-09", "publisher": "CNOOC Limited", "locator": "Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026", "access": "Parent retrieved original PDF and checked the relevant page on 15 October 2026; results remain source-reported."}], "disclosed_results": [{"metric": "Integrated field outcomes; not isolated AI benefits", "value": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.", "period": "2025 report; measurement period not supplied", "attribution": "CNOOC; source-reported", "source_ids": ["WATCH15AM-CNOOC"]}], "timeline": [{"date": "2026-04-09", "event": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.", "source_ids": ["WATCH15AM-CNOOC"]}], "field_sources": {"business_problem": ["WATCH15AM-CNOOC"], "workflow": ["WATCH15AM-CNOOC"], "technology_data": ["WATCH15AM-CNOOC"], "deployment_scope": ["WATCH15AM-CNOOC"], "integration_human_role": ["WATCH15AM-CNOOC"], "economics": ["WATCH15AM-CNOOC"]}}, {"packet_id": "watch17am-ai-279", "key": "watch17am-ai-279", "record_ids": ["AI-279"], "company": "Canadian Natural Resources", "title": "CNRL puts an annual savings figure on hydrogen optimization", "category": "Internal adoption", "department": "Refining, LNG & process operations", "report_date": "2026-10-17", "verification": "verified primary", "stage": "Operating use reported; site and start date unspecified", "business_problem": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.", "workflow": "The presentation names the task but does not explain the model, inputs, vendor or approval process.", "technology_data": "The presentation names the task but does not explain the model, inputs, vendor or approval process.", "deployment_scope": "Presentation dated 7 November 2025; exact posting and deployment dates unknown. File path is dated December 2025.", "integration_human_role": "Hydrogen approval and control rights are not disclosed. Do not transfer conveyor control rights to hydrogen optimization.", "economics": "The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.", "practical_relevance": "How much hydrogen was avoided, against which baseline, and what does the system cost to run?", "assessment": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.", "limitations": ["No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate."], "source_ids": ["WATCH17AM-CNRL"], "sources": [{"id": "WATCH17AM-CNRL", "title": "Investor Open House · 7 November 2025", "url": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "date": "2025-11-07", "publisher": "Canadian Natural Resources", "locator": "PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention", "review_status": "Original source read by editor", "supports": ["CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.", "The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI."], "accessed": "2026-10-17"}], "disclosed_results": [{"metric": "Company-reported annual hydrogen savings; not net ROI", "value": "Approximately CAD 8 million per year", "period": "Measurement period not disclosed", "attribution": "operator", "source_ids": ["WATCH17AM-CNRL"]}], "timeline": [{"date": "2025-11-07", "event": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.", "source_ids": ["WATCH17AM-CNRL"]}], "field_sources": {"business_problem": ["WATCH17AM-CNRL"], "workflow": ["WATCH17AM-CNRL"], "technology_data": ["WATCH17AM-CNRL"], "deployment_scope": ["WATCH17AM-CNRL"], "integration_human_role": ["WATCH17AM-CNRL"], "economics": ["WATCH17AM-CNRL"]}}, {"packet_id": "watch17am-ai-061", "key": "watch17am-ai-061", "record_ids": ["AI-061"], "company": "Cenovus", "title": "Governed productivity and decision-support pilots", "category": "Strategy only", "department": "IT, data & enterprise transformation", "report_date": "2026-10-17", "verification": "verified primary", "stage": "Governed pilot program; individual use cases not described", "business_problem": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "workflow": "It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.", "technology_data": "It is starting with productivity and decision-support pilots; individual applications, assets, models and vendors are not named.", "deployment_scope": "Program-level pilots described in the 2025 MD&A dated 18 February 2026; no facility, user count or go-live date supplied.", "integration_human_role": "Governed introduction is stated; specific human review and approval boundaries are not disclosed.", "economics": "No measured outcome, cost or ROI disclosed.", "practical_relevance": "This resolves whether the company disclosed AI pilots. It leaves open which business tasks reached regular use.", "assessment": "This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established.", "limitations": ["This verifies a program-level pilot disclosure, not a named deployment, user count, control boundary or operating result. The document is dated 18 February 2026; a separate posting date was not established."], "source_ids": ["S40"], "sources": [{"id": "S40", "title": "2025 Management’s Discussion and Analysis", "url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "date": "2026-02-18", "publisher": "Cenovus Energy / SEC", "locator": "Printed MD&A page 47: Technology, Information Systems and Data Privacy; page 2: document date", "review_status": "Original source read by editor", "supports": ["Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "No measured outcome, cost or ROI disclosed."], "accessed": "2026-10-17"}], "disclosed_results": [], "timeline": [{"date": "2026-02-18", "event": "Cenovus describes a governed introduction of AI, starting with productivity and decision-support pilots.", "source_ids": ["S40"]}], "field_sources": {"business_problem": ["S40"], "workflow": ["S40"], "technology_data": ["S40"], "deployment_scope": ["S40"], "integration_human_role": ["S40"], "economics": ["S40"]}}, {"packet_id": "watch18am-ai-295", "key": "watch18am-ai-295", "record_ids": ["AI-295"], "company": "Kuwait Oil Company", "title": "SANAD helps drilling engineers find the procedure they need", "category": "Internal adoption", "department": "Drilling & completions", "report_date": "2026-10-18", "verification": "verified primary", "stage": "Pilot and validation; later field test reported, wider rollout unverified", "business_problem": "Find relevant drilling procedures, well knowledge and lessons without relying on repeated manual searches.", "workflow": "The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision.", "technology_data": "LLM and retrieval-augmented generation over selected drilling procedures and records. SLB supported model training and data processing; KOC Innovation & Digitalization provided technical support. OpenWELL context was tested during the pilot; full integration is not established.", "deployment_scope": "Development reportedly began March 2025. KOC describes a 2025–26 pilot; its April–June 2026 issue has no verified posting day. The interview was published 29 September 2026. Field-test dates and well ID are not supplied. The pilot belonged to Drilling Group North & West Kuwait. The reported field test concerns one Lower Burgan production well; its identifier is undisclosed.", "integration_human_role": "KOC says its standards team validated every pilot output against official procedures. Engineering judgment and responsibility remain with the specialist. Retrieval and recommendations do not establish authority to control a rig.", "economics": "A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.", "practical_relevance": "Can engineers trace each answer to the right procedure, and does it improve decisions after review time and errors are included?", "assessment": "The useful step is a drilling answer grounded in the right procedure and checked by an engineer. The public evidence establishes a bounded pilot and an attributed later field test, not a company-wide brain or isolated AI savings.", "limitations": ["No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "SOP savings are not an isolated SANAD effect. No complete cost or net ROI established."], "source_ids": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "sources": [{"id": "WATCH18AM-KOC-DIGEST", "title": "SANAD AI: The Generative AI Drilling Assistant", "url": "https://www.kockw.com/sites/EN/Periodicals/Digest%20April%20-%20June%202026.pdf#page=28", "date": "2026-Q2", "publisher": "Kuwait Oil Company", "locator": "PDF pages 28–31 / printed 26–29; 14 SOPs on PDF 30; validation, ownership and roadmap on PDF 31", "access": "Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.", "excerpt": "Every output was validated", "review_status": "Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.", "accessed": "2026-10-18"}, {"id": "WATCH18AM-KOC-INTERVIEW", "title": "KOC’s Sanad uses AI to unlock drilling knowledge", "url": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "date": "2026-09-29", "publisher": "KUNA interview, republished by Kuwait Times", "locator": "Timestamp; lower-Burgan field test, SOP-related savings, RAG description and human responsibility", "access": "Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.", "excerpt": "Engineering judgment and responsibility remain with the specialist", "review_status": "Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.", "accessed": "2026-10-18"}], "disclosed_results": [], "timeline": [{"date": "2026-Q2", "event": "KOC Digest describes the 2025–26 pilot, 14 North Kuwait SOPs, standards-team validation and future modules. Exact posting day unknown.", "source_ids": ["WATCH18AM-KOC-DIGEST"]}, {"date": "2026-09-29", "event": "Attributed engineer interview reports a Lower Burgan field test. SOP-related savings do not isolate AI contribution.", "source_ids": ["WATCH18AM-KOC-INTERVIEW"]}], "field_sources": {"business_problem": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "workflow": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "technology_data": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "deployment_scope": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "integration_human_role": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "economics": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"], "practical_relevance": ["WATCH18AM-KOC-DIGEST", "WATCH18AM-KOC-INTERVIEW"]}}, {"packet_id": "watch20pm-ai-296", "key": "watch20pm-ai-296", "record_ids": ["AI-296"], "company": "U.S. Energy Development Corporation", "title": "Twelve wells moved from recommendations to automatic adjustments", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-20", "verification": "verified partner", "stage": "Operating field deployment; operator/vendor-reported results", "business_problem": "Well conditions change faster than an engineering team can manually review and execute every small ESP adjustment.", "workflow": "Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed.", "technology_data": "Physics-aware AI, live operating data, edge hardware and bounded ESP setpoint control. No language model identified.", "integration_human_role": "Engineers define operating envelopes and review exceptions. In-envelope adjustments execute automatically; exact override, shutdown and fallback procedures are undisclosed.", "deployment_scope": "12 Permian Basin wells; activated December 2025. Six-month evaluation against a 90-day pre-activation baseline; exact activation and publication days unknown.", "economics": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "practical_relevance": "Ask for the all-in cost and benefit bridge, the baseline calculation and the log of exceptions, overrides and rejected adjustments.", "assessment": "Shows how execution permissions and exception handling can matter as much as recommendation quality. Test the baseline and all-in economics before copying the workflow.", "limitations": ["Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case. The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion. Return definition, full initial/recurring costs, realized cash, adjustment for other operating changes and matched controls are undisclosed. The article says engineering work changed toward exceptions, not that headcount or total workload fell."], "source_ids": ["WATCH20PM-USEDC"], "partners": ["ROAM-AI"], "disclosed_results": [{"metric": "Author-reported production and return", "value": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "period": "First six months; production comparator is preceding 90 days", "attribution": "Operator/vendor coauthored; not independent", "source_ids": ["WATCH20PM-USEDC"]}], "timeline": [{"date": "2025-12", "event": "12-well activation, per operator coauthor.", "source_ids": ["WATCH20PM-USEDC"]}, {"date": "2026-07", "event": "July cover story; exact publication day not established.", "source_ids": ["WATCH20PM-USEDC"]}], "sources": [{"id": "WATCH20PM-USEDC", "title": "The Bottleneck Isn’t the Model. It’s the Execution.", "url": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "date": "2026-07", "publisher": "The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors", "locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "excerpt": "We activated the system across 12 Permian wells beginning in December 2025.", "access": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date.", "type": "Operator/vendor-coauthored field account", "review_status": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date."}], "field_sources": {"business_problem": ["WATCH20PM-USEDC"], "workflow": ["WATCH20PM-USEDC"], "technology_data": ["WATCH20PM-USEDC"], "integration_human_role": ["WATCH20PM-USEDC"], "deployment_scope": ["WATCH20PM-USEDC"], "economics": ["WATCH20PM-USEDC"]}}, {"packet_id": "watch20pm-ai-148", "key": "watch20pm-ai-148", "record_ids": ["AI-148"], "company": "Ecopetrol", "title": "Rubiales adds a dated operating footprint to CADE", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-20", "verification": "verified primary", "stage": "CADE operating coverage; specific AI-model commissioning unverified", "business_problem": "Choose the timing and mix of self-generation and purchased electricity for operating assets.", "workflow": "CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.", "technology_data": "2025 report describes analytics and GenAI in CADE, with a separate Rubiales dispatch-model MVP under construction. May 2026 release adds platform coverage, not model-specific commissioning.", "integration_human_role": "60 workers trained, not 60 verified active users. Source does not define approval, override or control permissions.", "deployment_scope": "Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. Issuer announcement May 4, 2026; Rubiales coverage reported from March 2026. Older evidence newly checked, not a new October launch.", "economics": "The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.", "practical_relevance": "Which decisions are actually automated, and what changes in delivered electricity cost after all platform and operating costs?", "assessment": "Trace whether an operating decision desk translates into lower delivered electricity cost. Capacity and a name implying autonomy cannot answer that question.", "limitations": ["No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.", "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given. May 2026 follow-up: No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population."], "source_ids": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "partners": ["ISA"], "disclosed_results": [], "timeline": [{"date": "2025", "event": "Rubiales model proof of concept and MVP construction described in earlier annual report.", "source_ids": ["OCT14C-ECOPETROL-CADE"]}, {"date": "2026-05-04", "event": "Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid.", "source_ids": ["WATCH20PM-CADE"]}], "sources": [{"id": "OCT14C-ECOPETROL-CADE", "title": "Ecopetrol Integrated Management Report 2025", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 (2025 annual report)", "publisher": "Ecopetrol", "type": "Operator annual report", "locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "review_status": "Relevant original source reviewed 14 October 2026; claims remain source-reported."}, {"id": "WATCH20PM-CADE", "title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "url": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "date": "2026-05-04", "publisher": "Ecopetrol", "type": "Operator announcement", "locator": "May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs", "review_status": "Full original operator page retrieved (HTTP200) and read by editor. Coverage and plans remain company-reported; no independent cost or performance validation."}], "field_sources": {"business_problem": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "workflow": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "technology_data": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "integration_human_role": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "deployment_scope": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"], "economics": ["OCT14C-ECOPETROL-CADE", "WATCH20PM-CADE"]}}, {"packet_id": "watch21am-ai-297", "key": "watch21am-ai-297", "record_ids": ["AI-297"], "company": "PTTEP", "title": "PTTEP describes an in-house engine for specialist agents", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-21", "verification": "verified primary", "stage": "Engine launch described; eight-agent rollout unverified", "business_problem": "Make company-specific knowledge and data available to specialist assistants across operating and corporate work.", "workflow": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live.", "technology_data": "In-house Enterprise AI Engine grounded in PTTEP knowledge and internal data. Model provider, retrieval stack, permissions and runtime architecture not disclosed.", "integration_human_role": "Analysis and recommendations are described. Approval, override and execution rights are undisclosed; no autonomous executive is established.", "deployment_scope": "Engine launch described; eight specialist agents planned across five domains. No active-user count, asset population or completed eight-agent rollout established.", "economics": "No X.brain benefit or full-cost return disclosed. DORA and LogSight appear separately in the same feature; their metrics are not credited to the engine.", "practical_relevance": "Ask which agents people actually use, what data they can reach and which decisions still need a named human approver.", "assessment": "An internal engine is a useful architectural direction. Ask for the agents people use and the decisions they improve before treating a launch as a working company brain.", "limitations": ["No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement."], "source_ids": ["WATCH21AM-PTTEP"], "partners": [], "disclosed_results": [], "timeline": [{"date": "2025-11-03", "event": "Operator feature describes launched engine and planned eight-agent scope; exact launch date unknown.", "source_ids": ["WATCH21AM-PTTEP"]}], "sources": [{"id": "WATCH21AM-PTTEP", "title": "PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy", "url": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "date": "2025-11-03", "publisher": "PTTEP", "locator": "Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples", "excerpt": "PTTEP has launched its in-house Enterprise AI Engine, “X.brain,”", "access": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.", "type": "Operator technology feature", "review_status": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation."}], "field_sources": {"business_problem": ["WATCH21AM-PTTEP"], "workflow": ["WATCH21AM-PTTEP"], "technology_data": ["WATCH21AM-PTTEP"], "integration_human_role": ["WATCH21AM-PTTEP"], "deployment_scope": ["WATCH21AM-PTTEP"], "economics": ["WATCH21AM-PTTEP"]}}, {"packet_id": "watch21extra-ai-086", "key": "watch21extra-ai-086", "record_ids": ["AI-086"], "company": "Devon Energy", "title": "Devon confirms ChatDVN is part of daily work", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-21", "verification": "verified primary", "stage": "Operating internal workflow; issuer-described daily use", "business_problem": "Reduce time spent finding company data and turn it into useful analysis.", "workflow": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.", "technology_data": "Firewalled internal AI tool and organized company data. Model provider, retrieval architecture and permissions are not disclosed.", "integration_human_role": "Engineers act on operating insights. The call does not establish ChatDVN execution rights, approval rules or any autonomous executive authority.", "deployment_scope": "Internal company-data workflow; user and asset populations undisclosed. Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch.", "economics": "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.", "practical_relevance": "Which jobs do employees finish through ChatDVN, and does the full review process improve time, accuracy and follow-through?", "assessment": "The issuer confirms an operating internal workflow. Ask what employees complete through the tool and measure quality and full review effort before crediting a productivity or cash benefit.", "limitations": ["Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status.", "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings."], "source_ids": ["FU001"], "partners": [], "disclosed_results": [], "timeline": [{"date": "2026-05-06", "event": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data.", "source_ids": ["FU001"]}], "sources": [{"id": "FU001", "title": "Devon Q1 2026 earnings call — internal ChatDVN workflow", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-05-06", "publisher": "Devon Energy / FactSet CallStreet", "locator": "PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI", "excerpt": "a fully-firewalled internal tool called ChatDVN", "access": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.", "type": "Issuer-hosted earnings transcript", "review_status": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation."}], "field_sources": {"business_problem": ["FU001"], "workflow": ["FU001"], "technology_data": ["FU001"], "integration_human_role": ["FU001"], "deployment_scope": ["FU001"], "economics": ["FU001"]}}, {"packet_id": "watch21pm-ai-052", "key": "watch21pm-ai-052", "record_ids": ["AI-052"], "company": "Quantum Capital Group", "title": "Quantum adds a relational layer to its deal-evaluation data", "category": "Business transformation foundation", "department": "Finance, investment & strategy", "report_date": "2026-10-21", "verification": "verified partner", "stage": "Vendor describes an adopted sponsor-level data layer", "business_problem": "Make geological, production and financial assumptions consistent across energy deal evaluation.", "workflow": "Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.", "technology_data": "Lakebase relational schemas, Postgres and PostGIS; Unity Catalog access controls and lineage; Lakeflow change-data capture; familiar SQL, Tableau and Spotfire tools.", "integration_human_role": "Engineers, analysts and deal teams query data. The source does not establish investment-approval rules, autonomous capital allocation or AI-agent execution rights.", "deployment_scope": "Quantum sponsor-level energy deal evaluation; asset and portfolio-company populations undisclosed. Undated source; adoption and evaluation periods unknown. Extends existing sponsor platform AI-052, separately from QB Energy’s field-planning workflow AI-054.", "economics": "Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.", "practical_relevance": "Can the team trace a deal assumption to its source and show whether shared data reduces reconciliation and review work?", "assessment": "Shared data can reduce inconsistent assumptions without proving AI value or investment performance. Test reconciliation, traceability and full review effort.", "limitations": ["No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution."], "source_ids": ["WATCH21PM-DBX"], "partners": ["Databricks"], "disclosed_results": [], "timeline": [{"date": "Unknown", "event": "Vendor describes adopted relational data layer; publication and adoption dates unknown.", "source_ids": ["WATCH21PM-DBX"]}], "sources": [{"id": "WATCH21PM-DBX", "title": "Transforming energy deal evaluation with Lakebase", "url": "https://www.databricks.com/customers/quantum-capital-group", "date": "Unknown", "publisher": "Databricks", "locator": "Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs", "excerpt": "The firm now runs over 11,000 unique data queries per month", "access": "Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.", "type": "Vendor customer story", "review_status": "Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation."}], "field_sources": {"business_problem": ["WATCH21PM-DBX"], "workflow": ["WATCH21PM-DBX"], "technology_data": ["WATCH21PM-DBX"], "integration_human_role": ["WATCH21PM-DBX"], "deployment_scope": ["WATCH21PM-DBX"], "economics": ["WATCH21PM-DBX"]}}, {"packet_id": "watch22pm-ai-298", "key": "watch22pm-ai-298", "record_ids": ["AI-298"], "company": "Vista Energy", "title": "Vista uses AI monitoring to adjust completion designs", "category": "Internal adoption", "department": "Drilling & completions", "report_date": "2026-10-22", "verification": "verified primary", "stage": "Operating since March 2025, per issuer", "business_problem": "Recognize frac hits while a completion is still underway and adjust the design before more stages are lost.", "workflow": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025.", "technology_data": "Proprietary AI monitoring for real-time completion-design decisions; no LLM or autonomous pumping authority established.", "integration_human_role": "Engineering decision support is described. The slide does not define human approval, override or execution rights.", "deployment_scope": "Vaca Muerta, Argentina; specific block, pad and well count undisclosed. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "economics": "Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.", "practical_relevance": "What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?", "assessment": "What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?", "limitations": ["No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted.", "2025 technical measures; matched baseline and denominator undisclosed."], "source_ids": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "partners": [], "disclosed_results": [{"metric": "Issuer-reported completion-stage measures", "value": "Less than 1% stages lost in 2025; one runaway stage per well avoided on average", "period": "2025; exact measurement period unknown", "attribution": "Issuer; not independently validated or causally isolated", "source_ids": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"]}], "timeline": [{"date": "2025-11-12", "event": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta.", "source_ids": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"]}], "sources": [{"id": "WATCH22PM-VISTA", "title": "Vista Investor Day 2025", "url": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "date": "2025-11-12", "publisher": "Vista Energy", "locator": "Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.", "excerpt": "Implemented since March 2025", "access": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated.", "type": "Company investor presentation", "review_status": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated."}, {"id": "WATCH22PM-VISTA-SEC", "title": "Vista Investor Day 2025 — SEC exhibit", "url": "https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm", "date": "2025-11-12", "publisher": "Vista Energy / SEC EDGAR", "locator": "Exhibit heading: November 12, 2025; same slide 19 presentation.", "excerpt": "Vista Investor Day", "access": "Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes.", "type": "Company investor presentation", "review_status": "Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes."}], "field_sources": {"business_problem": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "workflow": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "technology_data": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "integration_human_role": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "deployment_scope": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"], "economics": ["WATCH22PM-VISTA", "WATCH22PM-VISTA-SEC"]}}, {"packet_id": "watch22pm-ai-299", "key": "watch22pm-ai-299", "record_ids": ["AI-299"], "company": "Saipem", "title": "Saipem installs a predictive-maintenance pilot on one drillship", "category": "Internal adoption", "department": "Maintenance & reliability", "report_date": "2026-10-22", "verification": "verified primary", "stage": "Installed pilot on Saipem 12000; fleet extension planned", "business_problem": "Find emerging equipment problems early enough to plan maintenance on a drillship.", "workflow": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests.", "technology_data": "Equipment-data anomaly detection and failure prediction; no LLM or autonomous maintenance execution established.", "integration_human_role": "Maintenance planning is described. Approval, execution, override and fallback boundaries are undisclosed.", "deployment_scope": "Saipem 12000 ultra-deepwater drillship. Saipem 7000 is separate future context. Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.", "economics": "No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.", "practical_relevance": "Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?", "assessment": "Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?", "limitations": ["The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed.", "No pre-pilot reliability comparator or evaluation period disclosed."], "source_ids": ["WATCH22PM-SAIPEM"], "partners": ["ADC Energy"], "disclosed_results": [], "timeline": [{"date": "2026-02-23", "event": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it.", "source_ids": ["WATCH22PM-SAIPEM"]}], "sources": [{"id": "WATCH22PM-SAIPEM", "title": "Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000", "url": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "date": "2026-02-23", "publisher": "Saipem", "locator": "Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.", "excerpt": "This pilot project", "access": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified.", "type": "Company announcement", "review_status": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified."}], "field_sources": {"business_problem": ["WATCH22PM-SAIPEM"], "workflow": ["WATCH22PM-SAIPEM"], "technology_data": ["WATCH22PM-SAIPEM"], "integration_human_role": ["WATCH22PM-SAIPEM"], "deployment_scope": ["WATCH22PM-SAIPEM"], "economics": ["WATCH22PM-SAIPEM"]}}, {"packet_id": "watch22pm-ai-300", "key": "watch22pm-ai-300", "record_ids": ["AI-300"], "company": "Worley", "title": "Worley puts four internal tools into engineering and bid work", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-22", "verification": "verified primary", "stage": "Four recently deployed internal solutions, per issuer", "business_problem": "Reduce repeated work finding engineering knowledge, checking documents and preparing technical or commercial responses.", "workflow": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest.", "technology_data": "Internal AI-assisted document review, search, comparison and drafting. Exact model architecture and execution permissions undisclosed.", "integration_human_role": "Tools support review and drafting. Named approval roles and execution rights are undisclosed.", "deployment_scope": "Internal engineering and commercial work; no client asset, basin or user population disclosed. FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.", "economics": "Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.", "practical_relevance": "Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?", "assessment": "Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?", "limitations": ["These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established.", "No tool-level before/after or quality-controlled comparator disclosed."], "source_ids": ["WATCH22PM-WORLEY"], "partners": [], "disclosed_results": [], "timeline": [{"date": "2025-08-27", "event": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting.", "source_ids": ["WATCH22PM-WORLEY"]}], "sources": [{"id": "WATCH22PM-WORLEY", "title": "FY2025 full-year results presentation", "url": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "date": "2025-08-27", "publisher": "Worley", "locator": "Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.", "excerpt": "Recently deployed solutions through our Advanced Development Lab", "access": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes.", "type": "Company investor presentation", "review_status": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes."}], "field_sources": {"business_problem": ["WATCH22PM-WORLEY"], "workflow": ["WATCH22PM-WORLEY"], "technology_data": ["WATCH22PM-WORLEY"], "integration_human_role": ["WATCH22PM-WORLEY"], "deployment_scope": ["WATCH22PM-WORLEY"], "economics": ["WATCH22PM-WORLEY"]}}, {"packet_id": "watch22pm-ai-199", "key": "watch22pm-ai-199", "record_ids": ["AI-199"], "company": "Technip Energies", "title": "Technip Energies reports progress, without proving 55 AI deployments", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-22", "verification": "verified primary excerpt", "stage": "Mixed digital and AI program; end-2025 output counts", "business_problem": "Move useful digital and AI ideas into working tools and train staff to use them.", "workflow": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established.", "technology_data": "Mixed digital and AI initiatives and staff upskilling; no inference of 55 AI production deployments.", "integration_human_role": "Program and upskilling outputs disclosed; project-level approval and execution boundaries unknown.", "deployment_scope": "Company transformation program; project and asset populations undisclosed. 2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.", "economics": "The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.", "practical_relevance": "Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?", "assessment": "Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?", "limitations": ["Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved.", "End-2025 program output counts, not a benefit baseline."], "source_ids": ["WATCH22PM-TEN"], "partners": [], "disclosed_results": [], "timeline": [{"date": "2026 (exact publication day unverified)", "event": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025.", "source_ids": ["WATCH22PM-TEN"]}], "sources": [{"id": "WATCH22PM-TEN", "title": "Technip Energies 2025 Annual Report", "url": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "date": "2026 (exact publication day unverified)", "publisher": "Technip Energies", "locator": "Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.", "excerpt": "more than 55 digital and AI initiatives", "access": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date.", "type": "Indexed original annual-report excerpt", "review_status": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date."}], "field_sources": {"business_problem": ["WATCH22PM-TEN"], "workflow": ["WATCH22PM-TEN"], "technology_data": ["WATCH22PM-TEN"], "integration_human_role": ["WATCH22PM-TEN"], "deployment_scope": ["WATCH22PM-TEN"], "economics": ["WATCH22PM-TEN"]}}, {"packet_id": "watch23am-ai-301", "key": "watch23am-ai-301", "record_ids": ["AI-301"], "company": "Azule Energy", "title": "Azule connects the well plan to drilling automation", "category": "Internal adoption", "department": "Drilling & completions", "report_date": "2026-10-23", "verification": "verified vendor", "stage": "Operating single-well example; wider rollout unverified", "business_problem": "Carry an engineering plan into rig execution with fewer manual transfers.", "workflow": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola. DrillPlan passes the plan into DrillOps automation. The supplier case identifies AI-driven operations; it does not isolate AI from engineering automation.", "technology_data": "AI-enabled DrillOps execution alongside digital well planning; underlying algorithms and model type undisclosed.", "integration_human_role": "Autonomous section execution is reported. Approval, override, stop and fallback boundaries are not disclosed.", "deployment_scope": "One offshore Angola well; no named field or basin disclosed. Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown.", "economics": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "practical_relevance": "Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?", "assessment": "Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?", "limitations": ["No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate.", "Planning before/after and an unspecified field ROP benchmark; no matched controlled comparison."], "source_ids": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "partners": ["SLB"], "disclosed_results": [{"metric": "Supplier-reported planning and section performance", "value": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "period": "Single-well example; exact dates unknown", "attribution": "SLB supplier account; not independently validated or AI-only", "source_ids": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"]}], "timeline": [{"date": "2026-10-23", "event": "Q3 release describes the named example or contract; exact operating/award dates unknown.", "source_ids": ["WATCH23AM-SLB-Q3"]}], "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / Azule Energy paragraph; same case and key measures.", "excerpt": "cut planning time from 192 to 31 hours", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf", "type": "Supplier earnings release / customer account", "review_status": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation."}, {"id": "WATCH23AM-AZULE", "title": "Azule Energy pioneers digital well planning and rig automation in Africa", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa", "date": "Unknown", "publisher": "SLB", "locator": "Location; Results: Achieving operational excellence; Products Used: DrillOps AI-driven workflows.", "excerpt": "digitally ingested the well plan", "access": "Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration.", "type": "Supplier earnings release / customer account", "review_status": "Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration."}], "field_sources": {"business_problem": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "workflow": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "technology_data": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "integration_human_role": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "deployment_scope": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"], "economics": ["WATCH23AM-SLB-Q3", "WATCH23AM-AZULE"]}}, {"packet_id": "watch23am-ai-302", "key": "watch23am-ai-302", "record_ids": ["AI-302"], "company": "Brunei Shell Petroleum", "title": "BSP awards an on-premise software contract", "category": "Internal adoption", "department": "IT, data & enterprise transformation", "report_date": "2026-10-23", "verification": "verified vendor", "stage": "Contract awarded; go-live and usage unverified", "business_problem": "Connect subsurface and production software with a consistent implementation and support arrangement.", "workflow": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work. Analyst takeaway: a broad contract establishes buying intent and implementation scope; it does not show which AI tasks employees complete.", "technology_data": "Contract includes AI capabilities; no particular live model or LLM workflow established.", "integration_human_role": "No project-level AI approval or execution boundaries disclosed.", "deployment_scope": "BSP, Brunei; subsurface and production scope, no specific asset or user count. Published 23 October 2026. Award date, implementation timetable and active-use dates unknown.", "economics": "No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.", "practical_relevance": "Which workflows will go live first, and what acceptance test separates installed software from useful work?", "assessment": "Which workflows will go live first, and what acceptance test separates installed software from useful work?", "limitations": ["Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout.", "No live-use or outcome baseline established."], "source_ids": ["WATCH23AM-SLB-Q3"], "partners": ["SLB"], "disclosed_results": [], "timeline": [{"date": "2026-10-23", "event": "Q3 release describes the named example or contract; exact operating/award dates unknown.", "source_ids": ["WATCH23AM-SLB-Q3"]}], "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / Brunei BSP software contract paragraph.", "excerpt": "on-premises software contract", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf", "type": "Supplier earnings release / customer account", "review_status": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation."}], "field_sources": {"business_problem": ["WATCH23AM-SLB-Q3"], "workflow": ["WATCH23AM-SLB-Q3"], "technology_data": ["WATCH23AM-SLB-Q3"], "integration_human_role": ["WATCH23AM-SLB-Q3"], "deployment_scope": ["WATCH23AM-SLB-Q3"], "economics": ["WATCH23AM-SLB-Q3"]}}, {"packet_id": "watch23pm-ai-303", "key": "watch23pm-ai-303", "record_ids": ["AI-303"], "company": "Petrobras", "title": "AI checks supplier documents. Petrobras staff make the decision.", "category": "Internal adoption", "department": "Supply chain & logistics", "report_date": "2026-10-23", "verification": "verified primary", "stage": "Wave 2 pilot; rollout unconfirmed", "business_problem": "Check supplier evidence during registration.", "workflow": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them. The evaluation team records its decision and feeds it back to the AI. Petronect handles registration; the notice does not identify the AI model or developer.", "technology_data": "Preliminary document analysis; model type and developer undisclosed.", "integration_human_role": "Technical evaluators perform the final check and retain the decision.", "deployment_scope": "Petrobras corporate supplier registration; no producing asset or basin applies. Participating supplier volume is undisclosed. Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.", "economics": "No accuracy, workload, completed-case count, cost, saving or ROI reported.", "practical_relevance": "Does the first AI pass reduce total review work while preserving the quality of supplier decisions?", "assessment": "Does the first AI pass reduce total review work while preserving the quality of supplier decisions?", "limitations": ["Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.", "No before/after comparison or measurement period disclosed."], "source_ids": ["WATCH23PM-PETROBRAS"], "partners": ["Petronect (registration portal, not established AI developer)"], "disclosed_results": [], "timeline": [{"date": "Unknown", "event": "Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.", "source_ids": ["WATCH23PM-PETROBRAS"]}], "sources": [{"id": "WATCH23PM-PETROBRAS", "title": "Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice", "url": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "date": "Unknown", "publisher": "Petrobras", "locator": "Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.", "excerpt": "A supervisão e a decisão final permanecem sempre com nossos avaliadores técnicos.", "access": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied.", "type": "Company supplier-channel pilot notice", "review_status": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied."}], "field_sources": {"business_problem": ["WATCH23PM-PETROBRAS"], "workflow": ["WATCH23PM-PETROBRAS"], "technology_data": ["WATCH23PM-PETROBRAS"], "integration_human_role": ["WATCH23PM-PETROBRAS"], "deployment_scope": ["WATCH23PM-PETROBRAS"], "economics": ["WATCH23PM-PETROBRAS"]}}, {"packet_id": "watch24am-ai-304", "key": "watch24am-ai-304", "record_ids": ["AI-304"], "company": "Halliburton", "title": "A maintenance alert only helps if a crew can act on it.", "category": "Commercial AI product", "department": "Maintenance & reliability", "report_date": "2026-10-24", "verification": "verified vendor", "stage": "Supplier reports four deployments; customers unidentified", "business_problem": "Find developing equipment problems early enough for a useful maintenance intervention.", "workflow": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs. Sensors feed more than 400 equipment parameters into cloud models. Staff check warnings, arrange parts and schedule corrective work. The product page separately describes transferring approved job designs from VIDA and iCem into controllers; this is not proof of autonomous maintenance.", "technology_data": "Statistical models and machine learning for equipment condition; no LLM identified.", "integration_human_role": "Crews investigate alerts, obtain parts and schedule repairs; authorization, override and fallback rules are not specified.", "deployment_scope": "Cementing equipment; four deployments reported, with no identified basin, asset or customer. Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.", "economics": "The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.", "practical_relevance": "Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?", "assessment": "Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?", "limitations": ["No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "No comparator, productivity definition or cost model disclosed."], "source_ids": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "partners": ["Halliburton"], "disclosed_results": [], "timeline": [{"date": "2026-10-22", "event": "Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.", "source_ids": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"]}], "sources": [{"id": "WATCH24AM-HAL-RELIABILITY", "title": "How predictive intelligence reduces downtime and improves equipment reliability", "url": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "date": "2026-10-22", "publisher": "Halliburton", "locator": "A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.", "excerpt": "helped teams avoid an estimated $125,000 to $150,000 in downtime costs", "access": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided.", "type": "Supplier product account", "review_status": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided."}, {"id": "WATCH24AM-HAL-PRODUCT", "title": "LOGIX Unit Vitality — cement equipment health monitoring", "url": "https://www.halliburton.com/en/products/logix-unit-vitality", "date": "Unknown", "publisher": "Halliburton", "locator": "Automated cement equipment health assessment; AI-driven models; Real-time monitoring closes the loop.", "excerpt": "more than 400 real-time parameters", "access": "Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments.", "type": "Supplier product account", "review_status": "Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments."}], "field_sources": {"business_problem": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "workflow": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "technology_data": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "integration_human_role": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "deployment_scope": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"], "economics": ["WATCH24AM-HAL-RELIABILITY", "WATCH24AM-HAL-PRODUCT"]}}, {"packet_id": "slb-data-ai-thesis", "key": "slb-data-ai-thesis", "record_ids": ["AI-009"], "company": "SLB", "title": "What holds up in SLB’s data and AI investment thesis?", "category": "Commercial AI product", "department": "IT, data & enterprise transformation", "report_date": "2026-10-24", "verification": "verified primary", "stage": "Existing commercial platform; investor valuation argument remains unproven", "business_problem": "Evaluate whether SLB can convert specialized data, technical software and customer reach into durable incremental AI earnings.", "workflow": "Menon describes a seismic model trained on public and SLB multiclient datasets, with customers able to fine-tune for their own assets. Lumi connects governed data; domain models handle technical formats; Tela provides interaction and workflow orchestration. SLB describes subscription upgrades, metered usage, tailored workflows and marketplace revenue sharing. These are commercial mechanisms, not disclosed AI-only earnings.", "technology_data": "Technical domain models plus general LLMs and agents; not a single text model trained without restriction on all downhole data.", "integration_human_role": "Presentation describes engineering recommendations under human control and more autonomous workflows under physical constraints. It does not establish one common approval boundary for every customer deployment.", "deployment_scope": "Global Digital business; named operating cases retain their own asset and evidence boundaries. Original investor-day scripts: 17 June 2026. FY2025 baseline released 23 January 2026; Q3 results released 23 October 2026. Terms dates unknown. Investor interview framing published 20 October 2026.", "economics": "FY2025 Digital revenue was $2.660bn and adjusted EBITDA $928m (34.9% margin). Q3 2026 Digital revenue was $693m, adjusted EBITDA $228m and ARR $1.04bn. These are division-wide figures, not AI revenue or customer ROI. The CFO’s 2030 $1.8–2.0bn adjusted EBITDA and approximately $2bn ARR are ambitions, not results.", "practical_relevance": "SLB has a credible route to earning more from AI: it can put specialized models into software and operating workflows that customers already use. The public evidence supports a platform, engineering and governed-access advantage. It does not support treating every customer’s downhole data as an unrestricted SLB-owned training pool.", "assessment": "Ask for AI-specific paid usage and retention, incremental profit after compute and support, training rights by data category, held-out-basin model performance, customer-specific control boundaries, and a valuation using post-elimination earnings. For the full evidence map and interactive revenue chart, read the SLB briefing linked from the homepage.", "limitations": ["SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "Segment accounts are not AI-only economics. No pooled value or return estimate is admitted.", "Full source-linked briefing: https://energyai.sunya.ai/assets/slb-data-ai.html"], "source_ids": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "partners": ["SLB and its customer/technology ecosystem"], "disclosed_results": [], "timeline": [{"date": "2026-06-17", "event": "Management explains training provenance, commercial mechanisms and 2030 ambitions.", "source_ids": ["SLBTH-MENON", "SLBTH-BIGUET"]}, {"date": "2026-10-23", "event": "Q3 segment results update; not AI-only revenue.", "source_ids": ["SLBTH-Q3"]}], "sources": [{"id": "SLBTH-Q3", "title": "SLB third-quarter 2026 results — Digital economics and category definitions", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "date": "2026-10-23", "publisher": "SLB", "locator": "Digital results; Supplementary Information questions 10–13; Digital customer highlights.", "excerpt": "$693 million decreased 1% sequentially", "type": "Supplier earnings release", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-FY25", "title": "SLB full-year 2025 results — Digital financial baseline", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "date": "2026-01-23", "publisher": "SLB", "locator": "Supplementary Information question 12, Digital FY2025 table.", "excerpt": "Digital revenue increased 9% on a full-year basis", "type": "Supplier earnings release", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-MENON", "title": "Shashi Menon: The Race to Scale Digital Operations and AI", "url": "https://investorcenter.slb.com/static-files/cf773c3d-348f-479f-aeb0-81f5418844ca", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF pp. 3–5: data access and control; p. 11: seismic training; pp. 14–15: monetization.", "excerpt": "we trained it with public and SLB multiclient seismic data sets", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-BIGUET", "title": "Stephane Biguet: Proven Growth. Accretive Returns.", "url": "https://investorcenter.slb.com/static-files/0835f6dc-d2fb-4bc0-b85e-afecadcf36fe", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF p. 4: customer base; pp. 7–10: commercial models; p. 11: 2030 ambitions.", "excerpt": "more than 1,500 customers, including 90 of the world’s top 100 oil and gas producers", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-TERMS", "title": "SLB Online Services Terms and Conditions", "url": "https://www.software.slb.com/slb-online-services-terms-and-conditions", "date": "Unknown", "publisher": "SLB", "locator": "Definitions: Customer Data and Interaction Data; sections 8(a)–(b), 9(a), 10(a).", "excerpt": "SLB claims no ownership rights in Customer Data.", "type": "Published online-service terms", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement."}, {"id": "SLBTH-GENAI", "title": "SLB Online Services Service Specific Terms", "url": "https://www.software.slb.com/slb-online-services-service-specific-terms", "date": "Unknown", "publisher": "SLB", "locator": "Generative AI Services: definitions; Connected Gen AI A(i), A(iv); Embedded Gen AI B(iii).", "excerpt": "SLB shall not use Customer Prompts or Output to train models or further develop the Online Services without Customer’s consent.", "type": "Published online-service terms", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement."}, {"id": "SLBTH-MULLIN", "title": "Robert Mullin interview: 3.5% of the S&P, 20% of the Cash", "url": "https://georgenoble.substack.com/p/35-of-the-s-and-p-20-of-the-cash", "date": "2026-10-20", "publisher": "The Noble Update", "locator": "Host introduction and linked interview; exact SLB interview timestamp not established.", "excerpt": "The oil service giant hiding a real AI business that could be worth more than the whole company", "type": "Investor opinion / interview", "access": "Original host introduction reviewed; confirms the investment framing and interview identity. Full interview passage and exact timestamp not independently checked. An Eisman episode containing the same argument remains unconfirmed.", "review_status": "Original host introduction reviewed; confirms the investment framing and interview identity. Full interview passage and exact timestamp not independently checked. An Eisman episode containing the same argument remains unconfirmed."}, {"id": "SLBTH-RANNEN", "title": "Trygve Randen: Our Digital Advantage", "url": "https://investorcenter.slb.com/static-files/7aa731c6-2f38-47e4-9aa2-473a8d04ec6e", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF pp. 2–3: governed platforms; pp. 5–6: hosting and trust; p. 7: operational footprint.", "excerpt": "more than ninety percent of global production is simulated or modelled using at least one of our digital solutions", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation.", "review_status": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}], "field_sources": {"business_problem": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "workflow": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "technology_data": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "integration_human_role": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "deployment_scope": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"], "economics": ["SLBTH-Q3", "SLBTH-FY25", "SLBTH-MENON", "SLBTH-BIGUET", "SLBTH-TERMS", "SLBTH-GENAI", "SLBTH-MULLIN", "SLBTH-RANNEN"]}}, {"packet_id": "watch24pm-ai-305", "key": "watch24pm-ai-305", "record_ids": ["AI-305"], "company": "Petroleum Development Oman", "title": "PDO has moved ESP prediction from pilot to operating use", "category": "Internal adoption", "department": "Production & field operations", "report_date": "2026-10-24", "verification": "verified vendor", "stage": "Operating deployment, supplier-reported; scale unknown", "business_problem": "Help production teams anticipate electric submersible pump failures and estimate remaining run life.", "workflow": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework. The on-premise installation provides failure forecasts, run-life estimates and real-time pump-performance insight while meeting the customer’s cybersecurity requirements. The release does not say whether forecasts trigger automatic changes or how people review alerts and approve maintenance.", "technology_data": "PetroVisor machine-learning framework supports pump failure/run-life prediction. No LLM or autonomous pump control established.", "integration_human_role": "Human review, override, alert triage and maintenance approval rules are undisclosed; do not infer autonomous control.", "deployment_scope": "Oman; exact field, basin, pump population and operating period undisclosed. Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.", "economics": "No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.", "practical_relevance": "Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.", "assessment": "Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.", "limitations": ["A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.", "No counterfactual, baseline or result denominator disclosed."], "source_ids": ["WATCH24PM-WFRD-Q1"], "partners": ["Weatherford", "PetroVisor"], "disclosed_results": [], "timeline": [{"date": "2026-04-21", "event": "Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.", "source_ids": ["WATCH24PM-WFRD-Q1"]}], "sources": [{"id": "WATCH24PM-WFRD-Q1", "title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "date": "2026-04-21", "publisher": "Weatherford", "locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "excerpt": "moving the technology from pilot to operational use", "access": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown.", "type": "Supplier earnings release", "review_status": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown."}], "field_sources": {"business_problem": ["WATCH24PM-WFRD-Q1"], "workflow": ["WATCH24PM-WFRD-Q1"], "technology_data": ["WATCH24PM-WFRD-Q1"], "integration_human_role": ["WATCH24PM-WFRD-Q1"], "deployment_scope": ["WATCH24PM-WFRD-Q1"], "economics": ["WATCH24PM-WFRD-Q1"]}}, {"packet_id": "watch25am-ai-306", "key": "watch25am-ai-306", "record_ids": ["AI-306"], "company": "PTTEP", "title": "PTTEP reports faster purchase-order processing with X.Bot", "category": "Internal adoption", "department": "Supply chain & logistics", "report_date": "2026-10-25", "verification": "verified operator", "stage": "Operating procurement workflow; scale undisclosed", "business_problem": "Purchase-order processing and a backlog of pending orders; underlying task bottlenecks are not described.", "workflow": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing. The annual report describes improved processing, decision support and compliance. It does not identify model, vendor, user population or approval permissions.", "technology_data": "AI plus automation; no LLM, autonomous purchasing or X.brain integration established.", "integration_human_role": "Who approves or releases purchase orders, reviews the output or overrides it is not disclosed.", "deployment_scope": "Procurement process; geography, business-unit and supplier coverage unknown. Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.", "economics": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "practical_relevance": "Which steps changed, who approves orders, and do the time gains persist after all operating costs?", "assessment": "Which steps changed, who approves orders, and do the time gains persist after all operating costs?", "limitations": ["No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "Average per-PO time and six-month backlog change; baseline values and cohort unknown."], "source_ids": ["WATCH25AM-PTTEP-XBOT", "WATCH25AM-PTTEP-OPEX"], "partners": [], "disclosed_results": [{"metric": "Purchase-order process measures", "value": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "period": "Evaluation dates undisclosed; backlog change within six months", "attribution": "Operator-reported; AI and conventional automation not isolated", "source_ids": ["WATCH25AM-PTTEP-XBOT"]}], "timeline": [{"date": "Unknown", "event": "Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.", "source_ids": ["WATCH25AM-PTTEP-XBOT", "WATCH25AM-PTTEP-OPEX"]}], "sources": [{"id": "WATCH25AM-PTTEP-XBOT", "title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "Unknown", "publisher": "PTTEP", "locator": "Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate.", "excerpt": "average processing time by five hours per purchase order", "access": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "format": "pdf", "type": "Operator annual report", "review_status": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes."}, {"id": "WATCH25AM-PTTEP-OPEX", "title": "PTTEP organizes OpEx Day 2025, advancing operational excellence", "url": "https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence", "date": "2025-07-09", "publisher": "PTTEP", "locator": "Press release, procurement AI sentence in operational-excellence examples.", "excerpt": "application of AI to streamline procurement processes", "access": "Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.", "type": "Operator press release", "review_status": "Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results."}], "field_sources": {"business_problem": ["WATCH25AM-PTTEP-XBOT"], "workflow": ["WATCH25AM-PTTEP-XBOT"], "technology_data": ["WATCH25AM-PTTEP-XBOT"], "integration_human_role": ["WATCH25AM-PTTEP-XBOT"], "deployment_scope": ["WATCH25AM-PTTEP-XBOT"], "economics": ["WATCH25AM-PTTEP-XBOT"]}}], "coverage": [{"company": "Chevron", "existing_ids": ["AI-001", "AI-002", "AI-003", "AI-004", "AI-005", "AI-074", "AI-075"], "searched": true, "queries": ["Chevron AI 2024", "Chevron AI 2025", "Chevron AI 2026 operations"], "documents_reviewed": [{"id": "majors-001", "title": "A smarter way to prospect for oil and gas — Chevron ApEX", "url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "publisher": "Chevron", "date": "2025-11-12", "type": "company announcement", "locator": "How does ApEX work? / Advantages in exploration", "review_status": "full text reviewed", "supports": ["ApEX launched in August 2024 and was first used for deepwater exploration in the Gulf of America.", "The in-house multi-agent GenAI system searches more than one million files (4 TB) for explorationists, combining data from surveys, well logs, maps and technical reports.", "Use supports prospect screening and investment decisions by making subsurface materials easier to retrieve; experts remain the users and decision-makers."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Official ApEX update reviewed. Existing APOLO and exploration/workforce records share Investor Day source but that PDF was not reopened in this pass; AI-005 is AI-demand power infrastructure, and AI-075 vendor-supported inspection claims need separate review.", "cohort": "majors"}, {"company": "ExxonMobil", "existing_ids": ["AI-022", "AI-023", "AI-024", "AI-072", "AI-073"], "searched": true, "queries": ["ExxonMobil AI 2024", "ExxonMobil AI 2025", "ExxonMobil AI 2026 operations"], "documents_reviewed": [{"id": "majors-003", "title": "Digital technologies — ExxonMobil", "url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "publisher": "ExxonMobil", "date": "2026-03-27", "type": "company announcement", "locator": "Autonomous drilling", "review_status": "full text reviewed", "supports": ["ExxonMobil describes a proprietary AI drilling advisory system in use in Guyana that determines drilling parameters and enables closed-loop control without human intervention."], "accessed": "2026-09-27"}, {"id": "majors-004", "title": "How ExxonMobil is unleashing the power of open data", "url": "https://corporate.exxonmobil.com/locations/asia-pacific-region/ap-newsroom/2026/unleashing-the-power-of-open-data", "publisher": "ExxonMobil Malaysia", "date": "2026-07-10", "type": "company announcement", "locator": "Real-world wins", "review_status": "full text reviewed", "supports": ["ExxonMobil says OSDU-based data standardization has shortened evaluation times for deepwater assets including Guyana; platform is a foundation, not itself an AI application."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Reviewed current ExxonMobil digital technology page and 2026 OSDU account; AI-072/073 specific 2026 claims not verified against matching primary releases; existing transcript assertions not reopened.", "cohort": "majors"}, {"company": "Shell", "existing_ids": ["AI-025", "AI-026", "AI-076"], "searched": true, "queries": ["Shell AI 2024", "Shell AI 2025", "Shell AI 2026 operations"], "documents_reviewed": [{"id": "majors-018", "title": "C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations", "url": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "publisher": "C3 AI", "date": "2026-06-04", "accessed": "2026-09-27", "type": "partner case study", "locator": "News Details, lines 285–297", "review_status": "full text reviewed", "supports": ["C3 AI says it has worked with Shell since 2018 on predictive maintenance monitoring over 13,000 equipment items.", "A new multi-year agreement announced June 2026 extends reliability AI and introduces agent-based root cause analysis and remediation; the agreement is prospective.", "Program uses C3 AI Reliability and C3 Agentic AI Platform on Microsoft Azure.", "Hundreds of millions of dollars value is a C3 AI executive claim, not independent or Shell-attributed validation."]}, {"id": "majors-019", "title": "Safety at Pearl: Every second counts, every role matters", "url": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "publisher": "Shell", "date": "2025-09", "accessed": "2026-09-27", "type": "company announcement", "locator": "Being smart about safety", "review_status": "full text reviewed", "supports": ["Shell says Pearl GTL uses machine vision to flag unsafe behaviors and hazardous conditions in real time and alert supervisors.", "Pearl uses AI for near-real-time equipment health insights to reduce manual inspections and support earlier intervention.", "The plant was recognized by the World Economic Forum Global Lighthouse Network in September 2025."]}, {"id": "majors-022", "title": "Artificial Intelligence — Shell Global", "url": "https://www.shell.com/what-we-do/digitalisation/artificial-intelligence.html", "publisher": "Shell", "date": null, "accessed": "2026-09-27", "type": "company announcement", "locator": "Perdido pump AI / FAQ", "review_status": "full text reviewed", "supports": ["Shell official page describes AI at Perdido identifying a chemical signature ahead of 70% of pump disruptions with field operator input.", "Page publication date is not stated; not counted as in-window initiative."]}], "status": "partial", "gaps": "Partner announcement verifies AI-076 with vendor evidence; official Shell 2025 story verifies Pearl GTL safety and equipment-health AI. Official AI page reviewed but its Perdido workflow is undated. Existing AI-025/026 were not reconciled to a 2025 source in this pass.", "cohort": "majors"}, {"company": "bp", "existing_ids": [], "searched": true, "queries": ["bp AI 2024", "bp AI 2025", "bp AI 2026 operations"], "documents_reviewed": [{"id": "majors-002", "title": "bp Sustainability Report 2024", "url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "publisher": "bp", "date": "2025-03-11", "type": "annual report", "locator": "Advancing safety through AI, bpx energy", "review_status": "excerpt reviewed", "supports": ["In 2024 bpx energy implemented AI technology across Permian operations to assess driving hazards and risks in real time and provide in-vehicle alerts for preventative action."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "One qualifying driver-safety workflow verified from 2024 sustainability report. Broader bp/bpx AI use cases need dedicated 2025/2026 source search.", "cohort": "majors"}, {"company": "TotalEnergies", "existing_ids": ["AI-030", "AI-077", "AI-078"], "searched": true, "queries": ["TotalEnergies AI 2024", "TotalEnergies AI 2025", "TotalEnergies AI 2026 operations"], "documents_reviewed": [{"id": "majors-005", "title": "Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI", "url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "publisher": "TotalEnergies / Cognite", "date": "2025-09-26", "type": "company announcement", "locator": "PDF pp. 1–2", "review_status": "full text reviewed", "supports": ["Three-year rollout of Cognite industrial data and AI platform across all TotalEnergies-operated upstream assets worldwide, drilling through production.", "The initiative contextualizes industrial data, supports dynamic asset visualization and monitoring of critical equipment, and aims to accelerate applications."], "accessed": "2026-09-27"}, {"id": "majors-006", "title": "2025 Results and 2026 Objectives Presentation", "url": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf", "publisher": "TotalEnergies", "date": "2026-02-18", "type": "investor presentation", "locator": "PDF slides 28–29", "review_status": "full text reviewed", "supports": ["Reported data platform footprint: 40 upstream, 16 refining and chemicals, and 500 integrated power sites; platform captures 10x real-time data points and contextualizes 35 million points.", "The deck names three AI strategic programs and a new global competency center in India, with target production and availability uplift but no quantified realized uplift."], "accessed": "2026-09-27"}, {"id": "majors-007", "title": "TotalEnergies to Collaborate with Mistral AI to Increase the Application of Artificial Intelligence", "url": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "publisher": "TotalEnergies", "date": "2025-06-12", "type": "company announcement", "locator": "A shared AI innovation lab", "review_status": "full text reviewed", "supports": ["Joint lab use cases include an assistant for 1,000 researchers, decision-support for industrial performance and CO2 emissions, and customer energy-saving support.", "TotalEnergies reported Digital Factory had developed over 100 solutions, 60 using ML through generative AI; these are aggregate portfolio counts, not outcome metrics."], "accessed": "2026-09-27"}, {"id": "majors-020", "title": "TotalEnergies Announces a Partnership with Mistral to Develop Frontier AI Models for Reservoir Exploration and Engineering", "url": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "publisher": "TotalEnergies", "date": "2026-09-15", "accessed": "2026-09-27", "type": "company announcement", "locator": "Driving Innovation by Combining AI, Geoscience Expertise and Subsurface Data", "review_status": "full text reviewed", "supports": ["Three-year joint program represents over €100 million investment to develop frontier AI models supporting exploration, characterization and development of oil and gas reservoirs.", "Program will use agentic AI and combine almost 10 petaflops of data with geoscience/reservoir expertise; develop multiple scenarios for exploration opportunities and optimizing existing projects.", "Scientific lab will combine TotalEnergies subsurface experts with Mistral AI science and technology."]}], "status": "reviewed", "gaps": "Reviewed 2025 Mistral and Cognite announcements, Feb 2026 results deck, and Sep 2026 Mistral frontier-model announcement.", "cohort": "majors"}, {"company": "Eni", "existing_ids": ["AI-027"], "searched": true, "queries": ["Eni AI 2024", "Eni AI 2025", "Eni AI 2026 operations"], "documents_reviewed": [{"id": "majors-013", "title": "Eni opens its supercomputing capabilities to the innovation ecosystem", "url": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "publisher": "Eni", "date": "2026-07-28", "type": "company announcement", "locator": "Research and Development", "review_status": "full text reviewed", "supports": ["Eni opened HPC infrastructure, expertise and tools to companies, startups and research centres to develop AI and digital innovation; HPC6 and HPC7 together exceed exascale-class performance."], "accessed": "2026-09-27"}, {"id": "majors-014", "title": "Eni for 2025 — Just Transition", "url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "publisher": "Eni", "date": "2026-04-22", "type": "annual report", "locator": "Data and artificial intelligence", "review_status": "excerpt reviewed", "supports": ["Eni reports around 300 AI use cases (+40% vs. 2024), around 20 AI agents; digital plant target ranges of 2–3% production enhancement and emissions reduction and drilling automation target of 35% less drilling time."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Annual sustainability excerpt and July 2026 HPC access source reviewed; existing AI-027 capital-markets portfolio disclosure not reopened; operational use-case specifics/metric attribution need further review.", "cohort": "majors"}, {"company": "Repsol", "existing_ids": ["AI-032"], "searched": true, "queries": ["Repsol AI 2024", "Repsol AI 2025", "Repsol AI 2026 operations"], "documents_reviewed": [{"id": "majors-010", "title": "Repsol launches intelligent multi-agent platform", "url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "publisher": "Repsol", "date": "2025-10-15", "type": "company announcement", "locator": "Pilot project / GenAI competence center", "review_status": "full text reviewed", "supports": ["Accenture collaboration produced an agent-construction platform and pilot in Digitalization and Services department.", "Three teams with 22 agents supported processes for over four months; over 50 people supervised the agent teams.", "Repsol reports 60 production GenAI use cases and 22 scaled as of Oct 2025; these are company-reported portfolio measures."], "accessed": "2026-09-27"}, {"id": "majors-011", "title": "Consolidated Management Report 2025 — Repsol", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "publisher": "Repsol", "date": "2026-02-19", "type": "annual report", "locator": "5.5 Technology and digitalization", "review_status": "excerpt reviewed", "supports": ["Repsol reports Automated Well Data Ingestion automates daily drilling and geology reports and integrates data into corporate systems to speed exploration decisions.", "Annual report later lists 34 agents in three workflows, distinct from the Oct 2025 pilot snapshot of 22 agents."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "2025 agent-platform announcement and 2025 annual report excerpt reviewed; workflow-specific outcome metrics are absent.", "cohort": "majors"}, {"company": "OMV", "existing_ids": ["AI-084", "AI-132", "AI-133", "AI-134", "AI-135"], "searched": true, "queries": ["OMV AI 2024", "OMV AI 2025", "OMV AI 2026 operations"], "documents_reviewed": [{"id": "majors-012", "title": "Digitalization — OMV Combined Annual Report 2025", "url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "publisher": "OMV", "date": "2026-04-02", "type": "annual report", "locator": "Digitalization; AI Across the Value Chain", "review_status": "full text reviewed", "supports": ["OMV AI Hub launched 2025 as secure entry point built on Azure; in-house GPT used monthly by 2,500+ employees; Regulations Assistant searches 1,500+ internal rules.", "OMV reports 50+ AI projects in use, 40+ in development, >210 ideas, and 10,000 employees trained.", "Norwegian AI Companion searches 2.5m pages of internal/public geological and engineering information.", "iDEA surfaces global analogues and lessons for drilling engineers; Romanian pump failure ML pilot predicts failures up to 90 days ahead.", "Agent-based field development in low-carbon projects is being tested with Stanford University and TerraAI; early value claims remain preliminary."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "2025 annual-report digitalization section reviewed in full; AI-132-135 are supported, while report does not quantify outcome measures for these workflows.", "cohort": "majors"}, {"company": "Equinor", "existing_ids": ["AI-033", "AI-034", "AI-035", "AI-036"], "searched": true, "queries": ["Equinor AI 2024", "Equinor AI 2025", "Equinor AI 2026 operations"], "documents_reviewed": [{"id": "majors-008", "title": "Use of artificial intelligence saved Equinor USD 130 million in 2025", "url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "publisher": "Equinor", "date": "2026-01-07", "type": "company announcement", "locator": "Three valuable contributors implemented", "review_status": "full text reviewed", "supports": ["Equinor attributes $130 million value creation/savings in 2025 to AI across company and partners; >$330 million since 2020.", "Predictive maintenance monitors 700+ rotating machines using 24,000 sensors and is attributed $120 million value since 2020.", "AI well/field planning generated thousands of alternatives; Equinor says a Johan Sverdrup phase 3 solution saved the partnership $12 million.", "AI seismic tool increases interpretation capacity 10x; 2 million km² interpreted in 2025."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Jan 2026 official release re-verifies the three practical categories in AI-034/035/036 and reports value; older source locations for legacy IDs were not re-opened. No independent corroboration of reported value or valuation method.", "cohort": "majors"}, {"company": "Woodside", "existing_ids": [], "searched": true, "queries": ["Woodside AI 2024", "Woodside AI 2025", "Woodside AI 2026 operations"], "documents_reviewed": [{"id": "majors-009", "title": "Right fix at right time – AI delivers for Woodside", "url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "publisher": "Woodside", "date": "2026-01-05", "type": "company announcement", "locator": "Maint Intel", "review_status": "full text reviewed", "supports": ["Maint Intel analyses maintenance records and failure modes using statistical models to recommend equipment maintenance intervals.", "Built by Woodside Digital with AWS India; tested on Angel platform, launched at North West Shelf and scheduled for global rollout during 2026.", "Company reports model processing shortened from five days to under two hours and identifies potential 15% yearly maintenance-hours reduction at Angel; latter is potential, not achieved."], "accessed": "2026-09-27"}, {"id": "majors-021", "title": "Woodside Artificial Intelligence — Startup Advisor", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "publisher": "Woodside", "date": null, "accessed": "2026-09-27", "type": "company announcement", "locator": "Startup Advisor", "review_status": "full text reviewed", "supports": ["Startup Advisor uses AI, automation and digital procedures to guide LNG plant operators through startup steps and compare real-time performance to a standard.", "Woodside says it has been proven in live operations and used during critical situations to support decisions; control-room leads and panel operators endorsed the tool."]}, {"id": "majors-023", "title": "Artificial Intelligence — Woodside Energy", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "publisher": "Woodside", "date": null, "accessed": "2026-09-27", "type": "company announcement", "locator": "AI value; Startup Advisor", "review_status": "full text reviewed", "supports": ["Woodside reports >50 AI agents in Lumina, and Startup Advisor guides LNG plant operators with AI, automation and digital procedures.", "Page does not state publication or first deployment date; Startup Advisor is not counted as time-confirmed 24-month deployment."]}], "status": "reviewed", "gaps": "Maint Intel and general AI pages reviewed. Maint Intel is dated Jan 2026. Startup Advisor is concrete, but the source and first-deployment dates are unstated; it is not time-confirmed.", "cohort": "majors"}, {"company": "Santos", "existing_ids": ["AI-083"], "searched": true, "queries": ["Santos AI 2024", "Santos AI 2025", "Santos AI 2026 operations"], "documents_reviewed": [{"id": "majors-017", "title": "Santos 2026 Investor Briefing Day", "url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "publisher": "Santos", "date": "2026-05-20", "type": "investor presentation", "locator": "Technology program", "review_status": "excerpt reviewed", "supports": ["Santos says technology program applies digital and AI capability across integrated value chain, including cost reduction and production uplift across drilling, completions, projects and facilities; specific quantified AI-only impacts not isolated."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "2026 investor briefing excerpt reviewed. Need primary source review of Xecta named digital-twin initiative (AI-083 is secondary) and detailed AI tech program slides.", "cohort": "majors"}, {"company": "Aker BP", "existing_ids": ["AI-131"], "searched": true, "queries": ["Aker BP AI 2024", "Aker BP AI 2025", "Aker BP AI 2026 operations"], "documents_reviewed": [{"id": "majors-015", "title": "Aker BP Q4 2025 transcript", "url": "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf", "publisher": "Aker BP", "date": "2026-02-11", "type": "earnings call", "locator": "PDF pp. 5–6", "review_status": "full text reviewed", "supports": ["CEO describes AI-enabled earlier exploration decisions and AI agents reducing operational troubleshooting time and improving uptime.", "The CEO describes high-quality structured real-time data connection across the company; 10 years of digital investment is contextual foundation.", "Microsoft speaker (partner attribution) says Copilot adoption nearly 100%; not independently audited."], "accessed": "2026-09-27"}, {"id": "majors-016", "title": "Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation", "url": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "publisher": "Aker BP", "date": "2026-04-29", "type": "company announcement", "locator": "Yggdrasil remote operations", "review_status": "full text reviewed", "supports": ["Yggdrasil operations from onshore centre planned from 2027; AI agents planned from day one for condition-based maintenance and planning, with scaling to other workflows as mature.", "Aker BP says industrial data platform Cognite Data Fusion structures thousands of sensor streams and data is to be shared with people and AI agents."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "2026 Q4 transcript and Yggdrasil official page reviewed. AI-131 Copilot profile remains partner case study and workforce productivity evidence is partial.", "cohort": "majors"}, {"company": "Aramco", "existing_ids": ["AI-037", "AI-038", "AI-039", "AI-070"], "searched": true, "queries": ["Aramco AI gas plant RL acid gas removal Fadhili 2025", "site:aramco.com AI artificial intelligence 2025 operational deployment"], "documents_reviewed": [{"title": "Aramco AI and Big Data", "url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "date": "2026-09-27", "type": "company announcement"}], "status": "reviewed", "gaps": "Existing AI-037/038/070 profile URLs not all reviewed individually; this pass verifies a specific Fadhili control deployment and cites official AI portfolio source. AI-039 investment plan not independently checked.", "cohort": "nocs"}, {"company": "ADNOC", "existing_ids": ["AI-011", "AI-040", "AI-071", "AI-108"], "searched": true, "queries": ["site:adnoc.ae AI artificial intelligence 2025 AIQ agent operations", "ADNOC AiPSO SLB launch 2025"], "documents_reviewed": [{"title": "ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution", "url": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "date": "2025-01-16", "type": "company announcement"}, {"title": "ADNOC and SLB Launch Transformative AI-Powered Solution to Boost Upstream Productivity", "url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "date": "2025-11-03", "type": "company announcement"}], "status": "reviewed", "gaps": "ENERGYai scope and AiPSO were verified from ADNOC sources. Existing training-hours metric remains secondary-source only; company-wide 200+ use cases and forecast-accuracy claims were not separately validated.", "cohort": "nocs"}, {"company": "PETRONAS", "existing_ids": ["AI-050", "AI-079", "AI-080"], "searched": true, "queries": ["site:petronas.com AI 2025 subsurface PETRONAS", "PETRONAS ERMAI AI.SEEK Global Exploration Basin 2025"], "documents_reviewed": [{"title": "PETRONAS Integrated Report 2025", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "date": "2026-02-27", "type": "annual report"}, {"title": "PETRONAS Carigali Unveils Pioneering AI Venture - TriCipta AI", "url": "https://www.petronas.com/bm/node/1688", "date": "2025-06-18", "type": "company announcement"}, {"title": "TIRAI - Your New PETRONAS myPROdata Intelligence Assistance is Live!", "url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "date": "2025-11-03", "type": "company announcement"}], "status": "reviewed", "gaps": "The report is excerpt-reviewed for ERMAI; no full engineering documentation or independent validation. Existing AI-079 Microsoft collaboration is ecosystem strategy, not counted as an operating workflow.", "cohort": "nocs"}, {"company": "PetroChina / CNPC", "existing_ids": ["AI-081"], "searched": true, "queries": ["site:cnpc.com.cn AI artificial intelligence 2025 Kunlun oil gas", "CNPC Kunlun 2026 applications official Chinese"], "documents_reviewed": [{"title": "2024 Annual Report", "url": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "date": "2025-08-01", "type": "annual report"}, {"title": "中国石油昆仑大模型迈入主动智能新阶段", "url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml", "date": "2026-05-29", "type": "company announcement"}], "status": "reviewed", "gaps": "CNPC official annual report and Chinese company news reviewed. Kunlun 2024 launch is outside window; only its 2025-26 scaling and deployments are current-window findings.", "cohort": "nocs"}, {"company": "CNOOC", "existing_ids": ["AI-068", "AI-082"], "searched": true, "queries": ["site:cnooc.com.cn AI 2025 artificial intelligence intelligent oilfield", "CNOOC interim results AI+ smart drilling 2025"], "documents_reviewed": [{"title": "2025 Interim Results", "url": "https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html", "date": "2025-08-27", "type": "company announcement"}], "status": "reviewed", "gaps": "AI+ and smart-field disclosures verified from official interim results. Existing predictive-maintenance ~20% claim is secondary and was not confirmed by that source; vendor statements in AI-068 were not reviewed directly.", "cohort": "nocs"}, {"company": "Sinopec", "existing_ids": [], "searched": true, "queries": ["site:sinopec.com AI 2025 artificial intelligence oil gas", "Sinopec Jianghan oilfield AI 2025 Chinese"], "documents_reviewed": [{"title": "〖中国石化报〗多场景推进人工智能技术应用", "url": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "date": "2025-05-28", "type": "company announcement"}], "status": "reviewed", "gaps": "Specific subsidiary/field case reviewed in Chinese. Does not establish all-China group deployment or validate vendor/engineering metrics independently.", "cohort": "nocs"}, {"company": "Kuwait Oil Company", "existing_ids": ["AI-014", "AI-066", "AI-085"], "searched": true, "queries": ["Kuwait Oil Company AI seismic G Agent 2025 official", "site:kpc.com.kw AI models seismic drilling"], "documents_reviewed": [], "status": "partial", "gaps": "Existing AI-014/066 rely on SLB vendor disclosures, and AI-085 on secondary press. KPC sustainability report surfaced, but KOC-specific operating details still require company-primary technical or procurement source review.", "cohort": "nocs"}, {"company": "OQ Exploration & Production", "existing_ids": ["AI-013"], "searched": true, "queries": ["OQ Exploration Production SLB autonomous drilling Oman AI 2025", "site:oq.com AI artificial intelligence upstream"], "documents_reviewed": [], "status": "partial", "gaps": "Only vendor earnings disclosure in existing AI-013 was available in this pass; no OQ corporate primary source was located.", "cohort": "nocs"}, {"company": "PTTEP", "existing_ids": ["AI-067"], "searched": true, "queries": ["PTTEP offshore production AI pilot SLB 2026", "site:pttep.com artificial intelligence AI 2025"], "documents_reviewed": [], "status": "partial", "gaps": "Only vendor earnings disclosure in existing AI-067 was available in this pass; direct PTTEP primary-source confirmation remains needed.", "cohort": "nocs"}, {"company": "Petrobras", "existing_ids": [], "searched": true, "queries": ["site:petrobras.com.br inteligência artificial 2025 IA Petrobras", "Petrobras Smart Tocha AI flare official"], "documents_reviewed": [{"title": "Inovação e tecnologia: nossa fonte de energia", "url": "https://petrobras.com.br/inovacao-e-tecnologia", "date": "2026-09-27", "type": "company announcement"}], "status": "reviewed", "gaps": "Smart Tocha official company technology page reviewed. Other digital/GenAI product pages surfaced but many were launched before the window; portfolio-wide search is incomplete.", "cohort": "nocs"}, {"company": "Ecopetrol", "existing_ids": [], "searched": true, "queries": ["site:ecopetrol.com.co inteligencia artificial 2025 IA Ecopetrol", "Ecopetrol annual report 2025 AI planner AIOps"], "documents_reviewed": [{"title": "Informe de gestión y resultados 2025", "url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "date": "2026-03-31", "type": "annual report"}, {"title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/", "date": "2026-05-04", "type": "company announcement"}], "status": "reviewed", "gaps": "Official 2025 report excerpt and 2026 CADE release reviewed. Benefits from group science/technology portfolio are not attributed to AI. CADE’s AI component is prospective.", "cohort": "nocs"}, {"company": "YPF", "existing_ids": [], "searched": true, "queries": ["site:ypf.com inteligencia artificial IA 2025 YPF", "YPF sustainability report 2024 AI chatbot refinery inspection"], "documents_reviewed": [{"title": "Reporte de Sustentabilidad 2024", "url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "date": "2025-01-01", "type": "annual report"}], "status": "reviewed", "gaps": "YPF sustainability report documents 2024 projects and governance; exact operating scale and impact remain undisclosed.", "cohort": "nocs"}, {"company": "ONGC", "existing_ids": [], "searched": true, "queries": ["ONGC AI artificial intelligence oilfield 2025 official ONGC AI gas lift tender", "site:ongcindia.com AI Artificial Intelligence ONGC"], "documents_reviewed": [{"title": "ONGC tender for automated gas-lift system", "url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "date": "2025-11-03", "type": "company announcement"}], "status": "partial", "gaps": "Only procurement specification verified, not an award or operating deployment. Older ONGC AI seismic software references surfaced but were outside the requested period.", "cohort": "nocs"}, {"company": "Pertamina", "existing_ids": [], "searched": true, "queries": ["site:pertamina.com artificial intelligence AI 2025 Pertamina", "Pertamina Hulu Rokan PINTAR ChatGPT HSE 2024", "Pertamina Geothermal machine learning digital twin 2025"], "documents_reviewed": [{"title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "date": "2025-09-01", "type": "company announcement"}, {"title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "date": "2025-10-27", "type": "company announcement"}], "status": "reviewed", "gaps": "Specific upstream HSE GenAI and geothermal digital program reviewed. Geothermal results cover a bundled program and are not AI-only; parent-group other assets not comprehensively searched.", "cohort": "nocs"}, {"company": "QatarEnergy", "existing_ids": [], "searched": true, "queries": ["site:qatarenergy.qa artificial intelligence AI 2025 QatarEnergy", "QatarEnergy AI 2024 sustainability report"], "documents_reviewed": [], "status": "no qualifying initiative found", "gaps": "Search found broad digitalization/AI analytics language, hiring for AI/BI roles and AI infrastructure exposure, but no named practical operating workflow with adequate primary evidence in this pass. Absence is not established.", "cohort": "nocs"}, {"company": "Devon Energy", "existing_ids": ["AI-006", "AI-007", "AI-086"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Devon Energy Stays Ahead Of Technology Curve", "url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "date": "2025-07", "type": "partner case study", "accessed": "2026-09-27"}, {"title": "Devon and Coterra to Combine, Creating a Premier Shale Operator", "url": "https://investors.devonenergy.com/investors/press-releases/press-release-details/2026/Devon-Energy-and-Coterra-Energy-to-Combine-Creating-a-Premier-Shale-Operator/", "date": "2026-02-01", "type": "company announcement", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; AI workflows found. Q2 transcript is third-party-hosted; issuer-hosted slide 9 exists according to transcript but not separately reopened. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "EOG Resources", "existing_ids": ["AI-088", "AI-089"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "EOG Q3 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "date": "2025-11-27", "type": "secondary", "accessed": "2026-09-27"}, {"title": "EOG Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "date": "2026-02-25", "type": "secondary", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; detailed claims remain secondary-only. EOG investor materials for 2025-26 were searched; exact claims need issuer audio/transcript verification. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "ConocoPhillips", "existing_ids": ["AI-090"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "ConocoPhillips Annual Meeting 2026 Transcript", "url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "date": "2026", "type": "company announcement", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; enterprise AI strategy disclosure, no named workflow. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Occidental Petroleum", "existing_ids": ["AI-044", "AI-091"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Q4 2024 Occidental Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf", "date": "2025-02-19", "type": "earnings call", "accessed": "2026-09-27"}, {"title": "Q1 2026 Occidental Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "date": "2026-05-06", "type": "earnings call", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; named functions and AI Center of Excellence found; detailed workflow metrics absent. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Diamondback Energy", "existing_ids": ["AI-045"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Diamondback Q2 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "date": "2026-08-10", "type": "secondary", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; AI/lift claim not issuer-source verified. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Coterra Energy", "existing_ids": ["AI-087"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Devon and Coterra to Combine, Creating a Premier Shale Operator", "url": "https://investors.devonenergy.com/investors/press-releases/press-release-details/2026/Devon-Energy-and-Coterra-Energy-to-Combine-Creating-a-Premier-Shale-Operator/", "date": "2026-02-01", "type": "company announcement", "accessed": "2026-09-27"}, {"title": "Devon Energy Q2 2026 Earnings Call Transcript", "url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "date": "2026-08", "type": "secondary", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; Coterra models are described in merger integration disclosure, but only third-party transcript details them; no pre-merger primary AI profile found. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Ovintiv", "existing_ids": ["AI-092", "AI-093", "AI-094"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Ovintiv Q3 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "date": "2025-11-06", "type": "secondary", "accessed": "2026-09-27"}, {"title": "Ovintiv Q4 2025 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "date": "2026-02-24", "type": "secondary", "accessed": "2026-09-27"}, {"title": "Ovintiv Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "date": "2026-05-12", "type": "secondary", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; practical workflows described in secondary transcripts; primary confirmation outstanding. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Chord Energy", "existing_ids": ["AI-097"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Chord Energy Q1 2026 Earnings Call Transcript", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "date": "2026-05-06", "type": "secondary", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; combined intervention and AI effect not separated; third-party only. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Matador Resources", "existing_ids": ["AI-096"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Matador Q4 and FY2025 Results and 2026 Operating Plan", "url": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "date": "2026-02-24", "type": "company announcement", "accessed": "2026-09-27"}, {"title": "Matador Q4 2025 Earnings Call", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "date": "2026-02-25", "type": "secondary", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; AI integration is strategy-level; vendor detail secondary-only. AI gas demand is infrastructure. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "EQT", "existing_ids": ["AI-063", "AI-095"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "EQT Q4 2025 Earnings Call", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "date": "2026-02-18", "type": "secondary", "accessed": "2026-09-27"}, {"title": "Homer City Redevelopment announces agreement in principle for EQT Corporation to supply natural gas", "url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "date": "2025-07", "type": "company announcement", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; procurement claim secondary-only; Homer City claim categorized as AI-demand infrastructure. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Canadian Natural Resources", "existing_ids": ["AI-098"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Canadian Natural 2025 Annual Information Form", "url": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "date": "2026-03-25", "type": "annual report", "accessed": "2026-09-27"}, {"title": "Governance and Risk Management", "url": "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/", "date": "2026", "type": "company announcement", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; primary policy detail confirms cost-justified use and production-critical gate. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "Suncor Energy", "existing_ids": ["AI-099"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "The Future of Canadian Mining is Autonomous — and Suncor is Leading the Way", "url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "date": "2026", "type": "company announcement", "accessed": "2026-09-27"}, {"title": "Dispatching innovation: Suncor's new dispatch system improves mine operations", "url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "date": "2025", "type": "company announcement", "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed; concrete dispatch and autonomous fleet workflows. Company source updates deployment count. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use. Suncor haulage fleet and dynamic dispatch are separate initiatives; both cited company pages are undated, so the deployment snapshot cannot be pinned to the 24-month window.", "cohort": "north_america"}, {"company": "Cenovus Energy", "existing_ids": ["AI-061"], "searched": true, "queries": ["company + AI/artificial intelligence + 2024/2025/2026 earnings, investor presentation and annual filing", "issuer-hosted investor relations/annual report/quarterly earnings review", "existing profile URL review where available"], "documents_reviewed": [{"title": "Cenovus 2025 Management’s Discussion and Analysis", "url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "date": "2026", "type": "annual report", "accessed": "2026-09-27"}], "status": "partial", "gaps": "Partial; existing pilot claim needs cited MD&A passage reopened; no primary evidence of named pilots found in current snippets. Not an exhaustive review of every quarterly filing; absent detail is not evidence of no use.", "cohort": "north_america"}, {"company": "SLB", "existing_ids": ["AI-008", "AI-009", "AI-010", "AI-011", "AI-012", "AI-013", "AI-014", "AI-015", "AI-046", "AI-047", "AI-048", "AI-066", "AI-067", "AI-068", "AI-069", "AI-105", "AI-106", "AI-107"], "searched": true, "queries": ["SLB Tela catalog 33 workflows official", "SLB 2025 Q4 results Neuro OQ KOC seismic bp remote drilling", "SLB Q2 2026 earnings KOC ONGC PTTEP CNOOC AI", "SLB Petoro AI well planning case study", "SLB Digital Marketplace June 2026 official"], "documents_reviewed": [{"title": "Tela agentic-AI assistant catalog", "url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "date": "Undated; accessed 2026-09-27", "type": "company product documentation"}, {"title": "SLB Unveils Groundbreaking New Agentic AI Technology for the Energy Industry", "url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "date": "2025-11-03", "type": "company announcement"}, {"title": "SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy", "url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "date": "2026-06-15", "type": "company announcement"}, {"title": "Petoro and SLB: Pioneering AI-driven well planning on the Norwegian continental shelf", "url": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "date": "2026-06-05", "type": "partner case study"}, {"title": "SLB Announces Third-Quarter 2025 Results", "url": "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609", "date": "2025-10-17", "type": "earnings release"}, {"title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "date": "2026-01-23", "type": "earnings release"}, {"title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "date": "2026-07-24", "type": "earnings release"}], "status": "reviewed", "gaps": "Portfolio and partner contracts are documented from primary SLB sources. Partner-side verification was not located for individual customer deployments. Product catalog is capability, not proof of deployment.", "cohort": "services"}, {"company": "Halliburton", "existing_ids": ["AI-016", "AI-017", "AI-018", "AI-109", "AI-110", "AI-111"], "searched": true, "queries": ["Halliburton Coterra Octiv Auto Frac first fully automated hydraulic fracturing 2025", "Halliburton Chevron closed-loop intelligent fracturing 2025", "Halliburton Sekal automated on-bottom Equinor drilling 2025"], "documents_reviewed": [{"title": "Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program", "url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "date": "2025-01-06", "type": "company announcement"}, {"title": "Chevron and Halliburton Enable Intelligent Hydraulic Fracturing", "url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "date": "2025-04-12", "type": "company announcement"}, {"title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "date": "2025-02-26", "type": "company announcement"}], "status": "reviewed", "gaps": "Customer metrics are supplier-reported. Chevron workflow evidence establishes automation, not AI.", "cohort": "services"}, {"company": "Baker Hughes", "existing_ids": ["AI-019", "AI-020", "AI-021", "AI-112"], "searched": true, "queries": ["Baker Hughes Repsol Leucipa GenAI and Lift Optimizer 2025"], "documents_reviewed": [{"title": "Baker Hughes, Repsol Launching New AI-Powered Functionality in Leucipa", "url": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "date": "2025-06-10", "type": "company announcement"}], "status": "reviewed", "gaps": "Announcement says Repsol was in early-stage pilot; measured outcomes and scale remain unknown.", "cohort": "services"}, {"company": "Weatherford", "existing_ids": ["AI-113"], "searched": true, "queries": ["Weatherford Industrial Intelligence AI 2025 launch", "Weatherford 2025 sustainability AI agents internal workflows"], "documents_reviewed": [{"title": "Weatherford Launches Industrial Intelligence at FWRD 2025", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "date": "2025-10-01", "type": "company announcement"}, {"title": "Weatherford 2025 Sustainability Report: Artificial Intelligence", "url": "https://www.weatherford.com/documents/sustainability/sustainability-report/", "date": "2026-04-01", "type": "sustainability report"}], "status": "reviewed", "gaps": "Launch release documents product capabilities; adoption at customers not established. Internal employee agents are a separate workflow.", "cohort": "services"}, {"company": "NOV", "existing_ids": ["AI-061", "AI-114", "AI-122", "AI-123"], "searched": true, "queries": ["NOV Egypt Drilling Beliefs Analytics AI record bit run 2025", "NOV AIQ edge drilling pilot 2026"], "documents_reviewed": [{"title": "NOV supports record bit run and Egypt’s first remote operations center", "url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center", "date": "2025-06-27", "type": "company announcement"}, {"title": "NOV and AIQ to Advance AI-Enabled Drilling Optimization", "url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "date": "2026-01-12", "type": "company announcement"}], "status": "reviewed", "gaps": "Egypt case is vendor-authored; operator identity and independent performance validation unavailable. AIQ work is at planned pilot stage.", "cohort": "services"}, {"company": "Nabors", "existing_ids": ["AI-115"], "searched": true, "queries": ["Nabors Corva alliance expansion RigCLOUD Powered by Corva 2025", "Nabors Predictive Drilling field result 2023"], "documents_reviewed": [{"title": "Nabors and Corva Expand Strategic Alliance to Accelerate Digital Innovation", "url": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "date": "2025-04-10", "type": "company announcement"}, {"title": "Predictive Drilling - Nabors", "url": "https://www.nabors.com/drilling-automation/predictive-drilling/", "date": "Undated; accessed 2026-09-27", "type": "company product documentation"}], "status": "reviewed", "gaps": "Within-window alliance expansion is verified. The cited 36% ROP/9.7% vibration metrics come from an older 2023 case and are excluded from current-window results.", "cohort": "services"}, {"company": "Helmerich & Payne", "existing_ids": ["AI-116"], "searched": true, "queries": ["H&P FlexRobotics supermajor Permian deployment 2026", "H&P FlexRobotics rig-floor system AI zone monitoring"], "documents_reviewed": [{"title": "Helmerich & Payne, Inc. Announces Fiscal First Quarter Results", "url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "date": "2026-02-04", "type": "earnings release"}, {"title": "FlexRobotics Hands-Free Rig Floor", "url": "https://www.hpinc.com/technologies/flexrobotics", "date": "Undated; accessed 2026-09-27", "type": "company product documentation"}], "status": "reviewed", "gaps": "Rig-floor handling is robotics/automation; only ancillary zone monitoring is described as AI-enabled.", "cohort": "services"}, {"company": "Patterson-UTI", "existing_ids": ["AI-117"], "searched": true, "queries": ["Patterson-UTI Q2 2025 Cortex REX AI machine learning", "Patterson-UTI Q4 2025 nearly all rigs Cortex automation"], "documents_reviewed": [{"title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended June 30, 2025", "url": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "date": "2025-07-30", "type": "earnings release"}, {"title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025", "url": "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm", "date": "2026-02-04", "type": "earnings release"}], "status": "reviewed", "gaps": "Cortex autodriller is explicitly ML-enabled. Near-all-rig penetration applies to the broader automation suite, not AI model operation on every rig.", "cohort": "services"}, {"company": "Liberty Energy", "existing_ids": ["AI-118"], "searched": true, "queries": ["Liberty Forge LLM StimCommander rate pressure control Q3 2025"], "documents_reviewed": [{"title": "Liberty Energy Inc. Announces Third Quarter 2025 Financial and Operational Results", "url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "date": "2025-10-16", "type": "earnings release"}], "status": "reviewed", "gaps": "Performance metrics are company-reported for StimCommander automation; source does not isolate Forge contribution.", "cohort": "services"}, {"company": "Precision Drilling", "existing_ids": [], "searched": true, "queries": ["Precision Drilling 2024 ESG Alpha AI predictive maintenance", "Precision Drilling Alpha technology 2025"], "documents_reviewed": [{"title": "Precision Drilling 2024 ESG Report", "url": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "date": "2025-04-01", "type": "sustainability report"}], "status": "partial", "gaps": "Company report confirms 78 Alpha digital rigs and AI use, but no AI-specific outcome measure was found; rig-level deployment details are thin.", "cohort": "services"}, {"company": "Transocean", "existing_ids": [], "searched": true, "queries": ["Transocean 2025 Form 10-K AI use office administration offshore drilling"], "documents_reviewed": [{"title": "Transocean Ltd. 2025 Form 10-K", "url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "date": "2026-02-19", "type": "annual report"}], "status": "reviewed", "gaps": "Form 10-K explicitly says no AI is currently used in offshore drilling; office/admin use only, as described.", "cohort": "services"}, {"company": "Saipem", "existing_ids": [], "searched": true, "queries": ["Saipem AI for Engineering 2025 AI Project Schedule Onshore"], "documents_reviewed": [{"title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "date": "2026-04-01", "type": "sustainability report"}], "status": "partial", "gaps": "Sustainability-report excerpt identifies use cases and one named scheduling tool; implementation scale and results were not quantified.", "cohort": "services"}, {"company": "TechnipFMC", "existing_ids": [], "searched": true, "queries": ["TechnipFMC Subsea Studio AI deployment operations digital twin 2025", "TechnipFMC 2025 annual report AI"], "documents_reviewed": [], "status": "no qualifying initiative found", "gaps": "Generic AI and remote-operations material surfaced, but no named dated customer workflow with adequate primary detail was located in this pass. Search absence is not proof of absence.", "cohort": "services"}, {"company": "Blackstone / Enverus", "existing_ids": ["AI-051", "AI-065"], "searched": true, "queries": ["Blackstone / Enverus AI energy 2024", "Blackstone / Enverus AI oil and gas 2025", "Blackstone / Enverus AI energy 2026"], "documents_reviewed": [{"id": "sc-001", "title": "Blackstone announces agreement to acquire Enverus", "url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "publisher": "Blackstone", "date": "2025-08-06", "type": "company announcement", "locator": "Transaction announcement", "review_status": "full text reviewed", "supports": ["Blackstone funds agreed to acquire Enverus from Hellman & Friedman and Genstar; this is a financial transaction, not an AI deployment.", "Enverus provides energy data analytics and insights used by energy companies and investors."], "accessed": "2026-09-27"}, {"id": "sc-002", "title": "Enverus ONE is live: governed AI platform built for energy industry", "url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "publisher": "Enverus", "date": "2026-04-07", "type": "company announcement", "locator": "Launch, platform design and four Flows", "review_status": "full text reviewed", "supports": ["Enverus ONE launched with Astra energy model paired with general frontier AI, proprietary energy datasets and governed access controls.", "AFE Evaluation, Current Production Valuation, Project Siting and QuickStart workflows are available.", "AFE evaluation automates ingestion, ownership validation, offset economics and comparisons; vendor says weeks can become hours."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Enverus ONE is product launch with no named customer outcome. Blackstone deal is a separate ownership record; agreement source does not establish closing.", "cohort": "software_capital"}, {"company": "Quantum Capital Group", "existing_ids": ["AI-052", "AI-053", "AI-054", "AI-055"], "searched": true, "queries": ["Quantum Capital Group AI energy 2024", "Quantum Capital Group AI oil and gas 2025", "Quantum Capital Group AI energy 2026"], "documents_reviewed": [{"id": "sc-003", "title": "Quantum Capital Group Digital Mission", "url": "https://www.quantumcap.com/digital-mission/", "publisher": "Quantum Capital Group", "date": "undated", "type": "company webpage", "locator": "QEC, QIQ and AI at Work", "review_status": "full text reviewed", "supports": ["Quantum Energy Cloud unifies portfolio financial, operational and market data using Microsoft and Databricks.", "QIQ is a proprietary generative AI interface that answers questions, automates analyses and initiates workflows over portfolio data.", "Named workflows include single-well economics, geospatial capital allocation, completion design, artificial lift, anomaly detection and downtime prediction."], "accessed": "2026-09-27"}, {"id": "sc-004", "title": "2024 ESG Report", "url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "publisher": "Quantum Capital Group", "date": "2025-05-02", "type": "annual report", "locator": "AI/data infrastructure section (search excerpt; PDF open failed)", "review_status": "excerpt reviewed", "supports": ["QEC and private generative AI QIQ are described; report claims over 100 AI accelerators used by portfolio companies.", "Operational as-of date in excerpt is May 2024, prior to research window; report publication itself is in-window."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Digital Mission page is undated, and 2024 ESG PDF could not be opened; accelerator operations described as of May 2024 are outside window. No workflow-level ROI or independent validation.", "cohort": "software_capital"}, {"company": "Novi Labs / Invictus", "existing_ids": ["AI-122", "AI-123"], "searched": true, "queries": ["Novi Labs / Invictus AI energy 2024", "Novi Labs / Invictus AI oil and gas 2025", "Novi Labs / Invictus AI energy 2026"], "documents_reviewed": [{"id": "sc-005", "title": "Novi Labs raises $35 million from Invictus Growth Partners", "url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "publisher": "Novi Labs", "date": "2025-06-25", "type": "company announcement", "locator": "Funding and platform overview", "review_status": "full text reviewed", "supports": ["Invictus Growth Partners invested $35 million to support Novi product innovation, expansion and acquisitions.", "Novi describes platform combining machine learning, proprietary datasets and reservoir science for upstream capital decisions."], "accessed": "2026-09-27"}, {"id": "sc-006", "title": "Novi Labs unveils Novi Intelligence", "url": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "publisher": "Novi Labs", "date": "2025-09-10", "type": "company announcement", "locator": "Novi Intelligence product", "review_status": "full text reviewed", "supports": ["Novi Intelligence joins operator-sourced data, machine learning and forecasts with proprietary metrics such as rock quality, completion intensity, depletion and well density.", "Product exposes underlying data for client validation; accuracy claims are vendor claims."], "accessed": "2026-09-27"}, {"id": "sc-007", "title": "Novi Labs acquires Turing Analytics", "url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "publisher": "Novi Labs", "date": "2026-04-29", "type": "company announcement", "locator": "Acquisition rationale and geographic coverage", "review_status": "excerpt reviewed", "supports": ["Novi acquired Turing Analytics/EVA to combine US and Canadian upstream datasets and analytics for investment evaluation.", "Novi says Turing clients represent over 50% of Canadian oil and gas production; company-reported."], "accessed": "2026-09-27"}, {"id": "sc-008", "title": "Novi Labs acquires RBN Energy", "url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "publisher": "Novi Labs", "date": "2026-05-28", "type": "company announcement", "locator": "Combination and market workflows", "review_status": "full text reviewed", "supports": ["RBN acquisition combines Novi upstream AI forecasts with commodity, midstream and demand intelligence.", "Decision context includes basis, transport capacity, commodity flows and end-market pricing."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Company releases support funding, products and acquisitions. No independent forecast accuracy test or proof acquisitions are technically integrated.", "cohort": "software_capital"}, {"company": "Quorum / Francisco Partners", "existing_ids": ["AI-056"], "searched": true, "queries": ["Quorum / Francisco Partners AI energy 2024", "Quorum / Francisco Partners AI oil and gas 2025", "Quorum / Francisco Partners AI energy 2026"], "documents_reviewed": [{"id": "sc-009", "title": "Quorum Software acquires Streamba", "url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "publisher": "Quorum Software", "date": "2026-06-25", "type": "company announcement", "locator": "Streamba/VOR and DaWinci", "review_status": "full text reviewed", "supports": ["Streamba AI-native supply chain platform and Visibility, Orchestration and Response system unify logistics data, workflows and teams.", "Quorum says agentic capabilities can identify disruption and propose or execute plan changes with human oversight."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Reviewed Quorum disclosure concerns acquired Streamba; no named customer deployment or measured outcome. Francisco Partners ownership transaction not separately audited here.", "cohort": "software_capital"}, {"company": "PakEnergy / Charlesbank", "existing_ids": ["AI-057", "AI-058"], "searched": true, "queries": ["PakEnergy / Charlesbank AI energy 2024", "PakEnergy / Charlesbank AI oil and gas 2025", "PakEnergy / Charlesbank AI energy 2026"], "documents_reviewed": [{"id": "sc-010", "title": "PakEnergy launches embedded AI-powered lease provision capability", "url": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "publisher": "PakEnergy", "date": "2025-09-25", "type": "company announcement", "locator": "Land Management lease provision", "review_status": "full text reviewed", "supports": ["Embedded AI processes lease provision research and data inside PakEnergy Land Management.", "Company says work previously taking hours can be done in seconds; no accuracy or independent outcome data disclosed."], "accessed": "2026-09-27"}, {"id": "sc-011", "title": "PakEnergy launches AI-powered Ticket-N-Receipt Management", "url": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "publisher": "PakEnergy", "date": "2026-05-20", "type": "company announcement", "locator": "PakCAPTURE workflow", "review_status": "full text reviewed", "supports": ["Drivers photograph receipts or paper tickets; PakAI extracts data, populates pickup/drop forms and associates information with a load.", "Feature embedded in PakTransportation and available to customers."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Two named capabilities are covered; underlying model, review safeguards and accuracy metrics are not disclosed. Charlesbank ownership/transaction not treated as AI.", "cohort": "software_capital"}, {"company": "Collide / Battalion", "existing_ids": ["AI-124", "AI-125", "AI-126"], "searched": true, "queries": ["Collide / Battalion AI energy 2024", "Collide / Battalion AI oil and gas 2025", "Collide / Battalion AI energy 2026"], "documents_reviewed": [{"id": "sc-012", "title": "Battalion Oil announces strategic investment in Collide", "url": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "publisher": "Battalion Oil Corporation", "date": "2026-09-08", "type": "company announcement", "locator": "Strategic investment, intended deployment and Riggs", "review_status": "full text reviewed", "supports": ["Battalion made strategic equity investment and will deploy Collide Operations System under subscription.", "Initial plan unites more than 100 TB of records, contracts and production history; planned use cases include gas reconciliation, field invoices, power, and failure diagnosis.", "Riggs is expected to be available to production engineers around day 90; human review/control retained."], "accessed": "2026-09-27"}, {"id": "sc-013", "title": "Collide product: Operations System", "url": "https://collide.io/product", "publisher": "Collide", "date": "undated", "type": "company webpage", "locator": "Architecture/capability page", "review_status": "excerpt reviewed", "supports": ["Collide describes petroleum-specific connected asset model and grounded answers with source citations.", "Company describes deterministic tested calculations and deployment in customer cloud/data platform; page undated."], "accessed": "2026-09-27"}, {"id": "sc-033", "title": "Collide secures $5m to scale first GenAI platform for energy sector", "url": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "publisher": "Collide (via Business Wire)", "date": "2025-04-15", "type": "company announcement", "locator": "Seed funding and energy knowledge platform", "review_status": "full text reviewed", "supports": ["Collide announced a $5m seed led by Mercury Fund; proceeds support platform and engineering expansion.", "Platform uses retrieval augmented generation and LLMs over curated technical papers, internal processes and Q&A to provide cited, domain-specific energy answers.", "Collide reported 6,000+ community users and estimated 25%-33% time savings for petroleum engineers; these are company estimates and community users are not equivalent to deployed enterprise users."], "accessed": "2026-09-27"}, {"id": "sc-034", "title": "Winn Resources partners with Collide to turn operational data into a competitive advantage", "url": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html", "publisher": "Winn Resources (issuer release syndicated by PRNewswire/StreetInsider)", "date": "2026-03-05", "type": "company announcement", "locator": "W-10/G-10 filing automation", "review_status": "full text reviewed", "supports": ["Winn says Collide automates monthly Texas RRC W-10/G-10 filings using existing well test/production data; task moves from multi-hour effort to under 30 minutes.", "Winn says workflow surfaces missing-data gaps earlier and reports >95% time reduction. This is customer-reported; no independent audit."], "accessed": "2026-09-27"}, {"id": "sc-035", "title": "How Winn Resources Slashed Regulatory Filing Time by 95%", "url": "https://collide.io/case-studies/winn-resources-ai-regulatory-filing-oil-gas", "publisher": "Collide", "date": "undated", "type": "partner case study", "locator": "Case-study page title; body extraction unavailable", "review_status": "excerpt reviewed", "supports": ["Collide has an official named Winn Resources case study for regulatory-filing time reduction; accessible page confirms title only."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Funding/platform announcement, Winn filing customer workflow, and Battalion strategic investment/rollout are separate initiatives. Winn metrics are issuer-reported; Collide case-study page body extraction was unavailable.", "cohort": "software_capital"}, {"company": "Applied Computing", "existing_ids": ["AI-127"], "searched": true, "queries": ["Applied Computing AI energy 2024", "Applied Computing AI oil and gas 2025", "Applied Computing AI energy 2026"], "documents_reviewed": [{"id": "sc-014", "title": "Applied Computing raises £9m seed to build Orbital AI", "url": "https://appliedcomputing.com/press-release-28052025", "publisher": "Applied Computing", "date": "2025-05-28", "type": "company announcement", "locator": "Orbital architecture, financing and initial refinery", "review_status": "full text reviewed", "supports": ["Applied Computing raised £9m seed funding for Orbital, a process-industry model using language, time-series, physics and chemical engineering models.", "Company says it is on edge, can work on sensitive process data, and reports one large refinery live.", "Cost savings, energy reduction and payback cited are company claims/potential, not independently supported."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Primary £9m seed source reviewed; alleged later $20m Series A remains secondary-only in the legacy profile and is not reverified. No independently named refinery.", "cohort": "software_capital"}, {"company": "Cognite", "existing_ids": ["AI-049", "AI-120", "AI-121"], "searched": true, "queries": ["Cognite AI energy 2024", "Cognite AI oil and gas 2025", "Cognite AI energy 2026"], "documents_reviewed": [{"id": "sc-015", "title": "Cognite Atlas AI major release accelerates industrial AI", "url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "publisher": "Cognite", "date": "2025-09-09", "type": "company announcement", "locator": "Agent workbench, templates, governance", "review_status": "full text reviewed", "supports": ["Atlas AI provides low-code agent workbench over industrial OT/IT/engineering data.", "Templates cover root-cause analysis, work package generation, cross-source research, handover summaries and work-order classification.", "Cognite names energy customers; announcement is product update not independent evaluation of customer results."], "accessed": "2026-09-27"}, {"id": "sc-030", "title": "Cognite and NVIDIA partner to advance industrial AI performance", "url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "publisher": "Cognite", "date": "2025-10-14", "type": "company announcement", "locator": "Celanese time-series anomaly detection", "review_status": "full text reviewed", "supports": ["Atlas AI/Data Fusion integrated with NVIDIA accelerated computing and software.", "Initial Celanese project targets real-time anomaly detection in industrial time-series data.", "Celanese reports over 50 global Cognite sites, but release does not say all sites use the NVIDIA model."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Atlas AI, Aker BP expansion and NVIDIA/Celanese use case reviewed. Celanese is chemicals, not O&G; NVIDIA announcement does not prove 50+ sites use the NVIDIA model.", "cohort": "software_capital"}, {"company": "Seeq", "existing_ids": ["AI-128"], "searched": true, "queries": ["Seeq AI energy 2024", "Seeq AI oil and gas 2025", "Seeq AI energy 2026"], "documents_reviewed": [{"id": "sc-016", "title": "Introducing Seeq Intelligence", "url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "publisher": "Seeq", "date": "2026-03-03", "type": "company announcement", "locator": "Agent Q, build-your-own agents and integrations", "review_status": "full text reviewed", "supports": ["Agent Q synthesizes real-time operational context, prior analyses, historical events, documents and domain expertise into investigations and recommended actions.", "Custom agents run multistep workflows on demand, scheduled or triggered; secure extensions can retrieve customer-system context and initiate workflows."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Product launch reviewed; no named oil/gas customer Agent Q deployment or measured operational result found in launch.", "cohort": "software_capital"}, {"company": "AVEVA / Emerson", "existing_ids": ["AI-129"], "searched": true, "queries": ["AVEVA / Emerson AI energy 2024", "AVEVA / Emerson AI oil and gas 2025", "AVEVA / Emerson AI energy 2026"], "documents_reviewed": [{"id": "sc-017", "title": "AVEVA announces new capabilities to embed AI", "url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "publisher": "AVEVA", "date": "2026-05-20", "type": "company announcement", "locator": "CONNECT, Flows and AI roadmap", "review_status": "full text reviewed", "supports": ["AVEVA announced AI data mapping with lineage/governance, industrial knowledge graph and Flows/connectors.", "Press release labels major CONNECT release for Q1 2027 as planned; roadmap is beyond research window and must not be counted as deployed in-window."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "AI announcement mixes current offerings, demonstration and future roadmap; no named oil/gas deployment. Q1 2027 CONNECT roadmap is out of window.", "cohort": "software_capital"}, {"company": "C3 AI", "existing_ids": ["AI-021", "AI-076"], "searched": true, "queries": ["C3 AI AI energy 2024", "C3 AI AI oil and gas 2025", "C3 AI AI energy 2026"], "documents_reviewed": [{"id": "sc-023", "title": "C3 AI and Baker Hughes renew and expand joint venture", "url": "https://ir.c3.ai/news-releases/news-release-details/c3-ai-and-baker-hughes-renew-and-expand-joint-venture-agreement", "publisher": "C3 AI", "date": "2025-05-28", "type": "company announcement", "locator": "Energy AI partnership / market deployment list", "review_status": "excerpt reviewed", "supports": ["C3 AI and Baker Hughes renewed agreement through June 2028 to develop and market energy/industrial AI.", "C3 AI says existing joint solutions had been deployed to named major operators; announcement gives broad use categories but little workflow implementation detail."], "accessed": "2026-09-27"}, {"id": "sc-024", "title": "Baker Hughes achieves bottom line savings with AI Sourcing Optimization", "url": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "publisher": "C3 AI", "date": "undated; CY2025 economics", "type": "partner case study", "locator": "AI sourcing optimization; value and scale", "review_status": "excerpt reviewed", "supports": ["Case study says C3 AI Sourcing Optimization supported 900,000 SKUs across 800+ sites and realized millions in economic benefit in CY2025.", "Workflow addresses fragmented procurement data and manual spreadsheet reconciliation.", "Vendor-published customer story; not independent assurance."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Renewed alliance and vendor/customer sourcing case reviewed; legacy named customer profiles are not individually re-audited here. C3/Baker Hughes outcomes are vendor-published.", "cohort": "software_capital"}, {"company": "Snowflake / Microsoft", "existing_ids": ["AI-119", "AI-108"], "searched": true, "queries": ["Snowflake / Microsoft AI energy 2024", "Snowflake / Microsoft AI oil and gas 2025", "Snowflake / Microsoft AI energy 2026"], "documents_reviewed": [{"id": "sc-018", "title": "Snowflake launches Energy Solutions", "url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "publisher": "Snowflake", "date": "2026-01-27", "type": "company announcement", "locator": "Energy Solutions, partner ecosystem, SAP and spatial data", "review_status": "full text reviewed", "supports": ["Energy Solutions is an industry offering combining Snowflake data platform and partner solutions to unify IT/OT/IoT data.", "Announcement names customers but does not substantiate a named oil and gas AI production deployment by them."], "accessed": "2026-09-27"}, {"id": "sc-019", "title": "Encino Energy streamlines operations with Microsoft Power Platform", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "publisher": "Microsoft", "date": "2025-04-29", "type": "partner case study", "locator": "Copilot Studio/data agent and invoice AI Builder details", "review_status": "full text reviewed", "supports": ["Encino plans direct Snowflake to Dataverse connector and is rolling out a Copilot Studio agent for questions across production data; a separate early prototype analyzes activity logs.", "Invoice app uses AI Builder document processing with confidence display and human approval routing.", "The same story describes many Power Platform apps and savings, most unrelated to AI; do not attribute the $1.2m winter maintenance savings to AI."], "accessed": "2026-09-27"}, {"id": "sc-020", "title": "TotalEnergies from Microsoft 365 Copilot to Copilot Studio", "url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "publisher": "Microsoft", "date": "2025-10-22", "type": "partner case study", "locator": "MARGE and BuyerCompanion", "review_status": "full text reviewed", "supports": ["MARGE uses Copilot Studio to classify intervention reports; company reports 25-45% improvement in fault-cause identification.", "BuyerCompanion streamlines purchases up to €50,000, operational in France, with company target of 10% savings on certain purchases.", "30,000 Copilot licenses, training and internal agent-validation committee described."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Snowflake launch is platform-level. Microsoft case stories add named Encino/TotalEnergies workflows; agent rollouts are early or not independently audited.", "cohort": "software_capital"}, {"company": "Ambyint", "existing_ids": [], "searched": true, "queries": ["Ambyint AI energy 2024", "Ambyint AI oil and gas 2025", "Ambyint AI energy 2026"], "documents_reviewed": [{"id": "sc-021", "title": "Infinity PL drives efficiency gains for Encino Energy", "url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "publisher": "Ambyint", "date": "2025-08-05", "type": "partner case study", "locator": "InfinityPL customer deployment", "review_status": "excerpt reviewed", "supports": ["Vendor case describes automated plunger-lift surveillance, anomaly detection and setpoint control at Encino Energy.", "Operators supervise exceptions and use dashboard/history; metrics are vendor-reported."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Encino customer case is vendor-published and page extraction was partial; exact public metrics could not be verified in accessible text.", "cohort": "software_capital"}, {"company": "Xecta", "existing_ids": ["AI-083"], "searched": true, "queries": ["Xecta AI energy 2024", "Xecta AI oil and gas 2025", "Xecta AI energy 2026"], "documents_reviewed": [{"id": "sc-029", "title": "Xecta news and resources", "url": "https://xecta.com/news-resources/", "publisher": "Xecta", "date": "June/July 2025 items", "type": "company webpage", "locator": "Santos Eastern Australia and PNG digital oilfield project", "review_status": "excerpt reviewed", "supports": ["Xecta news page lists Santos implementation of an Integrated Production System Model for continuous production optimization across Eastern Australia and Papua New Guinea.", "Underlying detailed release page could not be opened; vendor announcement, not independent customer validation."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "The named Santos deployment profile AI-083 is covered once in majors_backfill.json; this cohort reviewed the official Xecta index but did not find an additional distinct Xecta product/customer workflow to add.", "cohort": "software_capital"}, {"company": "Imubit", "existing_ids": [], "searched": true, "queries": ["Imubit AI energy 2024", "Imubit AI oil and gas 2025", "Imubit AI energy 2026"], "documents_reviewed": [{"id": "sc-022", "title": "Preem partners with Imubit for closed-loop AI optimization", "url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "publisher": "Imubit", "date": "2025-02-19", "type": "company announcement", "locator": "Lysekil FCC deployment", "review_status": "full text reviewed", "supports": ["Imubit and Preem announced Optimizing Brain at Lysekil refinery FCC to optimize process variables, yields and fuel-gas balance.", "Fuel savings and emissions reduction were anticipated; no measured post-deployment results were given."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Preem partnership describes intended FCC optimization; no post-deployment results or detailed control/safety design disclosed.", "cohort": "software_capital"}, {"company": "Kimmeridge", "existing_ids": [], "searched": true, "queries": ["Kimmeridge AI energy 2024", "Kimmeridge AI oil and gas 2025", "Kimmeridge AI energy 2026"], "documents_reviewed": [{"id": "sc-025", "title": "Kimmeridge research: Evolution of the APES", "url": "https://kimmeridge.com/research/", "publisher": "Kimmeridge", "date": "2025-03", "type": "company research page", "locator": "AI data center power market, strategic context", "review_status": "excerpt reviewed", "supports": ["Kimmeridge research addresses AI-related power demand; this is a market thesis / energy supply context, not AI adoption within an operating portfolio company."], "accessed": "2026-09-27"}, {"id": "sc-026", "title": "Kimmeridge selected investments", "url": "https://kimmeridge.com/select-investments/", "publisher": "Kimmeridge", "date": "undated", "type": "company webpage", "locator": "Portfolio screen", "review_status": "excerpt reviewed", "supports": ["Reviewed investment list for named operational AI deployments; no qualifying portfolio workflow found on page."], "accessed": "2026-09-27"}], "status": "no qualifying initiative found", "gaps": "Official research and investment pages screened; AI-related power demand is market research only, not an operating portfolio deployment. This bounded screen is recorded here and is not an initiative dossier.", "cohort": "software_capital"}, {"company": "Riverstone", "existing_ids": [], "searched": true, "queries": ["Riverstone AI energy 2024", "Riverstone AI oil and gas 2025", "Riverstone AI energy 2026"], "documents_reviewed": [{"id": "sc-027", "title": "Riverstone portfolio", "url": "https://www.riverstonerel.com/portfolio/", "publisher": "Riverstone", "date": "undated", "type": "company webpage", "locator": "Portfolio screen", "review_status": "excerpt reviewed", "supports": ["Current portfolio page reviewed for concrete oil and gas AI implementations; none identified on page."], "accessed": "2026-09-27"}], "status": "no qualifying initiative found", "gaps": "Only public portfolio page screened; no qualifying case found. Individual portfolio company materials were not comprehensively audited.", "cohort": "software_capital"}, {"company": "NGP", "existing_ids": [], "searched": true, "queries": ["NGP AI energy 2024", "NGP AI oil and gas 2025", "NGP AI energy 2026"], "documents_reviewed": [{"id": "sc-028", "title": "NGP Natural Resources portfolio", "url": "https://www.ngpenergycapital.com/portfolio-natural-resources/", "publisher": "NGP Energy Capital", "date": "undated", "type": "company webpage", "locator": "Portfolio screen", "review_status": "excerpt reviewed", "supports": ["Natural Resources portfolio list reviewed for concrete AI deployment evidence; page alone lists holdings and does not evidence deployments."], "accessed": "2026-09-27"}], "status": "no qualifying initiative found", "gaps": "Only Natural Resources portfolio listing screened; no qualifying case found. Individual portfolio company materials were not comprehensively audited.", "cohort": "software_capital"}, {"company": "Williams", "existing_ids": ["AI-064", "AI-103"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Williams Q2 2025 earnings call transcript", "url": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "date": "2025-08-05", "type": "earnings call transcript", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Williams Q2 2025 earnings event page", "url": "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025", "date": "2025-08-05", "type": "investor relations event page", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Williams Q2 2025 results", "url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "date": "2025-08-04", "type": "earnings release", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Williams Q2 2026 earnings presentation", "url": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "date": "2026-08-04", "type": "investor presentation", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Primary issuer Q2 2025 transcript verifies only one month-long DFW gas-marketing test; no broader live trading deployment or economics disclosed. Reviewed selected earnings/investor materials, not every quarterly source.", "cohort": "midstream"}, {"company": "Energy Transfer", "existing_ids": ["AI-062"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Energy Transfer and CloudBurst sign natural gas supply agreement", "url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "date": "2025-02-10", "type": "company announcement", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Entergy Louisiana and Energy Transfer sign agreement supporting Meta data center", "url": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "date": "2025-11-04", "type": "company announcement", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Energy Transfer Q1 2026 results and updates", "url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "date": "2026", "type": "earnings release", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Reviewed announced CloudBurst, Meta and Nexus gas/pipeline infrastructure for data-center demand. No qualifying internal AI workflow was identified in these reviewed sources; infrastructure is not adoption.", "cohort": "midstream"}, {"company": "Kinder Morgan", "existing_ids": ["AI-104"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Flyscan announces Kinder Morgan as strategic investor", "url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "date": "2025-01-14", "type": "partner announcement", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Kinder Morgan Q1 2025 earnings release", "url": "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx", "date": "2025-04-16", "type": "earnings release", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Flyscan investor-group participation verified in a partner release; no KMI-specific rollout found in reviewed Q1 earnings material. Other internal AI uses may exist.", "cohort": "midstream"}, {"company": "TC Energy", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "TC Energy 2025 Report on Sustainability", "url": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "date": "2025", "type": "annual report", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}, {"title": "TC Energy 2026 Management Information Circular", "url": "https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf", "date": "2026-03-05", "type": "proxy statement", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "TC Energy Pipeline IQ on AWS", "url": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "date": "undated", "type": "partner case study", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}], "status": "reviewed", "gaps": "Found company-reported ML commercial tool and AWS partner case for engineering agents. Pipeline IQ page lacks publication/launch date and partner case results are vendor/partner reported.", "cohort": "midstream"}, {"company": "Enbridge", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Enbridge 2025 Sustainability Report", "url": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "date": "2025", "type": "sustainability report", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}, {"title": "Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability", "url": "https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "date": "2026", "type": "company technical publication", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}, {"title": "Enbridge 2026 AI Policy", "url": "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152", "date": "2026", "type": "company policy", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}], "status": "reviewed", "gaps": "ML integrity prioritization reported; 2025 source describes development, while 2026 handbook describes predictive analysis without establishing it is the same tool. No measured outcomes disclosed.", "cohort": "midstream"}, {"company": "ONEOK", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "ONEOK Innovation Spotlight", "url": "https://sustainability.oneok.com/innovation-spotlight", "date": "undated", "type": "company sustainability webpage", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}, {"title": "ONEOK Q4 2025 earnings release", "url": "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638", "date": "2026", "type": "earnings release", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}], "status": "reviewed", "gaps": "Undated current sustainability page reports 48,872 AI virtual gauges. Current deployment scale is visible, but initial deployment/update date is not; cannot establish in-window launch.", "cohort": "midstream"}, {"company": "Enterprise Products Partners", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Enterprise Products Q4 2025 earnings release", "url": "https://ir.enterpriseproducts.com/news-releases/news-release-details/enterprise-reports-fourth-quarter-2025-earnings", "date": "2026", "type": "earnings release", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Enterprise Products Investor Presentation", "url": "https://ir.enterpriseproducts.com/static-files/5773e172-f0c8-4782-a251-410b43134201", "date": "undated", "type": "investor presentation", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Reviewed Q4 2025 issuer release and investor presentation; presentation treats AI/data-center load as market opportunity. No named internal AI workflow found; broader reporting review incomplete.", "cohort": "midstream"}, {"company": "Pembina Pipeline", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Pembina 2025 Annual Report", "url": "https://www.pembina.com/getmedia/8988c0d1-d213-4c6c-acdb-bea7c1be681e/Q4-2025-Annual-Report.pdf", "date": "2026", "type": "annual report", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Pembina 2025 Management Information Circular", "url": "https://www.pembina.com/getattachment/ff572a91-7f7b-4977-88ee-f0ae89d0106f/2025-information-circular-enhanced-for-website-.pdf", "date": "2025", "type": "proxy statement", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Reviewed annual report and proxy for AI risk/governance. No named operating AI workflow identified; search is not evidence of absence.", "cohort": "midstream"}, {"company": "Targa Resources", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Targa Principal Engineer, AI & Emerging Technology job posting", "url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "date": "2026-05-07", "type": "job posting", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Verified AI/technology operational hiring role, but no production use case. Broader earnings/annual-report review incomplete.", "cohort": "midstream"}, {"company": "Cheniere Energy", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Cheniere Q2 2025 results and updates", "url": "https://lngir.cheniere.com/news-events/press-releases/detail/324/cheniere-reports-second-quarter-2025-results-and-updates", "date": "2025", "type": "earnings release", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Reviewed Q2 2025 issuer earnings release; no named operational AI use found. LNG process applications beyond this source remain open.", "cohort": "midstream"}, {"company": "Venture Global", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Venture Global Q1 2025 quarterly investor presentation", "url": "https://s205.q4cdn.com/622838971/files/doc_financials/2025/q1/VG-Quarterly-Investor-Presentation_1Q2025_2025-05-13.pdf", "date": "2025-05-13", "type": "investor presentation", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Reviewed Q1 2025 investor presentation; no named AI workflow found. Annual reports and call transcripts were not comprehensively reviewed.", "cohort": "midstream"}, {"company": "Phillips 66", "existing_ids": ["AI-100", "AI-101"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Phillips 66 Q4 2025 earnings call transcript", "url": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "date": "2026-02-04", "type": "earnings call transcript", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Phillips 66 Q2 2026 earnings call transcript", "url": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "date": "2026-08", "type": "earnings call transcript", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Issuer Q4 2025 and Q2 2026 transcripts verify turnaround ML and broad frontline AI program. No AI-specific financial bridge or deployment inventory.", "cohort": "midstream"}, {"company": "Valero Energy", "existing_ids": ["AI-102"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Valero Q3 2025 earnings call transcript (The Motley Fool)", "url": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "date": "2025-10-23", "type": "secondary transcript", "accessed": "2026-09-27", "review_status": "full text reviewed"}, {"title": "Valero Q3 2025 earnings call event", "url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "date": "2025-10-23", "type": "investor relations event page", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "Issuer event page confirms Q3 2025 earnings call; detailed AI claims remain third-party transcript only and were not verified against issuer recording.", "cohort": "midstream"}, {"company": "Marathon Petroleum", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "Marathon Petroleum 2025 Annual Report", "url": "https://www.marathonpetroleum.com/content/documents/Invest/Annual_Reports/2025_Annual_Report.pdf", "date": "2026", "type": "annual report", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}, {"title": "Flyscan announces Kinder Morgan as strategic investor", "url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "date": "2025-01-14", "type": "partner announcement", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "partial", "gaps": "2025 annual report source and Flyscan investor announcement reviewed. No named current internal AI deployment verified; Flyscan operating footprint and timing for MPC not established.", "cohort": "midstream"}, {"company": "HF Sinclair", "existing_ids": [], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "HF Sinclair 2025 Annual Report", "url": "https://investor.hfsinclair.com/investor-relations/financial-information/sec-filings/sec-filings-details/default.aspx?FilingId=19194360", "date": "2026", "type": "annual report index", "accessed": "2026-09-27", "review_status": "excerpt reviewed"}], "status": "partial", "gaps": "Annual report index used as search lead; exact report and detailed in-window earnings materials not confirmed. No qualifying initiative asserted.", "cohort": "midstream"}, {"company": "ECP", "existing_ids": ["AI-059"], "searched": true, "queries": ["Company AI/ML + 2024/2025/2026 earnings calls, investor presentations, annual/sustainability reports", "Company + named AI workflow, operational technology, pipeline/refinery/LNG/trading use case", "Existing profile URL and claim verification where a record exists"], "documents_reviewed": [{"title": "ECP and KKR announce Texas hyperscale data-center campus", "url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "date": "2025-07-30", "type": "company announcement", "accessed": "2026-09-27", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Verified ECP/KKR/CyrusOne campus as AI-demand infrastructure only; no internal AI adoption source identified.", "cohort": "midstream"}, {"company": "Aramco", "existing_ids": ["AI-037", "AI-038", "AI-039", "AI-070"], "searched": true, "queries": ["site:aramco.com AI Technology Realized Value 2025 2026 MetaBrain"], "documents_reviewed": [{"title": "Aramco deploys AI and supercomputing to drive digital transformation", "url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "date": "2025-11-24", "type": "Company announcement / remarks"}, {"title": "CEO Opening Remarks from the 2025 Full Year Financial Results Media Call", "url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "date": "2026-03-10", "type": "Company announcement / remarks"}, {"title": "FY 2025 results presentation and script", "url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "date": "2026-03-10", "type": "Investor presentation"}, {"title": "Aramco signs MoU with Microsoft to help advance industrial AI and digital talent transformation", "url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "date": "2026-02-12", "type": "Company announcement / remarks"}, {"title": "Remarks by Ahmad O. Al-Khowaiter at LEAP 2025", "url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "date": "2025-02-09", "type": "Company announcement / remarks"}, {"title": "Aramco announces fourth quarter and full-year 2025 results", "url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "date": "2026-03-10", "type": "Company announcement / remarks"}], "status": "Targeted primary-source backfill complete", "gaps": "Not every quarterly call reviewed. HUMAIN closing and workflow-level TRV allocations not established.", "cohort": "root_backfill"}, {"company": "Chevron", "existing_ids": ["AI-001", "AI-002", "AI-003", "AI-004", "AI-005", "AI-075"], "searched": true, "queries": ["Chevron official AI 2024", "site:chevron operational AI 2025/2026", "legacy profile source audit AI-001,AI-002,AI-003,AI-004,AI-005,AI-075"], "documents_reviewed": [{"id": "mb-001", "title": "2025 Chevron Investor Day edited transcript", "url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "publisher": "Chevron", "date": "2025-11-12", "type": "investor presentation", "locator": "Slides 39–40; transcript pp. 21–22", "review_status": ["APOLO launched Aug 2024 (pre-window); Nov 2025 update describes >50,000 wells, finer prospectivity mapping and model training for spacing, completion and development timing.", "Chevron described an AI team and prioritization of high-value workflows including shale/tight and exploration.", "Chevron’s “billions” annual value is a potential decade-end ambition, not realized savings."], "supports": [], "accessed": "2026-09-27"}, {"id": "mb-002", "title": "DataRobot and Chevron Collaborate to Advance Agentic AI for Autonomous Inspections", "url": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "publisher": "DataRobot", "date": "2026-06-02", "type": "partner case study", "locator": "Safe Start agent, robot missions and edge systems", "review_status": ["Chevron and DataRobot collaboration applies edge agents to support autonomous inspection at Chevron facilities.", "Safe Start assesses conditions before/during robotic missions using gas sensors, vision and locally running models; integrates with existing operating systems and NVIDIA NIM.", "Vendor and Chevron program manager describe reduced manual point-in-time checks and human oversight/control; no measured outcomes."], "supports": [], "accessed": "2026-09-27"}], "status": "partial", "gaps": "APOLO was launched Aug 2024 before window; Nov 2025 update qualifies. AI-002 shares exploration context but not separately verified beyond APOLO. AI-003/004 are broad workforce/value ambitions; AI-005 is AI-demand power infrastructure. AI-075 has partner announcement detail but no measured results.", "cohort": "majors_backfill"}, {"company": "Shell", "existing_ids": ["AI-010", "AI-025", "AI-026"], "searched": true, "queries": ["Shell official AI 2024", "site:shell operational AI 2025/2026", "legacy profile source audit AI-010,AI-025,AI-026"], "documents_reviewed": [{"id": "mb-003", "title": "SLB Enters Collaboration Agreement to Accelerate New Digital Solutions", "url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "publisher": "SLB", "date": "2025-12-11", "type": "company announcement", "locator": "Shell collaboration agreement", "review_status": "full text reviewed", "supports": ["SLB and Shell agreed to develop upstream digital and AI solutions; roadmap aims for open data/AI infrastructure across subsurface, well construction and production on SLB Lumi.", "Announcement states ambition/aim, not a deployed customer workflow or realized results."], "accessed": "2026-09-27"}, {"id": "mb-004", "title": "SLB Q4 and full-year 2025 results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "publisher": "SLB", "date": "2026-01-23", "type": "earnings release", "locator": "Digital segment customer collaborations section", "review_status": "full text reviewed", "supports": ["Q4 report repeats Shell collaboration aims and scope; this is an in-window disclosure/update, not evidence the collaboration was launched in January or deployed."], "accessed": "2026-09-27"}, {"id": "mb-017", "title": "Shell Capital Markets Day 2025", "url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "publisher": "Shell", "date": "2025-03", "type": "investor presentation", "locator": "p. 15 (cost reduction); AI infrastructure/cooling product", "review_status": "full text reviewed", "supports": ["Shell presents $5–7bn cumulative structural cost-reduction target from 2022–2028; examples include technology and AI, alongside procurement and corporate simplification.", "This is not an AI-only savings forecast or operational workflow. AI infrastructure cooling fluid appears as a separate commercial product/AI-demand case, not Shell AI adoption."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "AI-010 confirmed as agreement/roadmap, not deployment. AI-025 target includes many non-AI levers and is not AI-attributed. AI-026 concerns product for data-center cooling, not AI adoption.", "cohort": "majors_backfill"}, {"company": "ExxonMobil", "existing_ids": ["AI-022", "AI-023", "AI-024", "AI-069", "AI-073"], "searched": true, "queries": ["ExxonMobil official AI 2024", "site:exxonmobil operational AI 2025/2026", "legacy profile source audit AI-022,AI-023,AI-024,AI-069,AI-073"], "documents_reviewed": [{"id": "mb-010", "title": "ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence", "url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "publisher": "ExxonMobil", "date": "2025-10-23", "type": "company webpage", "locator": "Production bottleneck monitoring", "review_status": "full text reviewed", "supports": ["ExxonMobil India describes using AI with production staff to monitor equipment in real time and identify current or prospective bottlenecks affecting production.", "Engineers and asset operators use domain knowledge to investigate and remove bottlenecks; no basin/site counts or metrics."], "accessed": "2026-09-27"}, {"id": "mb-011", "title": "SLB Q2 2026 results: ExxonMobil Trinidad & Tobago seismic project", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "date": "2026-07-24", "type": "partner/customer announcement", "locator": "Digital/technology projects section", "review_status": "full text reviewed", "supports": ["ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Official operating bottleneck disclosure supports broad AI use but not Permian-specific scope of AI-022. AI-024 ERP/logistics claim remained secondary transcript-only and no primary workflow verified. AI-073 four Guyana prospects remained secondary only; no Exxon primary disclosure located. AI-023 is energy supply/data-center opportunity, not internal adoption.", "cohort": "majors_backfill"}, {"company": "TotalEnergies", "existing_ids": ["AI-030"], "searched": true, "queries": ["TotalEnergies official AI 2024", "site:totalenergies operational AI 2025/2026", "legacy profile source audit AI-030"], "documents_reviewed": [{"id": "mb-016", "title": "TotalEnergies to collaborate with Mistral AI", "url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "publisher": "TotalEnergies", "date": "2025-06-12", "type": "company announcement", "locator": "Digital Factory baseline and joint AI lab", "review_status": "full text reviewed", "supports": ["TotalEnergies says its Digital Factory had 300 digital experts and developed >100 solutions, 60 using technologies from ML to GenAI.", "The announcement creates a distinct Mistral joint innovation lab for researcher assistant, industrial asset decision support, customer and renewable use cases; it says those use cases will be designed/tested, not deployed."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Digital Factory counts disclosed; no workflow-level outcome. Mistral lab is a separate related AI-029 initiative and its named use cases were plans to test/design.", "cohort": "majors_backfill"}, {"company": "Repsol", "existing_ids": ["AI-032"], "searched": true, "queries": ["Repsol official AI 2024", "site:repsol operational AI 2025/2026", "legacy profile source audit AI-032"], "documents_reviewed": [{"id": "mb-005", "title": "Repsol launches intelligent multi-agent platform", "url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "publisher": "Repsol", "date": "2025-10-15", "type": "company announcement", "locator": "Competence Center, platform pilot and user adoption", "review_status": "full text reviewed", "supports": ["Repsol says its GenAI Competence Center had produced 60+ use cases in production and 22 scaled; 5,000+ employees trained.", "Platform was a pilot in Digitalization and Services with 3 teams and 22 agents supporting processes for more than four months; 50+ people supervised agents.", "A separate 2025 consolidated management report later says 34 agents across three workflows; dates/scope differ, so counts should not be blended."], "accessed": "2026-09-27"}, {"id": "mb-006", "title": "2025 Consolidated Management Report", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "publisher": "Repsol", "date": "2026; exact publication date not stated", "type": "annual report", "locator": "Section 5.5 Technology and digitalization", "review_status": "excerpt reviewed", "supports": ["2025 report says Competence Center promoted >75 initiatives after two years, and the multi-agent platform had 34 agents in three workflows.", "Report names automated well-data ingestion for daily drilling/geology reports, integration into corporate systems and faster exploration decision-making."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Competence center/program metrics vary by date and source; AI-031 multiagent platform is a separate exact initiative and related context.", "cohort": "majors_backfill"}, {"company": "Equinor", "existing_ids": ["AI-033", "AI-034", "AI-035", "AI-036"], "searched": true, "queries": ["Equinor official AI 2024", "site:equinor operational AI 2025/2026", "legacy profile source audit AI-033,AI-034,AI-035,AI-036"], "documents_reviewed": [{"id": "mb-007", "title": "Equinor: AI delivered USD 130m value in 2025", "url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "publisher": "Equinor", "date": "2026-01-07", "type": "company announcement", "locator": "Implemented applications and value disclosure", "review_status": "full text reviewed", "supports": ["Company says AI generated USD 130m value for Equinor and partners in calendar 2025, with >USD 330m cumulative since 2020.", "Three applications: 700+ rotating machines/24,000 sensors predictive maintenance (USD 120m since 2020); well and field planning, including USD12m Johan Sverdrup phase-3 saving; seismic interpretation of 2m km2 in 2025 with tenfold capacity."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "One official 2026 disclosure supports three implemented components and total 2025 value; the three subcase economics differ in period and may overlap.", "cohort": "majors_backfill"}, {"company": "Woodside", "existing_ids": ["AI-041", "AI-042", "AI-043"], "searched": true, "queries": ["Woodside official AI 2024", "site:woodside operational AI 2025/2026", "legacy profile source audit AI-041,AI-042,AI-043"], "documents_reviewed": [{"id": "mb-008", "title": "Woodside Artificial Intelligence", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "publisher": "Woodside", "date": "updated 2026-05-27; page undated", "type": "company webpage", "locator": "Lumina, Copilot, ChatGPT Enterprise and governance", "review_status": "full text reviewed", "supports": ["Woodside says Lumina supports >50 agents for business tasks; Microsoft Copilot supports Office productivity; ChatGPT Enterprise is an AI sandbox before custom agents integrate with Lumina.", "Responsible AI assessment, Centre of Excellence and executive AI Council described.", "Same page includes Maint Intel and Startup Advisor, already treated as separate exact profiles AI-042/AI-043; page does not state initial deployment dates."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Current AI webpage last updated 2026-05-27. Lumina covered; Maint Intel and Startup Advisor are separate exact IDs already covered in majors cohort. Initial deployment dates absent on current page.", "cohort": "majors_backfill"}, {"company": "Aker BP", "existing_ids": ["AI-049", "AI-131"], "searched": true, "queries": ["Aker BP official AI 2024", "site:akerbp operational AI 2025/2026", "legacy profile source audit AI-049,AI-131"], "documents_reviewed": [{"id": "mb-009", "title": "Aker BP leverages Cognite Atlas AI to pioneer AI-first future", "url": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "publisher": "Cognite", "date": "2025-09-16", "type": "partner case study", "locator": "RCA agents and additional workflows", "review_status": "full text reviewed", "supports": ["Aker BP expanded Atlas AI adoption. Described workflows include document processing for supply chain, barrier management, automatic equipment registration, data foundation for predictive maintenance and production chemistry.", "Root-cause agents identify relevant info, build cause maps and suggest steps; Cognite/Aker BP report >70% reduction in engineer RCA time."], "accessed": "2026-09-27"}, {"id": "mb-014", "title": "Aker BP builds AI-first future with Microsoft 365 Copilot and AI agents", "url": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "publisher": "Microsoft", "date": "2025-06-23", "type": "partner case study", "locator": "Employee Copilot rollout, governance, Knowledge Agents", "review_status": "full text reviewed", "supports": ["Aker BP started with 300 Copilot licenses and planned all-staff licensing in 2025; trained employees, set up help desk and champion/accountability network.", "Employee/small-team-built agents, including Knowledge Agents, answer repetitive documentation questions and route harder queries; several agents live.", "Microsoft case quotes 96% active users in prior four weeks; vendor-published customer metric, not independent audit."], "accessed": "2026-09-27"}], "status": "reviewed", "gaps": "Aker/Cognite source is partner-published and reports >70% RCA time reduction without sample/baseline. Microsoft story reports Copilot adoption; 96% activity is vendor-customer metric.", "cohort": "majors_backfill"}, {"company": "bp", "existing_ids": ["AI-060"], "searched": true, "queries": ["bp official AI 2024", "site:bp operational AI 2025/2026", "legacy profile source audit AI-060"], "documents_reviewed": [{"id": "mb-018", "title": "bp 3Q 2024 results Q&A transcript", "url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/investors/bp-third-quarter-2024-results-qa-transcript.pdf", "publisher": "bp", "date": "2024-10-29", "type": "earnings call", "locator": "Q&A; refining/digital remarks", "review_status": "full text reviewed", "supports": ["bp leadership said Palantir and Infosys had completed an ontology and were starting to make headway on digital in refining.", "Transcript language does not establish AI model or named deployed AI workflow; retain as digital transformation disclosure."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Official Q3 2024 transcript says Palantir/Infosys ontology was complete and digital work moving into action; it does not establish an AI workflow. AI classification remains unverified.", "cohort": "majors_backfill"}, {"company": "PETRONAS", "existing_ids": ["AI-079"], "searched": true, "queries": ["PETRONAS official AI 2024", "site:petronas operational AI 2025/2026", "legacy profile source audit AI-079"], "documents_reviewed": [{"id": "mb-012", "title": "PETRONAS and Microsoft expand collaboration to support Malaysia AI ecosystem", "url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "publisher": "PETRONAS", "date": "2025-07-31", "type": "company announcement", "locator": "MoU scope", "review_status": "full text reviewed", "supports": ["The MoU says parties will explore agentic AI, Microsoft Copilot, analytics, cloud and cybersecurity with Malaysia West cloud region.", "It is exploratory partnership language, not proof of rollout."], "accessed": "2026-09-27"}, {"id": "mb-013", "title": "PETRONAS-incubated AIngineer joins forces with Microsoft", "url": "https://www.petronas.com/bm/node/1668", "publisher": "PETRONAS", "date": "2025-06-10", "type": "company announcement", "locator": "AIngineer industrial engineer copilot", "review_status": "full text reviewed", "supports": ["AIngineer is a PETRONAS-incubated industrial AI copilot using LLMs and ML and integrating technical databases and standards.", "PETRONAS says it has technical use cases and prior deployments in energy/government; this release does not name a customer asset or quantify results.", "Microsoft collaboration includes cloud scaling and prospective marketplace commercialization."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "AI-079 MoU is exploratory. Additional AIngineer workflow/product disclosure is more concrete but does not identify an asset or quantified PETRONAS deployment.", "cohort": "majors_backfill"}, {"company": "Santos", "existing_ids": ["AI-083"], "searched": true, "queries": ["Santos official AI 2024", "site:santos operational AI 2025/2026", "legacy profile source audit AI-083"], "documents_reviewed": [{"id": "mb-015", "title": "Xecta news and resources index", "url": "https://xecta.com/news-resources/", "publisher": "Xecta", "date": "June/July 2025 listing; exact publication date unavailable", "type": "company webpage", "locator": "Santos Eastern Australia and PNG digital oilfield listing", "review_status": "excerpt reviewed", "supports": ["Xecta index identifies Santos integrated production-system model/digital oilfield across Eastern Australia and PNG.", "Underlying dated technical release did not open; current confirmation limited to vendor page index."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Xecta official index lists Santos production model; underlying dated release unavailable. Cooper Basin history, five-year terms and autonomous Kelvin.ai control remain unverified.", "cohort": "majors_backfill"}, {"company": "Emerson / AspenTech", "existing_ids": ["AI-130"], "searched": true, "queries": ["Emerson AspenTech acquisition completion March 2025 official AI", "AspenTech industrial AI workflow acquisition"], "documents_reviewed": [{"id": "mb-019", "title": "Emerson Completes Acquisition of Remaining Outstanding Shares of AspenTech", "url": "https://ir.emerson.com/news-events/press-releases/detail/11/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "publisher": "Emerson", "date": "2025-03-12", "type": "company announcement", "locator": "Transaction completion", "review_status": "excerpt reviewed", "supports": ["Emerson completed acquisition of remaining AspenTech shares on March 12, 2025; AspenTech became wholly owned.", "The release confirms ownership change and price, not a specific AI deployment or operational workflow."], "accessed": "2026-09-27"}], "status": "partial", "gaps": "Primary source confirms acquisition completion and $265/share cash consideration; AI-130 should remain transaction context, not a separate operational AI deployment. AspenTech product workflows were not audited in this bounded pass.", "cohort": "majors_backfill"}, {"company": "Indian Oil Corporation (IndianOil)", "existing_ids": [], "searched": true, "queries": ["site:iocl.com annual report 2025-26 AI Vishleshan; IndianOil AI predictive maintenance refinery"], "documents_reviewed": [{"title": "IndianOil – Intellectual Capital, Integrated Annual Report 2024-25", "url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "date": "2025 (FY2024-25)", "type": "annual report webpage"}], "status": "reviewed", "gaps": "Reviewed FY2024-25 AI/predictive-maintenance and FY2025-26 SPRINT material; predictive maintenance is described broadly without site, asset-class or outcome detail. The FY2025-26 page’s opened text did not expose the Vishleshan AI claim returned in search excerpts, so it is not treated as verified here.", "cohort": "regional"}, {"company": "Bharat Petroleum Corporation (BPCL)", "existing_ids": [], "searched": true, "queries": ["site:bharatpetroleum.in annual report 2024-25 AI predictive analytics UCVA"], "documents_reviewed": [{"title": "Annual Report 2024-25", "url": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "date": "2025 (FY2024-25 report)", "type": "annual report"}], "status": "reviewed", "gaps": "Reviewed FY2024-25 annual report. Two named refinery workflows are in scope; results and detailed technical architecture are not disclosed.", "cohort": "regional"}, {"company": "Hindustan Petroleum Corporation (HPCL)", "existing_ids": [], "searched": true, "queries": ["site:hindustanpetroleum.com 2026 LC-Max Digital Suite AI Visakh"], "documents_reviewed": [{"title": "HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery", "url": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "date": "2026-01-05", "type": "company announcement"}], "status": "reviewed", "gaps": "Reviewed HPCL’s Jan. 2026 commissioning announcement for the Visakh RUF and its Digital Suite; no AI-specific attribution for refinery-wide yield uplift.", "cohort": "regional"}, {"company": "Reliance Industries Limited (RIL)", "existing_ids": [], "searched": true, "queries": ["site:ril.com FY26 AI crude blend optimization O2C"], "documents_reviewed": [{"title": "Chairman’s Statement, 49th Annual General Meeting", "url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "date": "2026-06 (PDF path; exact AGM date not represented here)", "type": "shareholder statement"}], "status": "partial", "gaps": "Reviewed FY2025-26 chairman statement/transcript. It names an AI crude-feedstock blending optimizer but gives no method, scale, economics or outcome measure.", "cohort": "regional"}, {"company": "Sonatrach", "existing_ids": [], "searched": true, "queries": ["site:sonatrach.com AI artificial intelligence 2025 operations"], "documents_reviewed": [{"title": "Sonatrach News, issue 45 (AI seminar item)", "url": "https://sonatrach.com/wp-content/uploads/2025/06/SH-NEWS-VF-N%C2%B045.pdf", "date": "2025-06 (issue includes 2024 event)", "type": "company newsletter PDF"}], "status": "partial", "gaps": "Official 2024-26 search surfaced general digital/R&D material and an AI event outside the window; no named in-window operational AI workflow verified.", "cohort": "regional"}, {"company": "NNPC Limited", "existing_ids": [], "searched": true, "queries": ["site:nnpcgroup.com AI predictive maintenance 2025"], "documents_reviewed": [{"title": "NNPC Gas Master Plan 2026", "url": "https://cms1977.nnpcgroup.com/uploads/NNPC_GMP_2026_Report_Final_For_Upload_51bfeafda6.pdf", "date": "2026", "type": "company strategy report"}], "status": "partial", "gaps": "Official search surfaced conventional maintenance and broad digital-transformation material; no named in-window AI workflow verified.", "cohort": "regional"}, {"company": "Sonangol", "existing_ids": [], "searched": true, "queries": ["site:sonangol.co.ao inteligência artificial 2025 operations; annual report 2025"], "documents_reviewed": [{"title": "Annual Report and Accounts 2025", "url": "https://www.sonangol.co.ao/wp-content/uploads/2026/05/Relatorio-Anual-e-Contas-2025.pdf", "date": "2026-05 (publication path)", "type": "annual report PDF"}], "status": "partial", "gaps": "Reviewed the 2025 annual-report digital systems discussion and official strategy material. Digital modernization is described, but no operational AI workflow or AI deployment evidence found.", "cohort": "regional"}, {"company": "INPEX", "existing_ids": [], "searched": true, "queries": ["site:inpex.com AI generative AI digital strategy 2025"], "documents_reviewed": [{"title": "Digital Transformation Strategy", "url": "https://www.inpex.com/english/business/technology/dx/", "date": "Undated; accessed 2026-09-27", "type": "company strategy webpage"}], "status": "partial", "gaps": "Reviewed current official digital-strategy material describing internal GenAI and AI-enabled security examples, but the page is undated and in-window launch/deployment timing could not be verified.", "cohort": "regional"}, {"company": "PTT Public Company Limited (PTT)", "existing_ids": [], "searched": true, "queries": ["site:pttplc.com artificial intelligence operations 2025"], "documents_reviewed": [], "status": "partial", "gaps": "Official search did not identify a specific parent-company operating AI workflow in the period; PTTGC evidence is reported separately.", "cohort": "regional"}, {"company": "PTT Global Chemical (PTTGC / GC)", "existing_ids": [], "searched": true, "queries": ["site:pttgcgroup.com AI Smart Plant HERO 2026"], "documents_reviewed": [{"title": "GC Advances Holistic Optimization Through Digital Technology", "url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "date": "2026-07-20", "type": "company announcement"}], "status": "reviewed", "gaps": "Reviewed GC’s 20 July 2026 operational technology disclosure naming plant-wide optimization, asset intelligence/maintenance, AI vision and HERO heat integration.", "cohort": "regional"}, {"company": "Petrovietnam", "existing_ids": [], "searched": true, "queries": ["site:petrovietnam.vn AI trí tuệ nhân tạo 2025 oil gas operations"], "documents_reviewed": [{"title": "Artificial intelligence in the oil and energy sector: roadmap and realization", "url": "https://petrovietnam.petrotimes.vn/tri-tue-nhan-tao-trong-linh-vuc-dau-khi-nang-luong-lo-trinh-va-hien-thuc-hoa-732289.html", "date": "2025-09-16", "type": "secondary article"}], "status": "partial", "gaps": "Group-affiliated coverage points to subsidiary AI plans and applications, but no suitable primary company source verifying a named operational workflow in-window was established.", "cohort": "regional"}, {"company": "Gazprom Neft", "existing_ids": [], "searched": true, "queries": ["site:gazprom-neft.com AI autonomous drilling 2025"], "documents_reviewed": [{"title": "Gazprom Neft official Telegram post", "url": "https://t.me/s/gazpromneft_official/503", "date": "Undated; accessed 2026-09-27", "type": "official social post"}], "status": "partial", "gaps": "Official search surfaced smart-drilling/autopilot claims without a clearly dated in-window primary disclosure; no qualifying named workflow independently verified for this screen.", "cohort": "regional"}, {"company": "Rosneft", "existing_ids": [], "searched": true, "queries": ["site:rosneft.ru EchoTools искусственный интеллект 2025; LLM ассистент 2025"], "documents_reviewed": [{"title": "Rosneft improves well research methods (EchoTools)", "url": "https://rosneft.ru/press/news/item/222577", "date": "2025-07-17", "type": "company announcement"}, {"title": "Rosneft increases production through innovative software (EchoTools update)", "url": "https://www.rosneft.ru/press/news/item/224545/", "date": "2026-08-20", "type": "company announcement"}, {"title": "Rosneft introduces new technologies using artificial intelligence (LLM assistant)", "url": "https://rosneft.ru/press/news/item/222604/", "date": "2025-07-28", "type": "company announcement"}], "status": "reviewed", "gaps": "Reviewed official July 2025 and August 2026 EchoTools disclosures plus the July 2025 internal LLM-assistant announcement. EchoTools has workflow and company-reported metrics; assistant has task coverage and workload reduction claim.", "cohort": "regional"}, {"company": "APA Corporation", "existing_ids": [], "searched": true, "queries": ["APA Corporation AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "APA Corporation expands partnership with Palantir to leverage AI technology across oil and gas operations", "url": "https://www.nasdaq.com/press-release/apa-corporation-expands-partnership-palantir-leverage-ai-technology-across-their-oil", "date": "2024-09-25", "type": "company announcement", "locator": "", "review_status": "excerpt reviewed"}, {"title": "APA Corporation 2025 Form 10-K", "url": "https://www.sec.gov/Archives/edgar/data/1692787/000169278726000048/apa-20251231.htm", "date": "2026-03", "type": "SEC filing", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "The Palantir AI expansion announcement predates the window by one day (2024-09-25). 2025 10-K mentions generic algorithmic/AI evaluation and controls, but no named operating workflow surfaced. No qualifying new initiative added.", "cohort": "round2/additions-1"}, {"company": "Apache Corporation (historical APA subsidiary)", "existing_ids": [], "searched": true, "queries": ["Apache Corporation (historical APA subsidiary) AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Apache 2024 Form 10-K (APA annual filing)", "url": "https://www.sec.gov/Archives/edgar/data/1692787/000169278726000048/apa-20251231.htm", "date": "2026-03", "type": "SEC filing", "locator": "", "review_status": "full text reviewed"}, {"title": "APA Palantir partnership (announcement)", "url": "https://www.nasdaq.com/press-release/apa-corporation-expands-partnership-palantir-leverage-ai-technology-across-their-oil", "date": "2024-09-25", "type": "company announcement", "locator": "", "review_status": "excerpt reviewed"}], "status": "reviewed", "gaps": "No separate qualifying Apache-only initiative surfaced. Apache is an APA operating subsidiary, and the surfaced Palantir expansion is one day outside the window; do not duplicate it as a separate company record.", "cohort": "round2/additions-1"}, {"company": "Hess Corporation (standalone historical entity)", "existing_ids": [], "searched": true, "queries": ["Hess Corporation (standalone historical entity) AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Hess Corporation 2024 Form 10-K", "url": "https://www.sec.gov/Archives/edgar/data/4447/000162828025008716/hes-20241231.htm", "date": "2025-02-27", "type": "SEC filing", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "SEC filing AI mentions are risk/cybersecurity context, not evidence of Hess deploying AI. Hess standalone operations ceased after Chevron merger close, outside this reporting lookback for operating Hess workflows.", "cohort": "round2/additions-1"}, {"company": "Range Resources", "existing_ids": [], "searched": true, "queries": ["Range Resources AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Range Resources 2025 Form 10-K", "url": "https://ir.rangeresources.com/static-files/df4cbb92-55be-4ec9-9cd1-1a027e7e61ec", "date": "2026-02-24", "type": "SEC filing", "locator": "", "review_status": "excerpt reviewed"}, {"title": "Range Resources 2026 proxy", "url": "https://ir.rangeresources.com/static-files/8e35b66a-2542-485c-bc13-fb15745d940b", "date": "2026-03-31", "type": "company announcement", "locator": "", "review_status": "excerpt reviewed"}], "status": "reviewed", "gaps": "AI references concern external demand for reliable gas for data centers and cyber risk, not an internal AI workflow. Keep AI-demand context out of adoption records.", "cohort": "round2/additions-1"}, {"company": "Antero Resources", "existing_ids": [], "searched": true, "queries": ["Antero Resources AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Antero Resources Q2 2025 results", "url": "https://www.anteroresources.com/investors/press-releases/detail/247/antero-resources-announces-second-quarter-2025-financial", "date": "2025-07-30", "type": "company announcement", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "CEO links forecast gas demand to AI data centers; it is market context, not an operational AI deployment.", "cohort": "round2/additions-1"}, {"company": "SM Energy", "existing_ids": [], "searched": true, "queries": ["SM Energy AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "SM/Civitas merger joint conference-call transcript; SM 2025 10-K", "url": "https://www.sm-energy.com/investors/financial-information/all-sec-filings/content/0000893538-25-000149/smandcivitasmergerjointc.htm", "date": "2025-11-03", "type": "SEC filing", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Management described anticipated merger synergies from optimizing operations with ML and AI, but supplied no named workflow, implementation details, realized AI outcomes, or AI-attributable economics. Not elevated to a stand-alone initiative.", "cohort": "round2/additions-1"}, {"company": "Civitas Resources", "existing_ids": [], "searched": true, "queries": ["Civitas Resources AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "SM/Civitas merger joint conference-call transcript; joint merger presentation", "url": "https://www.sm-energy.com/investors/financial-information/all-sec-filings/content/0000893538-25-000149/smandcivitasmergerjointc.htm", "date": "2025-11-03", "type": "company announcement", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "The combined-company synergy statement refers to anticipated optimization using ML/AI; no defined use case or realized deployment. This is plan-level evidence, not a practical workflow dossier.", "cohort": "round2/additions-1"}, {"company": "Permian Resources", "existing_ids": [], "searched": true, "queries": ["Permian Resources AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Q3 2025 Earnings Presentation", "url": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "date": "2025-11-05", "type": "investor presentation", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Slide 5 connects AI-driven subsurface insight and precision wellbore placement to expanding play boundaries; it also describes an early-Q3 2025 Haley 17-well development and separate ~45% outperformance vs offsets. The presentation does not attribute that performance specifically to AI. A sparse, bounded initiative is added.", "cohort": "round2/additions-1"}, {"company": "Cheniere Energy", "existing_ids": [], "searched": true, "queries": ["Cheniere Energy AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Cheniere 2025 Form 10-K", "url": "https://lngir.cheniere.com/sec-filings/all-sec-filings/content/0000003570-26-000005/lng-20251231.htm", "date": "2026-02", "type": "SEC filing", "locator": "", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "The annual filing was searched for artificial intelligence; no matching AI text surfaced. No practical named AI initiative identified.", "cohort": "round2/additions-1"}, {"company": "Gulfport Energy", "existing_ids": [], "searched": true, "queries": ["Gulfport Energy AI artificial intelligence machine learning 2024 2025 2026", "site:{company.lower().replace(\" \",\"\")} AI operations annual report"], "documents_reviewed": [{"title": "Gulfport Energy 2025 Form 10-K", "url": "https://www.gulfportenergy.com/investors/financials/sec-filings/default.aspx", "date": "2026 (listing pages)", "type": "SEC filing", "locator": "", "review_status": "excerpt reviewed"}, {"title": "Gulfport Energy investor materials", "url": "https://www.gulfportenergy.com/investors/financials/quarterly-results/default.aspx", "date": "2026 (listing pages)", "type": "company announcement", "locator": "", "review_status": "excerpt reviewed"}], "status": "partial", "gaps": "Queries for Gulfport AI/ML operations did not surface a qualifying named workflow or primary-source case. Access to stable filing URL not confirmed in this limited pass; no initiative added.", "cohort": "round2/additions-1"}, {"company": "KazMunayGas", "existing_ids": [], "searched": true, "queries": ["site:kmg.kz AI predictive analytics refinery 2025", "site:anpz.kz AI predictive maintenance Honeywell refinery"], "documents_reviewed": [{"title": "КМГ внедряет предиктивную аналитику на НПЗ", "url": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "date": "2026-05-22", "type": "company announcement", "review_status": "full text reviewed"}, {"title": "AI control of equipment at Atyrau refinery: predictive analytics in action", "url": "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/", "date": "2025-12-25", "type": "subsidiary announcement", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Two refinery predictive-maintenance disclosures substantiate pilots and planned scale; no accuracy or financial result. Other KMG workflows such as ABAI are not included because the official holding-company article was inaccessible during this pass.", "cohort": "round2/additions-2"}, {"company": "QatarEnergy LNG", "existing_ids": [], "searched": true, "queries": ["site:qatarenergylng.qa AI machinery fleet predictive maintenance", "QatarEnergy LNG AI rotating equipment pilot"], "documents_reviewed": [{"title": "The Pioneer 167 — QatarEnergy LNG", "url": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "date": "2025-07", "type": "company magazine / strategy document", "review_status": "full text reviewed"}], "status": "partial", "gaps": "The full PDF is accessible; its relevant passage identifies pilot and expansion intent only. Pilot location, partner, model, rollout count and outcomes are not disclosed. QatarEnergy parent company had no additional named in-window AI operations disclosure confirmed in this targeted pass.", "cohort": "round2/additions-2"}, {"company": "Kuwait Petroleum Corporation / KIPIC / KNPC", "existing_ids": [], "searched": true, "queries": ["site:kpc.com.kw AI predictive maintenance seismic drilling 2025", "site:kipic.com.kw artificial intelligence refinery 2025", "site:knpc.com AI artificial intelligence 2025"], "documents_reviewed": [{"title": "KPC Sustainability Report 2023–25 (search excerpt only)", "url": "https://www.kpc.com.kw/AnnualReports/KPC-Sustainability-Report-2023-25.pdf", "date": "2025", "type": "sustainability report", "review_status": "excerpt reviewed"}], "status": "partial", "gaps": "Search excerpt attributed high-level AI for seismic interpretation, drilling optimization and predictive maintenance to KPC, but the 29 MB PDF failed full-text retrieval; no named deployment or measurable result added. KIPIC/KNPC-specific operational corroboration not found.", "cohort": "round2/additions-2"}, {"company": "Bapco Energies", "existing_ids": [], "searched": true, "queries": ["site:bapcoenergies.com 2025 AI refinery artificial intelligence", "site:bapcoenergies.com AI Bapco Modernization Program 2024"], "documents_reviewed": [{"title": "Annual Report 2024", "url": "https://www.bapcoenergies.com/media-centre/reports/download/annual-report-2024", "date": "2025", "type": "annual report", "review_status": "excerpt reviewed"}], "status": "partial", "gaps": "AI appears as an aim to optimize the Bapco Modernization Program, not a specific deployed workflow; excluded from operational adoption count.", "cohort": "round2/additions-2"}, {"company": "OQ Group / OQ Exploration & Production", "existing_ids": [], "searched": true, "queries": ["site:oq.com AI predictive maintenance 2025", "OQ Sustainability Report 2025 AI use cases"], "documents_reviewed": [{"title": "Digitalisation — OQ", "url": "https://oq.com/en/about-us/innovation/digitalisation", "date": "Undated; accessed 2026-09-27", "type": "company webpage", "review_status": "full text reviewed"}, {"title": "OQ Sustainability Report 2025", "url": "https://assets.oq.com/-/media/oq/files/sustainability/sustainability-at-oq/oq-sustainability-report--2025--english.pdf?rev=-1", "date": "2026", "type": "sustainability report", "review_status": "excerpt reviewed"}], "status": "partial", "gaps": "OQ webpage lists AI video analytics and recruiting capabilities but no dated in-window deployment. Search excerpt of 2025 report aggregates seven executed AI use cases without named workflows; no new dossier from an aggregate count.", "cohort": "round2/additions-2"}, {"company": "SOCAR", "existing_ids": [], "searched": true, "queries": ["site:socar.az artificial intelligence AI operations 2025", "site:socartech.az AI machine learning oil gas 2025"], "documents_reviewed": [{"title": "Data Management — SOCAR Tech", "url": "https://socartech.az/en/services/data-management", "date": "Undated; accessed 2026-09-27", "type": "company product page", "review_status": "full text reviewed"}], "status": "partial", "gaps": "SOCAR Tech current page describes ML product-quality prediction and refinery optimization capability but gives no launch/deployment date or named operating customer; not counted as in-window use.", "cohort": "round2/additions-2"}, {"company": "Sonatrach", "existing_ids": [], "searched": true, "queries": ["site:sonatrach.com intelligence artificielle 2025 exploitation", "site:sonatrach.com AI machine learning production refinery"], "documents_reviewed": [], "status": "no qualifying initiative found", "gaps": "Targeted public primary-source searches did not surface a named in-window operational AI workflow with deployment detail. Search limits apply; this is not evidence none exists.", "cohort": "round2/additions-2"}, {"company": "Pertamina subsidiaries", "existing_ids": [], "searched": true, "queries": ["site:pge.pertamina.com AI geothermal 2026 OR 2025", "site:phe.pertamina.com AI machine learning 2025 subsidiaries"], "documents_reviewed": [{"title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "date": "2025-09", "type": "company PDF", "review_status": "inaccessible"}, {"title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "date": "2025-10-27", "type": "subsidiary announcement", "review_status": "full text reviewed"}], "status": "partial", "gaps": "No additional named subsidiary case was found beyond assigned PINTAR and geothermal G-Bionic; PINTAR PDF could not be reopened and PGE results are geothermal/adjacent-energy, not oil-and-gas metrics.", "cohort": "round2/additions-2"}, {"company": "Other assigned entities: Aramco, PETRONAS, Ecopetrol, Petrobras, IndianOil, Reliance, SLB, Halliburton, NOV, Liberty, Cognite, Seeq, AVEVA, Snowflake, Enverus, Cenovus, Coterra, Woodside, OMV, PTTEP, Valero, Preem, Targa, ECP, Encino", "existing_ids": [], "searched": true, "queries": ["fresh company and partner-source queries across 2024–2026; see audit-2 source and record logs"], "documents_reviewed": [{"title": "Assignment-2 profile and packet source set", "url": "work/research-swarm/round2/assignment-2.json", "date": "2026-09-27", "type": "assigned source inventory", "review_status": "full text reviewed"}], "status": "reviewed", "gaps": "Existing-source re-review and record-level findings are captured in audit-2.json/.md; this row summarizes coverage only and makes no claim that all company disclosures were searched exhaustively.", "cohort": "round2/additions-2"}, {"company": "Ambyint", "existing_ids": [], "searched": true, "queries": ["site:ambyint.com AI methane plunger lift Tourmaline 2025", "site:eralberta.ca Ambyint 2025 methane AI", "Ambyint Encino Energy InfinityPL 2025 case study"], "documents_reviewed": [{"title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction", "url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "date": "2025-05-09", "type": "company announcement"}, {"title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector", "url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "date": "2025-04-16", "type": "public funding announcement"}, {"title": "Infinity PL Drives Efficiency Gains for Encino Energy", "url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "date": "2025-08-05", "type": "vendor case study"}], "status": "reviewed", "gaps": "Tourmaline rollout/up-to-90% methane reduction are prospective, not verified outcomes. Encino Energy already has exact baseline record AI-168; the source was reopened and packet replacement is supplied in audit-3.json rather than duplicated as a new addition.", "cohort": "round2/additions-3"}, {"company": "Kpler", "existing_ids": [], "searched": true, "queries": ["site:kpler.com AI chartering agent email parsing 2025", "Kpler AI Copilot energy trading 2025 official"], "documents_reviewed": [{"title": "Navigating Risk & Harnessing AI Briefing", "url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "date": "2025-10-16", "type": "company event announcement"}, {"title": "Kpler AI", "url": "https://www.kpler.com/solutions/ai", "date": null, "type": "product page"}], "status": "reviewed", "gaps": "The Oct 2025 event says AI chartering agent was launching Q4 2025; the present product page confirms AI features but supplies no dated launch record or adoption/outcome data.", "cohort": "round2/additions-3"}, {"company": "Peloton Computer Enterprises", "existing_ids": [], "searched": true, "queries": ["Peloton Computer Enterprises oil gas AI official platform 2025", "site:peloton.com oil gas AI"], "documents_reviewed": [{"title": "The Peloton Platform", "url": "https://www.peloton.com/peloton-platform-overview", "date": null, "type": "product page"}, {"title": "Carlyle Invests in Peloton to Drive Growth", "url": "https://www.peloton.com/peloton-in-the-news", "date": "2026-02-23", "type": "company news page"}], "status": "no qualifying initiative found", "gaps": "Oil-and-gas Peloton is a well/production/land data-management vendor. Carlyle describes AI as a forward growth driver, but no AI workflow or deployed customer case was identified. Fitness company Peloton Interactive is unrelated.", "cohort": "round2/additions-3"}, {"company": "W Energy Software", "existing_ids": [], "searched": true, "queries": ["site:wenergysoftware.com AI oil gas Stream+ 2025", "site:wenergysoftware.com machine learning oil gas software"], "documents_reviewed": [{"title": "How Stream+ Is Defining the Next Generation of Oil & Gas Software", "url": "https://wenergysoftware.com/resources/insights/introducing-streamplus-the-future-of-oil-gas-software/", "date": "2025-12-05", "type": "company product announcement"}], "status": "no qualifying initiative found", "gaps": "Reviewed materials describe integrated upstream/midstream accounting, land, production and field-service ERP/business intelligence. No specific AI method or AI-enabled operating workflow evidenced.", "cohort": "round2/additions-3"}, {"company": "IFS / Noble Corporation", "existing_ids": [], "searched": true, "queries": ["site:ifs.com oil gas AI Noble Corporation maintenance customer story", "site:ifs.com energy oil and gas AI customer case"], "documents_reviewed": [{"title": "Noble Corporation IFS.ai journey with IFS Cloud ERP", "url": "https://www.ifs.com/en/insights/customer-stories/noble-corporation", "date": null, "type": "vendor customer story"}], "status": "partial", "gaps": "IFS page presents predictive maintenance/FMECA benefits as expected upon a future IFS Cloud upgrade, not an implemented AI deployment. No replacement evidence of a qualifying deployed oil-and-gas customer workflow found in this pass.", "cohort": "round2/additions-3"}, {"company": "ComboCurve", "existing_ids": [], "searched": true, "queries": ["site:combocurve.com AI machine learning oil and gas software", "ComboCurve artificial intelligence official"], "documents_reviewed": [{"title": "About ComboCurve", "url": "https://combocurve.com/about-us/", "date": null, "type": "company page"}], "status": "no qualifying initiative found", "gaps": "Company describes technical forecasting and economic modeling software, with no named AI workflow or deployment-specific evidence surfaced.", "cohort": "round2/additions-3"}, {"company": "ChampionX", "existing_ids": [], "searched": true, "queries": ["site:championx.com digital solutions AI machine learning oil gas", "ChampionX AI machine learning SLB digital 2025"], "documents_reviewed": [{"title": "Midstream Data & Knowledge Management", "url": "https://www.championx.com/digital-solution/", "date": null, "type": "company product page"}, {"title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "date": "2026-01-23", "type": "earnings release"}], "status": "partial", "gaps": "ChampionX now operates within SLB (acquired July 2025). Public product pages describe AI/ML-backed analysis and digital optimization but omit named customer implementation and attributable results. Keep acquired scope distinct from standalone company.", "cohort": "round2/additions-3"}, {"company": "Verdane, Lime Rock Partners, EnCap, NGP Energy Capital, Warburg Pincus", "existing_ids": [], "searched": true, "queries": ["Verdane oil and gas energy software AI portfolio", "Lime Rock Partners AI oilfield software portfolio", "EnCap AI oil gas portfolio technology", "NGP energy portfolio AI software oil gas", "Warburg Pincus oil gas AI portfolio company"], "documents_reviewed": [{"title": "NGP Portfolio — Energy Transition", "url": "https://ngpenergycapital.com/portfolio-energy-transition/", "date": null, "type": "investor portfolio page"}, {"title": "Lime Rock Partners portfolio and approach", "url": "https://www.lrpartners.com/our-approach/", "date": null, "type": "investor page"}, {"title": "Warburg Pincus Energy Sector — Current Portfolio", "url": "https://warburgpincus.com/areas-of-focus/energy/current-portfolio/", "date": null, "type": "investor portfolio page"}, {"title": "EnCap Investments", "url": "https://encapinvestments.com/", "date": null, "type": "investor website"}, {"title": "Verdane", "url": "https://verdane.com/", "date": null, "type": "investor website"}], "status": "partial", "gaps": "No dated primary evidence found in this pass that these investors sponsored or operationally deployed a named oil-and-gas AI workflow. NGP portfolio does identify PhysicsX as AI-enabled engineering software and Orennia as energy-transition analytics; neither is evidence of oil-and-gas operator deployment. Warburg’s historic energy portfolio and Lime Rock/EnCap mandates alone do not establish AI adoption.", "cohort": "round2/additions-3"}, {"company": "IFS / Kodiak Gas Services", "existing_ids": [], "searched": true, "queries": ["site:ifs.com Kodiak Material Replenisher IFS Loops 2025", "site:ifs.com Kodiak field technicians AI agent"], "documents_reviewed": [{"title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI", "url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "date": "2025-11-27", "type": "company announcement"}, {"title": "IFS Loops launches Agent Studio", "url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "date": "2026-04-23", "type": "company announcement"}], "status": "reviewed", "gaps": "Deployment confirmed in dated IFS releases. The $3m ROI and 90,000-hour figures are conditional on half the technicians using the agent daily, not audited realized outcomes.", "cohort": "round2/additions-3"}, {"company": "Expand Energy", "existing_ids": [], "searched": true, "queries": ["Expand Energy AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "date": "2026-03-03", "type": "company announcement"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "SLB", "existing_ids": [], "searched": true, "queries": ["SLB AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "date": "2026-06-09", "type": "company announcement"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "Viridien", "existing_ids": [], "searched": true, "queries": ["Viridien AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "url": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "date": "2026-06-10", "type": "company announcement"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "TGS", "existing_ids": [], "searched": true, "queries": ["TGS AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "url": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "date": "2026-08-18", "type": "company announcement"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "Woodside", "existing_ids": [], "searched": true, "queries": ["Woodside AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "SLB Announces First-Quarter 2025 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "date": "2025-04-25", "type": "earnings release"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "Technip Energies", "existing_ids": [], "searched": true, "queries": ["Technip Energies AI 2025 2026 earnings investor operational workflows"], "documents_reviewed": [{"title": "Technip Energies shareholder information — 2025 chair and CEO messages", "url": "https://investors.technipenergies.com/shareholder-information", "date": null, "type": "annual report"}], "status": "partial", "gaps": "Targeted primary-source review; not a complete inventory of every reporting-period document.", "cohort": "round2/additions-root"}, {"company": "Vår Energi", "existing_ids": [], "searched": true, "queries": ["site:slb.com Vår Energi AI well planning2026"], "documents_reviewed": [{"title": "SLB and Vår Energi Expand Digital Collaboration to Scale Well and Integrated Field Development Planning", "url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0528-slb-var-energi", "date": "2026-05-28", "type": "company announcement"}], "status": "partial", "gaps": "The 2026 release establishes digital collaboration and planning-cycle claims, but not the AI contribution to the specific workflow. Earlier Neptune case describes AI but operating dates are historical. Not promoted into a fresh AI initiative count.", "cohort": "round2/additions-root"}, {"company": "Marathon Petroleum", "existing_ids": ["AI-203"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "AI-assisted pipeline right-of-way surveillance", "url": "https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/", "date": "2025-04-09", "type": "Operator interview / company news"}], "status": "partial", "gaps": "Company interview from a senior MPL field-services/planning executive; operational claims are company-reported and lack independent performance data or specific rollout coverage.", "cohort": "Gap review · 29 September 2026"}, {"company": "ORLEN", "existing_ids": ["AI-204"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "Group-wide AI rollout with market-monitoring and supply-chain assistants", "url": "https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment", "date": "2025-07-08", "type": "Operator press release"}], "status": "partial", "gaps": "Company announcement states scope and intent; no measured outcomes and no specific production asset use case attributed to the new AI assistants.", "cohort": "Gap review · 29 September 2026"}, {"company": "APA Corporation", "existing_ids": ["AI-205"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "AI use in emissions reporting, logistics and reservoir-management decisions", "url": "https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence-1.pdf", "date": "2024-11", "type": "Operator executive interview in industry supplement / original report"}], "status": "partial", "gaps": "Company interview in Egypt Oil & Gas industry supplement. It names AI functions (emissions reporting, supply-chain logistics, reservoir decisions) but gives no implementation example, product, asset by use case, maturity by use case, or measured result. These are high-level company-reported claims.", "cohort": "Gap review · 29 September 2026"}, {"company": "Murphy Oil", "existing_ids": ["AI-206"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "AI/ML-assisted continuous cyber monitoring and response", "url": "https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf", "date": "2025-08-06", "type": "Operator sustainability report"}], "status": "partial", "gaps": "The AI/ML use case is cybersecurity continuous monitoring/detection and response, not hydrocarbon production optimization. The report does not isolate model role, vendors, OT scope or performance. Separate GenAI Hub/Copilot deployment is reported on p. 79.", "cohort": "Gap review · 29 September 2026"}, {"company": "TotalEnergies", "existing_ids": ["AI-207"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "Experion Operations Assistant pilot at Port Arthur Delayed Coking Unit", "url": "https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial", "date": "2025-11-11", "type": "Operator-hosted joint pilot announcement"}], "status": "partial", "gaps": "Preliminary, company-reported; sample size is five events, with no false-alarm rate, missed-event rate, production or emissions quantity, counterfactual, deployment cost or payback. The source does not establish autonomous control authority.", "cohort": "Gap review · 29 September 2026"}, {"company": "Snam", "existing_ids": ["AI-208"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "AI extraction of gas-plant design data and equipment relationships", "url": "https://arxiv.org/html/2511.08609v1", "date": "2025-11-05", "type": "Operator/provider-authored technical preprint"}], "status": "partial", "gaps": "The abstract gives different aggregate accuracy figures. Complex diagrams remain a weakness. No independently validated performance, cash saving, measured time saving or fleet rollout is established; the 11,000-plant fleet is contextual.", "cohort": "Gap review · 29 September 2026"}, {"company": "PETRONAS", "existing_ids": ["AI-209"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "PROTEAN predictive rotating-equipment analytics", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "date": "2026 (PETRONAS Integrated Report 2025; precise publication date not established)", "type": "Operator integrated report"}], "status": "partial", "gaps": "The report establishes deployment breadth, not the number of sites, time in service, model performance, false alarms, realized avoided downtime or savings. It gives no implementation cost or payback and does not quantify the claimed outcome in the text.", "cohort": "Gap review · 29 September 2026"}, {"company": "Diversified Energy", "existing_ids": ["AI-210"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "OneDEC producer-statement automation and contract repository for revenue accounting and settlements", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "date": "2026-05-28", "type": "Operator-published sustainability report; issuer release confirms report date"}], "status": "partial", "gaps": "No quantified financial outcome or workflow-specific deployment denominator. Platform architecture does not prove every statement or contract step uses AI.", "cohort": "Gap review · 29 September 2026"}, {"company": "Trafigura", "existing_ids": ["AI-211"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "AI-enabled maritime voyage optimization across controlled fleet", "url": "https://www.trafigura.com/news-and-insights/press-releases/2025/trafigura-announces-strategic-alliance-with-maritime-technology-provider-zeronorth/", "date": "2025-07-09", "type": "Operator press release with named partner and attributed Global Head of Shipping"}], "status": "partial", "gaps": "Future-tense announcement. Fleet size is not a verified AI usage count, and expected savings are not operating results.", "cohort": "Gap review · 29 September 2026"}, {"company": "SM Energy", "existing_ids": ["AI-212"], "searched": true, "queries": ["Targeted missing-company / workflow follow-up"], "documents_reviewed": [{"title": "Machine-learning gas-lift injection-rate optimization in Midland Basin production operations", "url": "https://www.sec.gov/Archives/edgar/data/893538/000089353825000160/stephensannualinvestorco.htm", "date": "2025-11-21", "type": "Company conference transcript filed with SEC under Rule 425"}], "status": "partial", "gaps": "Company-authored conference transcript; no independently measured result. Recommendations do not establish autonomous valve control. Civitas extension was prospective.", "cohort": "Gap review · 29 September 2026"}, {"company": "Koch Energy Services", "existing_ids": ["AI-213"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "ipushpull launches LLM-enabled agent to automate physical gas trade booking via chat", "url": "https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat", "date": "2025-09-18", "type": "Vendor announcement with named customer quote"}, {"title": "Koch Energy Services and ipushpull automate physical gas trade booking through chat", "url": "https://ipushpull.com/case-studies/koch-energy-services-and-ipushpull-automate-physical-gas-trade-booking-through-chat", "date": "Undated", "type": "Vendor customer case study"}], "status": "partial", "gaps": "Platform trade count does not establish that every trade was AI-parsed. No baseline minutes or measurement method. Later undated 90,000-trade/85% figures are not treated as dated progress.", "cohort": "Second gap review · September 2026"}, {"company": "Sentinel Peak Resources", "existing_ids": ["AI-214"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Quantum Capital Group 2024 Stakeholder Report — Sentinel Peak case study", "url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "date": "2025-05-02 (file-indicated)", "type": "Sponsor report"}], "status": "partial", "gaps": "Results predate the 24-month disclosure window and include stewardship and analysis beyond AI. AI-only causality and baseline method are unproven. May 2025 upload is established by source path; exact May 2 date is file-indicated.", "cohort": "Second gap review · September 2026"}, {"company": "Occidental", "existing_ids": ["AI-215"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Occidental Q2 2025 Earnings Call Transcript", "url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy2q25transcript.pdf", "date": "2025-08-07", "type": "Issuer-hosted earnings-call transcript"}], "status": "partial", "gaps": "No route-level time, cost, failure-rate or production-uplift result. The contribution of AI cannot be separated from sensors and automation.", "cohort": "Second gap review · September 2026"}, {"company": "Occidental", "existing_ids": ["AI-153"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "2026 Oxy Proxy Statement", "url": "https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf", "date": "2026-03-19", "type": "Issuer proxy statement"}], "status": "partial", "gaps": "“Deployed” is not defined by asset count or rollout breadth. Incentive objectives refer to realized value, but the supporting results paragraph uses estimated benefits.", "cohort": "Second gap review · September 2026"}, {"company": "ADNOC Gas", "existing_ids": ["AI-216"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report"}], "status": "partial", "gaps": "No plant count, baseline, attribution method or value-component split. Future upside is excluded from reported delivered value.", "cohort": "Second gap review · September 2026"}, {"company": "ADNOC Gas", "existing_ids": ["AI-217"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report"}], "status": "partial", "gaps": "No quantified reduction in shutdowns or project economics. Neuron 5 appeared in an earlier unreviewed ADNOC portfolio disclosure; this is a dedicated, operator-attributed profile.", "cohort": "Second gap review · September 2026"}, {"company": "ADNOC Gas", "existing_ids": ["AI-218"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "ADNOC Gas Integrated Report 2025", "url": "https://adnocgas.ae/-/media/gas/investor-relation/2026/adnoc-gas_integrated-report-2025_final_english.ashx", "date": "2026", "type": "Issuer integrated report"}], "status": "partial", "gaps": "No detection accuracy, inspection-volume denominator, human sign-off process or measured safety effect.", "cohort": "Second gap review · September 2026"}, {"company": "Petrobras", "existing_ids": ["AI-219"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation"}], "status": "partial", "gaps": "No baseline, project duration, well count or value-calculation method. Separate from Petrobras flare control and monitoring.", "cohort": "Second gap review · September 2026"}, {"company": "Petrobras", "existing_ids": ["AI-220"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation"}], "status": "partial", "gaps": "No counterfactual, event count, baseline or implementation cost. Do not monetize without entitlement, pricing and realization evidence.", "cohort": "Second gap review · September 2026"}, {"company": "Petrobras", "existing_ids": ["AI-221"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation"}], "status": "partial", "gaps": "The 51% is not an absolute accuracy rate or a percentage-point change. No test population, horizon, baseline or validation method.", "cohort": "Second gap review · September 2026"}, {"company": "Petrobras", "existing_ids": ["AI-146"], "searched": true, "queries": ["Second targeted gap review"], "documents_reviewed": [{"title": "Petrobras Business Plan 2026–2030 investor presentation", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11", "type": "Issuer investor presentation"}], "status": "partial", "gaps": "Do not assign the full result specifically to Smart Tocha: the slide does not name it. No baseline, measurement period, plant count or emissions calculation method.", "cohort": "Second gap review · September 2026"}, {"company": "Enbridge", "existing_ids": ["AI-222"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization", "url": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "date": "2024-10-07", "type": "Operator press release"}], "status": "partial", "gaps": "No asset count, baseline or measured outcome. Enterprise cloud-migration statistics do not measure this tool’s deployment.", "cohort": "Third gap review · September 2026"}, {"company": "Enbridge", "existing_ids": ["AI-223"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "Enbridge collaborates with Microsoft to use AI for enhanced safety, emissions reduction and asset optimization", "url": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en", "date": "2024-10-07", "type": "Operator press release"}], "status": "partial", "gaps": "No accuracy, false-positive rate or response-time baseline. Separate from in-line inspection and deformation-feature assessment.", "cohort": "Third gap review · September 2026"}, {"company": "Williams", "existing_ids": ["AI-224"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "Williams tops 90% AI adoption as approved tools become part of everyday work", "url": "https://www.williams.com/2026/06/01/williams-tops-90-ai-adoption-as-approved-tools-become-part-of-everyday-work/", "date": "2026-06-01", "type": "Operator article"}], "status": "partial", "gaps": "Adoption is not a productivity measure. No department-level usage, frequency, workflow result or paid-license denominator.", "cohort": "Third gap review · September 2026"}, {"company": "ONGC", "existing_ids": ["AI-225"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "ONGC Integrated Annual Report 2024-25", "url": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "date": "2025-08-07", "type": "integrated annual report / statutory filing"}], "status": "partial", "gaps": "Implementation is described, but no quantified result or validation is reported. The pilot and dated South Heera start precede the window; only the August 2025 report disclosure is in-window. The source does not state well count, input data, decision rights, or model design.", "cohort": "Third gap review · September 2026"}, {"company": "ADNOC", "existing_ids": ["AI-226"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "ADNOC and SLB Deploy AI Platform Across Over 120 Drilling Rigs to Strengthen Upstream Performance", "url": "https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/", "date": "2026-08-04", "type": "Operator press release"}], "status": "partial", "gaps": "All performance figures are company-reported and lack measurement definitions, baselines, or independent validation. More than 120 rigs indicates platform deployment scope, not autonomous control. No model architecture, sensor list, or specific intervention examples are disclosed.", "cohort": "Third gap review · September 2026"}, {"company": "Devon Energy", "existing_ids": ["AI-227"], "searched": true, "queries": ["Third targeted gap review"], "documents_reviewed": [{"title": "Devon Energy Selects Harvey's Legal AI Platform", "url": "https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform", "date": "2025-08-14", "type": "Vendor announcement with named customer executive quote"}], "status": "partial", "gaps": "Vendor announcement with no task-level examples, rollout date, production evidence, controls or quantified results.", "cohort": "Third gap review · September 2026"}, {"company": "Moeve", "existing_ids": ["AI-228"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report"}], "status": "partial", "gaps": "No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.", "cohort": "Fourth gap review · October 2026"}, {"company": "Moeve", "existing_ids": ["AI-229"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report"}], "status": "partial", "gaps": "No validation baseline or autonomous-control evidence. Adjacent energy-recovery savings are not attributed to these AI models.", "cohort": "Fourth gap review · October 2026"}, {"company": "Moeve", "existing_ids": ["AI-230"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Moeve and Mistral AI join forces to spearhead generative AI solutions in the energy sector", "url": "https://www.moeveglobal.com/en/press/moeve-mistral-ai-alliance-ai-digital-assistant-industrial-efficiency", "date": "2025-12-04", "type": "Operator press release"}, {"title": "Moeve Consolidated Management Report 2025", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf", "date": "2026-02-26", "type": "Operator annual management report"}], "status": "partial", "gaps": "Announcement is prospective. Neither rollout across all sites nor energy/cost savings is evidenced.", "cohort": "Fourth gap review · October 2026"}, {"company": "Presidio Production", "existing_ids": ["AI-231"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Presidio Production Company Investor Presentation, dated June 2026", "url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026069875/ea029480701_ex99-1.htm", "date": "2026-06", "type": "issuer investor presentation filed with SEC"}], "status": "partial", "gaps": "Live surveillance is disclosed; fleet-wide orchestration remains a target. Use the explicit target/progress framing on slide 15. The column placement of the flattened slide 16 bullet was not independently verified. Visit reductions have a different scope from production uplift and must not be added to it. The $2.5 million sensitivity is illustrative. Presidio listed in March 2026; historical sponsor backing does not make it currently private.", "cohort": "Fourth gap review · October 2026"}, {"company": "SLB", "existing_ids": ["AI-232"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study with named SLB architect"}], "status": "partial", "gaps": "The August 8, 2025 date is the Microsoft Learn page's explicit last-updated date, not the deployment date. Vendor-published case; customer executive quote and detailed workflow are disclosed, but the source does not give accuracy, exception rate, baseline minutes, governance details for ambiguous records, or audited savings. Invoice scope is airline travel invoices and tax processing, not supplier AP/JIB/revenue accounting.", "cohort": "Fourth gap review · October 2026"}, {"company": "SLB", "existing_ids": ["AI-233"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study"}], "status": "partial", "gaps": "Source is a Microsoft-hosted customer story; no publication date beyond last-updated 2025-08-08, no deployment launch date, precision/recall, project-value attribution, costs, or independent verification. This is an internal project-discovery workflow, not a field production deployment.", "cohort": "Fourth gap review · October 2026"}, {"company": "SLB", "existing_ids": ["AI-234"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "SLB enhances productivity with Power Platform and AI", "url": "https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/slb-enhances-productivity", "date": "2025-08-08", "type": "Microsoft Learn customer case study"}], "status": "partial", "gaps": "Microsoft-hosted customer story, not independently validated. The $10 million figure has no stated period, reconciliation, or counterfactual. The exact publication date is not explicit beyond page last-updated 2025-08-08; no launch date or throughput/accuracy is disclosed. This is a distinct shipping-document/export-duty workflow, not the airline invoice GST bot also described on the same page.", "cohort": "Fourth gap review · October 2026"}, {"company": "ExxonMobil", "existing_ids": ["AI-235"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Gas Lift Optimization in the Permian Uses Machine Learning, Artificial Intelligence", "url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "date": "2025-10-01", "type": "SPE Journal of Petroleum Technology paper highlight"}, {"title": "Gas Lift Optimization in the Permian Using Machine Learning and Artificial Intelligence (SPE 219553-MS)", "url": "https://www.researchgate.net/publication/383230634_Gas_Lift_Optimization_in_the_Permian_Using_Machine_Learning_and_Artificial_Intelligence", "date": "2024-08", "type": "Operator-authored SPE conference paper abstract"}, {"title": "Unsung Hero: Artificial Lift Boosts Production Without Breaking the Bank", "url": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted", "date": "2024-10-01", "type": "SPE Journal of Petroleum Technology feature with operator-engineer interview"}], "status": "partial", "gaps": "The primary paper's accessible abstract states the full paper is unavailable, so full technical validation is not possible. The operator-reported uplift lacks a disclosed counterfactual or confidence interval. Gas-lift performance can drift when water cut or productivity changes; JPT says an existing setting may remain suboptimal for years without frequent monitoring. The control-boundary details and automatic-versus-human exception rates are undisclosed.", "cohort": "Fourth gap review · October 2026"}, {"company": "PetroChina", "existing_ids": ["AI-236"], "searched": true, "queries": ["Fourth targeted gap review"], "documents_reviewed": [{"title": "Machine-learning production forecasting across Dagang Oilfield blocks", "url": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml", "date": "2026-02-04", "type": "CNPC News Center / China Petroleum News Center operator-group news article"}], "status": "partial", "gaps": "The article says Dagang created 11 AI applications in 2025; this is a portfolio count, not 11 separately evidenced profiles. The 15% accuracy statement is an operator-group-reported model-performance claim without a disclosed denominator or validation method. Adjacent figures in the article—response speed up more than 60% and plan-preparation time down an average 30%—describe broader digital transformation/platform work and are not attributed to this ML forecast model.", "cohort": "Fourth gap review · October 2026"}, {"company": "Repsol", "existing_ids": ["AI-237"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "Harvey Power Users: “People Call Me Harvey”", "url": "https://www.harvey.ai/blog/harvey-power-users-people-call-me-harvey", "date": "2026-03-30", "type": "Harvey customer interview with Repsol Legal Counsel"}, {"title": "Repsol becomes the first Spanish in-house legal team to use Harvey", "url": "https://www.repsol.com/en/press-room/press-releases/2024/repsol-becomes-the-first-spanish-in-house-legal-team-to-use-harvey/index.cshtml", "date": "2024-04-04 (pre-window context only)", "type": "Repsol issuer press release"}], "status": "partial", "gaps": "The new dated source is a Harvey vendor interview, not a Repsol-issued performance report. It supports rollout across 150 people but not a completed adoption percentage or validated performance result. An undated Harvey Repsol customer page makes stronger claims (“96%” adoption; “4–6 hours” saved per lawyer/week, “two years later”) but has no verifiable publication date; treat those claims as an undated supplemental lead, not a dated in-window outcome.", "cohort": "Fifth gap review · October 2026"}, {"company": "Spearpoint Resources", "existing_ids": ["AI-238"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "Three entities, 27 purchaser formats, and 20 to 30 hours a week back", "url": "https://www.joltly.io/blog/spear-point-revenue-accounting-case-study", "date": "2026-09-05", "type": "Joltly customer case study quoting Spearpoint controller"}], "status": "partial", "gaps": "Vendor-authored case study; no audited results, accuracy/error rate, control testing, implementation cost, customer-specific pricing, or independent attribution. The output is reviewed/exported by people; do not describe it as fully autonomous posting. It is a named but small non-operated asset owner described without basin or asset-level details.", "cohort": "Fifth gap review · October 2026"}, {"company": "INPEX", "existing_ids": ["AI-239"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "INPEX Integrated Report 2024", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf", "date": "2025-07-04", "type": "Operator integrated report"}, {"title": "INPEX Integrated Report 2025 — Growth Strategy section", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "date": "2026-07-07", "type": "Operator integrated report (separate Growth Strategy PDF)"}, {"title": "INPEX Integrated Report index — publication date confirmation", "url": "https://www.inpex.com/english/ir/library/integrated_report/", "date": "Undated live index; accessed 2026-10-01", "type": "Issuer report index"}], "status": "partial", "gaps": "This is an engineering decision-support model, not evidence of AI control of an operating CCS asset. No specific field, injection project, go-live date, user role, validation method, accuracy, quantified time reduction or economic result is reported.", "cohort": "Fifth gap review · October 2026"}, {"company": "INPEX", "existing_ids": ["AI-240"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "INPEX Integrated Report 2025 — Growth Strategy section", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "date": "2026-07-07", "type": "Operator integrated report (separate Growth Strategy PDF)"}, {"title": "INPEX Integrated Report index — publication date confirmation", "url": "https://www.inpex.com/english/ir/library/integrated_report/", "date": "Undated live index; accessed 2026-10-01", "type": "Issuer report index"}], "status": "partial", "gaps": "Operator-reported implementation and process metrics. The reported >90% concerns classification accuracy, not overall answer accuracy. Faster information retrieval and record structuring are not quantified as downtime or production improvements. Engineer approval of search quality does not establish mandatory human review of every response or maintenance action.", "cohort": "Fifth gap review · October 2026"}, {"company": "Saipem", "existing_ids": ["AI-241"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "Saipem publishes 2025 sustainability results", "url": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "date": "2026-05-20", "type": "Issuer disclosure"}, {"title": "Saipem Smart Cameras privacy notice", "url": "https://www.saipem.com/sites/default/files/2026-01/75_Privacy_Notice_Smart_Cameras.pdf", "date": "2026-01-28", "type": "Issuer operational notice"}, {"title": "Saipem MISR Smart Cameras notice — Saipem 10000", "url": "https://www.saipem.com/sites/default/files/2025-04/Information%20notice%20Smart%20Cameras_Saipem%20MISR%20for%20Petroleum%20Services%20%28S.A.E.%29_0.pdf", "date": "2025-03-31", "type": "Issuer operational notice"}, {"title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf", "date": "2026", "type": "Issuer sustainability report"}, {"title": "Saipem Consolidated Sustainability Statement 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf", "date": "2026", "type": "Issuer annual-report extract"}], "status": "partial", "gaps": "Notices cover different entities/dates. The March 2025 vessel notice has inconsistent subsidiary naming; use it for the named vessel and stated controls only. No deployment-wide accuracy or causal safety result.", "cohort": "Fifth gap review · October 2026"}, {"company": "Saipem", "existing_ids": ["AI-242"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "Saipem publishes 2025 sustainability results", "url": "https://www.saipem.com/en/media/news/2025-sustainability-results-published", "date": "2026-05-20", "type": "Issuer disclosure"}, {"title": "Saipem Sustainability Overview 2025", "url": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf", "date": "2026", "type": "Issuer sustainability report"}, {"title": "Saipem Consolidated Sustainability Statement 2025", "url": "https://www.saipem.com/sites/default/files/2026-04/2025_Consolidated_Sustainability_Statement_0.pdf", "date": "2026", "type": "Issuer annual-report extract"}], "status": "partial", "gaps": "The readiness graphic is not used as an outcome or deployment percentage. Asset count and operator approval procedures remain unspecified.", "cohort": "Fifth gap review · October 2026"}, {"company": "MODEC", "existing_ids": ["AI-243"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "MODEC, Shell Brasil, Shape Digital and Unicamp safety-barrier AI project", "url": "https://www.modec.com/news/2025/20250813_pr_shape-digital.html", "date": "2025-08-13", "type": "Issuer project announcement"}], "status": "partial", "gaps": "Existing Shape Reef use in Brazil, Mexico and Africa does not establish deployment of the new Phase II models.", "cohort": "Fifth gap review · October 2026"}, {"company": "SBM Offshore", "existing_ids": ["AI-244"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "SBM Offshore Q3 2025 trading update", "url": "https://www.sbmoffshore.com/newsroom/sbm-offshore-third-quarter-2025-trading-update/", "date": "2025-11-13", "type": "Issuer disclosure"}, {"title": "SBM Offshore and Cognite accelerate AI and data collaboration", "url": "https://www.sbmoffshore.com/newsroom/sbm-offshore-and-cognite-partner-to-accelerate-ai-and-data-collaboration-to-optimize-floating-production-operations/", "date": "2025-10-02", "type": "Issuer disclosure"}, {"title": "SLB and SBM Offshore digital alliance", "url": "https://www.sbmoffshore.com/newsroom/slb-and-sbm-offshore-agree-digital-alliance-to-transform-fpso-production-performance/", "date": "2025-10-02", "type": "Issuer disclosure"}], "status": "partial", "gaps": "Fleet expansion is not proof that every asset or AI use case is live. No vessel-by-vessel rollout, acceptance tests or quantified benefit.", "cohort": "Fifth gap review · October 2026"}, {"company": "Sentinel Peak Resources", "existing_ids": ["AI-214"], "searched": true, "queries": ["Fifth targeted gap review"], "documents_reviewed": [{"title": "Environment — Sentinel Peak Resources", "url": "https://sentinelpeakresources.com/sustainability/environment/", "date": "undated; accessed 2026-10-01", "type": "Operator environment disclosure"}], "status": "partial", "gaps": "Do not add this estimate to the baseline’s historical 2023 figures or treat either as AI-only net savings.", "cohort": "Fifth gap review · October 2026"}, {"company": "PTTEP", "existing_ids": ["AI-245"], "searched": true, "queries": ["Sixth targeted gap review"], "documents_reviewed": [{"title": "PTTEP 56-1 One Report 2025 — internal audit", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "2026; referenced in invitation dated 27 February 2026", "type": "Operator annual report"}, {"title": "PTTEP 2026 AGM invitation", "url": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-invitation-en.pdf", "date": "2026-02-27", "type": "Issuer meeting invitation"}], "status": "partial", "gaps": "No model, deployment scale, accuracy, approval controls or measured productivity. Do not relabel the separately described audit software as AI.", "cohort": "Sixth gap review · October 2026"}, {"company": "MOL Group", "existing_ids": ["AI-246"], "searched": true, "queries": ["Sixth targeted gap review"], "documents_reviewed": [{"title": "MOL 2025 AGM documents — Consumer Services", "url": "https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf", "date": "2025 AGM material; meeting 24 April 2025", "type": "Issuer AGM management report"}], "status": "partial", "gaps": "Eight-market program reach is not eight-market AI deployment. No test population, conversion uplift, cost or model controls disclosed.", "cohort": "Sixth gap review · October 2026"}, {"company": "Ecopetrol", "existing_ids": ["AI-247"], "searched": true, "queries": ["Sixth targeted gap review"], "documents_reviewed": [{"title": "Ecopetrol 2025 Integrated Management Report — EVO talent map", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "type": "Operator annual report"}], "status": "partial", "gaps": "Do not attach adjacent SAP onboarding or Signavio results to EVO. Bias testing, privacy controls and human override are unspecified.", "cohort": "Sixth gap review · October 2026"}, {"company": "Ecopetrol", "existing_ids": ["AI-147"], "searched": true, "queries": ["Sixth targeted gap review"], "documents_reviewed": [{"title": "Ecopetrol 2025 Integrated Management Report — AI KPIs", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "type": "Operator annual report"}], "status": "partial", "gaps": "The 7% uses five-year consolidated, EBITDA-aligned figures with lagged benefits. The investment measure excludes other AI spending without direct financial impact. No project-level payback or independently verified causal return.", "cohort": "Sixth gap review · October 2026"}, {"company": "Ecopetrol", "existing_ids": ["AI-148"], "searched": true, "queries": ["Sixth targeted gap review"], "documents_reviewed": [{"title": "Ecopetrol 2025 Integrated Management Report — CADE", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 AGM report; covers 2025", "type": "Operator annual report"}], "status": "partial", "gaps": "An operating center does not establish deployed autonomous dispatch. No final model acceptance, production rollout, control bounds, cost saving or payback disclosed.", "cohort": "Sixth gap review · October 2026"}, {"company": "Continental Resources", "existing_ids": ["AI-248"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Continental Resources Streamlines Operations with Automation: From fragmented emails to orchestrated operations at Continental Resources", "url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "date": "No displayed publication/update date; accessed 2026-10-01", "type": "Vendor customer case"}, {"title": "Honoring the 2025 UiPath AI25 Winners", "url": "https://www.uipath.com/ai/ai25-awards/winners", "date": "Undated page; references 2025 award cohort; accessed 2026-10-01", "type": "Vendor customer case"}, {"title": "UiPath Announces 2025 AI25 Awards Winners", "url": "https://www.uipath.com/newsroom/uipath-announces-2025-ai25-awards-winners", "date": "2025-09-29", "type": "Vendor customer case"}, {"title": "UiPath + Databricks: Data Intelligence Meets Agentic Automation", "url": "https://www.uipath.com/platform/agentic-automation/ai-ecosystem/databricks", "date": "Undated page; accessed 2026-10-01", "type": "Vendor partner page"}], "status": "partial", "gaps": "2025-09-29 dates the award announcement only. Current workflow and metric dates are unknown; 40+ processes are not 40 AI initiatives. No test-set definition, error monitoring or independently verified cash savings.", "cohort": "Seventh gap review · October 2026"}, {"company": "Chord Energy", "existing_ids": ["AI-097"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Chord Energy FY2025 results release", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "date": "2026-02-25", "type": "issuer earnings release"}, {"title": "Chord Energy 1Q26 Earnings Presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2026/05/CHRD-Q1-2026-Earnings-Presentation.pdf", "date": "2026-05-05", "type": "issuer investor presentation"}, {"title": "Chord Energy 2Q26 results release", "url": "https://ir.chordenergy.com/2026-08-05-Chord-Energy-Reports-Second-Quarter-2026-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Updates-2026-Outlook", "date": "2026-08-05", "type": "issuer earnings release"}, {"title": "Chord Energy 2Q25 Investor Presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2025/08/CHRD-2Q25-Earnings-Presentation.pdf", "date": "2025-08-06", "type": "issuer investor presentation"}], "status": "partial", "gaps": "Company-reported early result without a disclosed comparison method. Broader workover, ESP and savings results cannot be isolated to AI.", "cohort": "Seventh gap review · October 2026"}, {"company": "Devon Energy", "existing_ids": ["AI-006"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Devon Energy Q1 2026 Earnings Call Corrected Transcript", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-05-06", "type": "Issuer disclosure"}], "status": "partial", "gaps": "Company-level CEO assertion; no baseline, period, control group, production uplift, uptime, savings, lift mix, well list or attribution. Do not allocate Devon's $1bn business-optimization target/achievement to AI.", "cohort": "Seventh gap review · October 2026"}, {"company": "Devon Energy", "existing_ids": ["AI-249"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Devon Energy 2025 Sustainability Report", "url": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "date": "2025 (report; covers calendar 2024 unless noted)", "type": "Issuer disclosure"}], "status": "partial", "gaps": "Deployment scope, model, alert precision/recall, response time and impact are undisclosed. Avoid conflating the continuous-monitoring work-order automation with AI decision support.", "cohort": "Seventh gap review · October 2026"}, {"company": "Devon Energy", "existing_ids": ["AI-250"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Devon Energy 2025 Sustainability Report", "url": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "date": "2025 (report; covers calendar 2024 unless noted)", "type": "Issuer disclosure"}], "status": "partial", "gaps": "No model accuracy, false-alert rate, equipment coverage, avoided emissions, uptime or financial attribution. General statement about capabilities, not proof of quantified impact.", "cohort": "Seventh gap review · October 2026"}, {"company": "Diamondback Energy", "existing_ids": ["AI-251"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Diamondback Energy 2025 Corporate Sustainability Report", "url": "https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418", "date": "2025 (2024 performance reporting)", "type": "Issuer disclosure"}], "status": "partial", "gaps": "Vendor product description can establish generic mechanism only; no Diamondback-specific performance or exact decision threshold. Do not infer the model's accuracy or causal safety benefit.", "cohort": "Seventh gap review · October 2026"}, {"company": "Diamondback Energy", "existing_ids": ["AI-045"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Diamondback Energy Letter to Stockholders (Q3 2025)", "url": "https://ir.diamondbackenergy.com/node/17156/pdf", "date": "2025-11-03", "type": "Issuer disclosure"}, {"title": "Diamondback Energy Letter to Stockholders (Q1 2026)", "url": "https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11", "date": "2026-05-04", "type": "Issuer disclosure"}], "status": "partial", "gaps": "Qualitative issuer disclosure only; implementation level and actual performance unquantified. May 2026 release explicitly frames uptime/LOE as expected in coming years.", "cohort": "Seventh gap review · October 2026"}, {"company": "Encino Energy", "existing_ids": ["AI-166"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Encino Energy streamlines operations with Microsoft Power Platform", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "date": "2025-04-29 (visible on page)", "type": "Vendor customer case"}], "status": "partial", "gaps": "Vendor/customer account. Maintenance-app savings are ordinary workflow automation, not AI-attributable. Snowflake supplies data; this is not proof of Snowflake-native inference.", "cohort": "Seventh gap review · October 2026"}, {"company": "Expand Energy", "existing_ids": ["AI-195"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "date": "No publication/update date displayed; accessed 2026-10-01", "type": "Vendor customer case"}], "status": "partial", "gaps": "Supplemental case has no visible date. Its metrics cannot inherit the existing profile’s March 2026 disclosure date; model accuracy and AI-only financial return undisclosed.", "cohort": "Seventh gap review · October 2026"}, {"company": "ConocoPhillips", "existing_ids": ["AI-202"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "Case Study: Field Deployments of AI-Based IOCaaS Advancing Artificial Lift and Flow Assurance", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "date": "2026-04-01", "type": "Society of Petroleum Engineers / Journal of Petroleum Technology article; vendor-contributed case study, coauthored by OPX Ai executives."}], "status": "partial", "gaps": "Forecast construction, well denominator, comparison group and evaluation dates undisclosed. Supplier executives authored the case; no independent validation. Monitoring does not establish closed-loop control.", "cohort": "Seventh gap review · October 2026"}, {"company": "ExxonMobil", "existing_ids": ["AI-072"], "searched": true, "queries": ["Seventh targeted gap review"], "documents_reviewed": [{"title": "ExxonMobil Guyana Limited leverages drilling automation to set new performance benchmarks in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "date": "2026 (URL path); no exact date displayed in page text", "type": "SLB vendor case study"}], "status": "partial", "gaps": "2026 URL path, no exact publication or campaign date. Vendor-reported results; 93% is the cited operating scope, not all Exxon drilling. No sample or counterfactual. Separate Halliburton campaign metrics must not be merged.", "cohort": "Seventh gap review · October 2026"}, {"company": "Vista Energy", "existing_ids": ["AI-298"], "searched": true, "queries": ["Bounded watch; exact searches retained in research/watch-2026-10-22-pm agent ledgers"], "documents_reviewed": [{"title": "Vista Investor Day 2025", "url": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "date": "2025-11-12", "type": "Company investor presentation"}, {"title": "Vista Investor Day 2025 — SEC exhibit", "url": "https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm", "date": "2025-11-12", "type": "Company investor presentation"}], "status": "partial", "gaps": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted.", "cohort": "22 October evening source review", "reviewed_on": "2026-10-22"}, {"company": "Saipem", "existing_ids": ["AI-299"], "searched": true, "queries": ["Bounded watch; exact searches retained in research/watch-2026-10-22-pm agent ledgers"], "documents_reviewed": [{"title": "Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000", "url": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "date": "2026-02-23", "type": "Company announcement"}], "status": "partial", "gaps": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed.", "cohort": "22 October evening source review", "reviewed_on": "2026-10-22"}, {"company": "Worley", "existing_ids": ["AI-300"], "searched": true, "queries": ["Bounded watch; exact searches retained in research/watch-2026-10-22-pm agent ledgers"], "documents_reviewed": [{"title": "FY2025 full-year results presentation", "url": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "date": "2025-08-27", "type": "Company investor presentation"}], "status": "partial", "gaps": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established.", "cohort": "22 October evening source review", "reviewed_on": "2026-10-22"}, {"company": "Whitecap Resources", "existing_ids": [], "searched": true, "queries": ["Bounded watch; exact searches retained in research/watch-2026-10-22-pm agent ledgers"], "documents_reviewed": [{"title": "Whitecap Resources customer story", "url": "https://www.shakudo.io/customers/whitecap-resources", "date": "2026-04-23", "type": "Vendor customer story"}, {"title": "Shakudo customer stories — Whitecap entry", "url": "https://www.shakudo.io/customers", "date": "2026-04-23", "type": "Vendor customer story"}], "status": "partial", "gaps": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved.", "cohort": "22 October evening source review", "reviewed_on": "2026-10-22"}, {"company": "Technip Energies", "existing_ids": ["AI-199"], "searched": true, "queries": ["Bounded watch; exact searches retained in research/watch-2026-10-22-pm agent ledgers"], "documents_reviewed": [{"title": "Technip Energies 2025 Annual Report", "url": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "date": "2026 (exact publication day unverified)", "type": "Indexed original annual-report excerpt"}], "status": "partial", "gaps": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved.", "cohort": "22 October evening source review", "reviewed_on": "2026-10-22"}, {"company": "Azule Energy", "existing_ids": ["AI-301"], "searched": true, "queries": ["Fresh SLB Q3 release; exact editor and Luna queries in dated ledger"], "documents_reviewed": [{"title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "type": "Supplier earnings release / customer account"}, {"title": "Azule Energy pioneers digital well planning and rig automation in Africa", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa", "date": "Unknown", "type": "Supplier earnings release / customer account"}], "status": "partial", "gaps": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate.", "cohort": "23 October morning source review", "reviewed_on": "2026-10-23"}, {"company": "SLB / unnamed Permian operator", "existing_ids": [], "searched": true, "queries": ["Fresh SLB Q3 release; exact editor and Luna queries in dated ledger"], "documents_reviewed": [{"title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "type": "Supplier earnings release / customer account"}], "status": "partial", "gaps": "Unnamed operator: no company attribution, core adoption or result-count addition. No controls, cost reconciliation, model error rate or approval/override policy. Do not infer a specific Permian producer or treat the figure as audited net cash ROI.", "cohort": "23 October morning source review", "reviewed_on": "2026-10-23"}, {"company": "Eni", "existing_ids": ["AI-027"], "searched": true, "queries": ["Fresh SLB Q3 release; exact editor and Luna queries in dated ledger"], "documents_reviewed": [{"title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "type": "Supplier earnings release / customer account"}], "status": "partial", "gaps": "Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications.", "cohort": "23 October morning source review", "reviewed_on": "2026-10-23"}, {"company": "Brunei Shell Petroleum", "existing_ids": ["AI-302"], "searched": true, "queries": ["Fresh SLB Q3 release; exact editor and Luna queries in dated ledger"], "documents_reviewed": [{"title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "type": "Supplier earnings release / customer account"}], "status": "partial", "gaps": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout.", "cohort": "23 October morning source review", "reviewed_on": "2026-10-23"}, {"company": "ADNOC", "existing_ids": ["AI-226"], "searched": true, "queries": ["Fresh SLB Q3 release; exact editor and Luna queries in dated ledger"], "documents_reviewed": [{"title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "type": "Supplier earnings release / customer account"}], "status": "partial", "gaps": "Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet.", "cohort": "23 October morning source review", "reviewed_on": "2026-10-23"}, {"company": "Petrobras", "existing_ids": ["AI-303"], "searched": true, "queries": ["Exact queries and exclusions in research/watch-2026-10-23-pm/global-limits.json"], "documents_reviewed": [{"title": "Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice", "url": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "date": "Unknown", "type": "Company supplier-channel pilot notice"}], "status": "partial", "gaps": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.", "cohort": "23 October evening source review", "reviewed_on": "2026-10-23"}, {"company": "Halliburton", "existing_ids": ["AI-304"], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "How predictive intelligence reduces downtime and improves equipment reliability", "url": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "date": "2026-10-22", "type": "Supplier product account"}, {"title": "LOGIX Unit Vitality — cement equipment health monitoring", "url": "https://www.halliburton.com/en/products/logix-unit-vitality", "date": "Unknown", "type": "Supplier product account"}], "status": "partial", "gaps": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "cohort": "24 October morning source review", "reviewed_on": "2026-10-24"}, {"company": "ONGC", "existing_ids": [], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "ONGC Technology — Project ICE", "url": "https://www.ongcindia.com/web/eng/about-ongc/technology", "date": "Unknown", "type": "Company technology overview"}], "status": "partial", "gaps": "Historical non-AI context only. AI-enabled innovation is a roadmap aim; no specific live AI workflow or decision authority established. A shared platform does not by itself prove clean data, adoption quality or AI readiness.", "cohort": "24 October morning source review", "reviewed_on": "2026-10-24"}, {"company": "Petroleum Development Oman", "existing_ids": ["AI-305"], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "date": "2026-04-21", "type": "Supplier earnings release"}], "status": "partial", "gaps": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.", "cohort": "24 October evening source review", "reviewed_on": "2026-10-24"}, {"company": "Weatherford", "existing_ids": [], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "Weatherford Announces Third Quarter 2026 Results — rod-lift energy recovery", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18871", "date": "2026-10-20", "type": "Supplier earnings release"}], "status": "partial", "gaps": "Anonymous single-installation supplier claim. Do not treat the percentage as company-wide energy savings, production uplift, net cash ROI or an AI result. The word ForeSite links a brand family, not proof this equipment used the PDO predictive model.", "cohort": "24 October evening source review", "reviewed_on": "2026-10-24"}, {"company": "Nabors", "existing_ids": ["AI-115"], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated services/private-platform/adverse ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No new qualifying deployment or adverse outcome retained in this bounded pass. Hilcorp recap inaccessible and AWS context undated; no-hit is not no activity.", "cohort": "24 October evening bounded research", "reviewed_on": "2026-10-24"}, {"company": "Patterson-UTI", "existing_ids": ["AI-117"], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated services/private-platform/adverse ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No new qualifying deployment or adverse outcome retained in this bounded pass. Hilcorp recap inaccessible and AWS context undated; no-hit is not no activity.", "cohort": "24 October evening bounded research", "reviewed_on": "2026-10-24"}, {"company": "Hilcorp Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated services/private-platform/adverse ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No new qualifying deployment or adverse outcome retained in this bounded pass. Hilcorp recap inaccessible and AWS context undated; no-hit is not no activity.", "cohort": "24 October evening bounded research", "reviewed_on": "2026-10-24"}, {"company": "Aethon Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated services/private-platform/adverse ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No new qualifying deployment or adverse outcome retained in this bounded pass. Hilcorp recap inaccessible and AWS context undated; no-hit is not no activity.", "cohort": "24 October evening bounded research", "reviewed_on": "2026-10-24"}, {"company": "Caerus Oil and Gas", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated services/private-platform/adverse ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No new qualifying deployment or adverse outcome retained in this bounded pass. Hilcorp recap inaccessible and AWS context undated; no-hit is not no activity.", "cohort": "24 October evening bounded research", "reviewed_on": "2026-10-24"}, {"company": "PTTEP", "existing_ids": ["AI-306"], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "Unknown", "type": "Operator annual report"}, {"title": "PTTEP organizes OpEx Day 2025, advancing operational excellence", "url": "https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence", "date": "2025-07-09", "type": "Operator press release"}], "status": "partial", "gaps": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "cohort": "25 October morning source review", "reviewed_on": "2026-10-25"}, {"company": "BKV Corporation", "existing_ids": [], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "BKV 2025 Sustainability Report — Project Runway", "url": "https://www.bkv.com/2025-sustainability-report-governance", "date": "Unknown", "type": "Operator annual report"}, {"title": "BKV 2025 Sustainability Report — letter date", "url": "https://www.bkv.com/2025-sustainability-report", "date": "Unknown", "type": "Operator report landing page"}], "status": "partial", "gaps": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date.", "cohort": "25 October morning source review", "reviewed_on": "2026-10-25"}, {"company": "Harbour Energy", "existing_ids": [], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "Harbour Energy Coupa Supplier Portal FAQ, version 4", "url": "https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf", "date": "2025-06", "type": "Operator supplier guide"}, {"title": "Harbour vendor relations — distinct supplier channels", "url": "https://www.harbourenergy.com/vendor-relations/", "date": "Unknown", "type": "Operator supplier webpage"}], "status": "partial", "gaps": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates.", "cohort": "25 October morning source review", "reviewed_on": "2026-10-25"}, {"company": "ConocoPhillips", "existing_ids": [], "searched": true, "queries": ["Exact editor and Luna queries in dated research ledger"], "documents_reviewed": [{"title": "ConocoPhillips GEP SMART Introduction", "url": "https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf", "date": "2024-12", "type": "Operator supplier guide"}, {"title": "ConocoPhillips technical upgrades to supplier systems", "url": "https://vendors.conocophillips.com/supplier-guide/technical-upgrades-to-supplier-systems/", "date": "Unknown", "type": "Operator supplier webpage"}], "status": "partial", "gaps": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption.", "cohort": "25 October morning source review", "reviewed_on": "2026-10-25"}, {"company": "ARC Resources", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Peyto", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Vermilion Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Tourmaline Oil", "existing_ids": ["AI-257"], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Kimmeridge", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Aethon Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Rockcliff Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Ascent Resources", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "Gulfport Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause assignments and access limits in dated canada/private/admin-platforms ledgers"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow retained. Tourmaline vintage curves do not establish a new software workflow; ARC job-card detail is inconsistent; no-hit is not no activity.", "cohort": "25 October morning bounded research", "reviewed_on": "2026-10-25"}, {"company": "YPF", "existing_ids": ["AI-149"], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [{"title": "YPF refinery blending — interview with BCG project lead Roberto Rodrigo", "url": "https://mase.lmneuquen.com/empresas/ia-ypf-la-tecnologia-que-gano-el-premio-digital-del-ano-y-ya-se-aplica-la-refineria-la-plata-n1253008", "date": "Unknown", "type": "Original supplier interview"}], "status": "partial", "gaps": "Original interview with the supplier project lead, not an operator filing or independent result. No realized saving, quality/error measure, cost or net ROI. No stream, tank, blend, user or throughput denominator. Further rollout is an ambition. Separate from the earlier YPF cataloging and inspection profile.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "Ecopetrol", "existing_ids": ["AI-147", "AI-148", "AI-247"], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [{"title": "Ecopetrol 2025 Integrated Management Report — digital transformation and AIOps", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 (exact publication day unknown)", "type": "Operator annual report"}], "status": "partial", "gaps": "No all-in cost, cash reconciliation, comparison design or complete net ROI. The US$1.5 million-plus may overlap the existing AI-147 portfolio benefits; do not add them. Neither 160 digital initiatives nor more than 100 operating solutions is a count of AI deployments.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "PRIO", "existing_ids": ["AI-277"], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "PetroReconcavo", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "GeoPark", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "Parex", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Latin American operators", "reviewed_on": "2026-10-25"}, {"company": "Seplat Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / African and private operators", "reviewed_on": "2026-10-25"}, {"company": "Tullow Oil", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / African and private operators", "reviewed_on": "2026-10-25"}, {"company": "Kosmos Energy", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / African and private operators", "reviewed_on": "2026-10-25"}, {"company": "Perenco", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / African and private operators", "reviewed_on": "2026-10-25"}, {"company": "Databricks", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Platform customer evidence", "reviewed_on": "2026-10-25"}, {"company": "Snowflake", "existing_ids": ["AI-119"], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Platform customer evidence", "reviewed_on": "2026-10-25"}, {"company": "C3 AI", "existing_ids": [], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Platform customer evidence", "reviewed_on": "2026-10-25"}, {"company": "Cognite", "existing_ids": ["AI-120", "AI-121"], "searched": true, "queries": ["Exact six-clause research assignment and screening decisions in the dated ledger"], "documents_reviewed": [], "status": "partial", "gaps": "No additional qualifying workflow admitted in this bounded assignment; duplicates and access limits are retained in the ledger. This does not establish inactivity.", "cohort": "25 October evening / Platform customer evidence", "reviewed_on": "2026-10-25"}], "excluded_or_unverified": [{"company": "Chevron", "existing_ids": ["AI-005"], "claim": "West Texas data-center power was presented in original register alongside AI adoption.", "reason": "This is energy supply/infrastructure for AI demand, not company AI adoption; classify separately as AI-demand infrastructure, outside internal-use initiative count.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "cohort": "majors"}, {"company": "ExxonMobil", "existing_ids": ["AI-073"], "claim": "AI identifies four new Guyana exploration opportunities (2026).", "reason": "Secondary news result found; did not locate/open matching ExxonMobil primary release in this pass. Do not certify without primary corroboration.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai", "cohort": "majors"}, {"company": "Shell", "existing_ids": ["AI-076"], "claim": "C3 AI and Shell expand root-cause analysis across global asset operations (2026).", "reason": "Existing record links BusinessWire; official Shell announcement not found/opened in pass. Secondary release needs verification; listed official Shell AI pages establish generic Shell e/model platform but not this named expansion.", "source_url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations", "cohort": "majors"}, {"company": "Santos", "existing_ids": ["AI-083"], "claim": "Xecta AI digital twin expanded to East Australia and PNG (2025).", "reason": "Existing source is trade press. Official 2026 investor briefing excerpt reviewed does not independently verify Xecta rollout claim.", "source_url": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta", "cohort": "majors"}, {"company": "ExxonMobil", "existing_ids": ["AI-072"], "claim": "First fully autonomous AI-drilled well offshore Guyana (2026).", "reason": "Official page confirms current AI advisory and closed-loop drilling in Guyana, but source does not verify the specific first-well milestone or timing.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "cohort": "majors"}, {"company": "Eni", "existing_ids": ["AI-027"], "claim": "Broad Eni AI portfolio/value claims from Capital Markets Update.", "reason": "Existing official PDF source remains not re-opened in this cohort pass; no contradiction found, but portfolio-level claims not re-verified against its source text.", "source_url": "https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "cohort": "majors"}, {"company": "ADNOC", "existing_ids": ["AI-071"], "claim": "40,000+ employees AI-trained; 70,000 productivity hours per month", "reason": "Existing profile cites a news source rather than primary company evidence; not re-verified in reviewed primary disclosures. Do not treat as disproved.", "source_url": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/", "cohort": "nocs"}, {"company": "CNOOC", "existing_ids": ["AI-082"], "claim": "Offshore predictive maintenance reduces failures by approximately 20%", "reason": "Existing claim cites secondary China Daily coverage; the reviewed official interim results discuss AI+ scenarios but do not substantiate this specific metric. Unverified, not disproved.", "source_url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "cohort": "nocs"}, {"company": "Kuwait Oil Company", "existing_ids": ["AI-014", "AI-066", "AI-085"], "claim": "AI seismic processing, Jurassic gas AI and G Agent drilling outcomes", "reason": "No official KOC source review completed in this pass to substantiate these entries; the SLB earnings source is vendor disclosure and Kuwait Times is secondary. Retain as reported/unverified pending primary KOC technical or procurement evidence.", "source_url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/", "cohort": "nocs"}, {"company": "OQ Exploration & Production", "existing_ids": ["AI-013"], "claim": "Autonomous drilling project in Oman", "reason": "Existing cited source is SLB earnings material; no OQ primary-source verification completed. Vendor-reported partnership, not disproved.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "cohort": "nocs"}, {"company": "PTTEP", "existing_ids": ["AI-067"], "claim": "Offshore production AI pilot", "reason": "Existing cited source is SLB earnings material; no PTTEP primary-source verification completed. Vendor-reported and not re-verified.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "cohort": "nocs"}, {"company": "QatarEnergy", "existing_ids": [], "claim": "No specific qualifying AI operating initiative located in this pass", "reason": "Search returned sustainability-report references to digitalization/AI analytics and AI/BI recruiting, but no adequately specific, named operational AI use case was verified. This is a search gap, not proof of absence.", "source_url": "https://www.qatarenergy.qa/en/MediaCenter/Publications/QatarEnergy%202024%20Sustainability%20Report.pdf", "cohort": "nocs"}, {"company": "ONGC", "existing_ids": [], "claim": "Automated Discontinuity Tracker and HYDRATEMIND", "reason": "Official government report surfaced, but appears to describe 2022-23 status, outside the requested time window; not counted as current initiative.", "source_url": "https://www.dpe.gov.in/static/uploads/2025/07/29d117aeeefcf739cbd9b0b937348d62.pdf", "cohort": "nocs"}, {"company": "Suncor Energy", "existing_ids": ["AI-099"], "claim": "Autonomous haul fleet “nears 140 trucks” in 2025.", "reason": "Official Suncor feature says nearly 120 were operating, with 140 planned by year end 2026; 140 is a target.", "source_url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "cohort": "north_america"}, {"company": "EQT Corporation", "existing_ids": ["AI-063"], "claim": "EQT gas supply is part of AI adoption.", "reason": "This is natural-gas supply infrastructure serving AI/HPC power demand; categorize separately from internal AI. Agreement was in principle.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "cohort": "north_america"}, {"company": "Matador Resources", "existing_ids": ["AI-096"], "claim": "Named vendor-led technical AI deployment.", "reason": "Official release only identifies AI integration as a focus area; CEO transcript is third-party-hosted and identifies no vendors or use case.", "source_url": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "cohort": "north_america"}, {"company": "Ovintiv", "existing_ids": ["AI-092", "AI-093", "AI-094"], "claim": "AI use cases and benefits across production, drilling and completions.", "reason": "Detailed workflow claims are currently supported by third-party transcript excerpts only; no primary issuer transcript was reviewed. Not disproved.", "source_url": "https://investors.ovintiv.com/", "cohort": "north_america"}, {"company": "EOG Resources", "existing_ids": ["AI-088", "AI-089"], "claim": "Generative AI, production optimizers, remote control deployment.", "reason": "Existing citations are third-party transcript copies; issuer-source corroboration outstanding. Not disproved.", "source_url": "https://investors.eogresources.com/Events-and-Presentations", "cohort": "north_america"}, {"company": "Cenovus Energy", "existing_ids": ["AI-061"], "claim": "Productivity and decision-support pilots.", "reason": "Cited MD&A text was not reopened; current evidence is insufficient to verify claim. Not disproved.", "source_url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "cohort": "north_america"}, {"company": "Diamondback Energy", "existing_ids": ["AI-045"], "claim": "AI-driven lift optimization produced 1% decline improvement.", "reason": "Existing source is secondary; decline-rate reference was illustrative and no measured AI-specific result was verified.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "cohort": "north_america"}, {"company": "Chord Energy", "existing_ids": ["AI-097"], "claim": "AI rod-pump optimization halted base decline.", "reason": "Third-party transcript combines AI controls with pump depth and workovers; AI-only causal effect is not established.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "cohort": "north_america"}, {"company": "SLB", "existing_ids": ["AI-008"], "claim": "33 deployment-ready Tela workflows", "reason": "Current SLB product page confirms 33 workflows; this is a catalog/capability count, not completed customer deployments. Product page is undated and mutable.", "source_url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "cohort": "services"}, {"company": "Halliburton", "existing_ids": ["AI-110"], "claim": "Chevron intelligent fracturing is AI", "reason": "Partner release confirms closed-loop, feedback-driven automation but does not establish that the system uses AI. Categorize as automation unless additional evidence confirms AI.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "cohort": "services"}, {"company": "Precision Drilling", "existing_ids": [], "claim": "Alpha suite quantified AI value", "reason": "2024 ESG report confirms AI and predictive-maintenance applications across a 78-rig digital fleet, but gives no AI-specific outcome metric.", "source_url": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "cohort": "services"}, {"company": "Transocean", "existing_ids": [], "claim": "AI in offshore drilling operations", "reason": "2025 Form 10-K explicitly says AI is not currently used in offshore drilling operations; office/admin use exists. Automated drilling control and robotics are separate.", "source_url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "cohort": "services"}, {"company": "TechnipFMC", "existing_ids": [], "claim": "Named AI deployment in customer operations", "reason": "Subsea Studio material discusses AI deployment and remote operations generally, but no dated named AI customer case was verified within the window. This does not imply absence.", "source_url": "https://www.technipfmc.com/en/what-we-do/subsea/subsea-studio/", "cohort": "services"}, {"company": "Nabors", "existing_ids": ["AI-115"], "claim": "Predictive Drilling 36% ROP improvement and 9.7% vibration reduction are current-window results", "reason": "Company product page reports these results, but Nabors primary case page dates first deployment and metrics to 2023; keep them as historical baseline only.", "source_url": "https://www.nabors.com/predictive-drilling-results-delaware/", "cohort": "services"}, {"company": "Applied Computing", "existing_ids": ["AI-127"], "claim": "Applied Computing $20m Series A and claims about broad whole-plant model deployments.", "reason": "No primary company/investor funding release reviewed; only secondary reporting surfaced. The May 2025 primary source confirms a £9m seed and one unnamed refinery claim, not a later round or independently validated whole-plant performance.", "source_url": "https://appliedcomputing.com/press-release-28052025", "cohort": "software_capital"}, {"company": "Quantum Capital Group", "existing_ids": ["AI-052", "AI-053", "AI-054", "AI-055"], "claim": "More than 100 portfolio AI accelerators represent in-window deployments.", "reason": "Official 2024 ESG PDF excerpt puts operating coverage at May 2024, before window; publication date in May 2025 does not change operational date. PDF could not be opened in full.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "cohort": "software_capital"}, {"company": "Encino Energy / Microsoft", "existing_ids": [], "claim": "Microsoft Power Platform saved $1.2m/year on winter maintenance as an AI outcome.", "reason": "Microsoft case attributes this to a rules-and-workflow Power Apps/cloud-flow plowing process, not AI; excluded from AI economics.", "source_url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "cohort": "software_capital"}, {"company": "AVEVA", "existing_ids": ["AI-129"], "claim": "Q1 2027 CONNECT major AI release was deployed by end of research window.", "reason": "2026 announcement states this major release is planned for Q1 2027; future roadmap is outside window.", "source_url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "cohort": "software_capital"}, {"company": "Novi Labs", "existing_ids": ["AI-123"], "claim": "Novi/Turing/RBN acquisitions already form one integrated product.", "reason": "Acquisition announcements describe strategy and intended unified platform; reviewed sources do not establish technical integration completion.", "source_url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "cohort": "software_capital"}, {"company": "Applied Computing", "existing_ids": ["AI-127"], "claim": "$20m Series A led by KBR and Databricks Ventures; entire-plant model performance.", "reason": "No primary financing announcement located in bounded review; legacy evidence is TechCrunch secondary. Primary Applied Computing/KBR Jan 2026 release verifies INSITE 3.0 powered by Orbital for ammonia process optimization, but does not verify $20m round or the full-plant claims. Preserve as secondary-only and distinct from the £9m seed/unnamed refinery record.", "source_url": "https://appliedcomputing.com/press-release-21012026", "cohort": "software_capital"}, {"company": "Emerson / AspenTech", "existing_ids": ["AI-130"], "claim": "Emerson completion of remaining AspenTech acquisition.", "reason": "Verified in Emerson primary release on 2025-03-12; acquisition transaction is ownership context and does not itself establish an AI deployment. Legacy classification as an AI initiative should be treated as transaction context.", "source_url": "https://ir.emerson.com/news-events/press-releases/detail/11/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "cohort": "software_capital"}, {"company": "Chevron", "existing_ids": ["AI-003", "AI-004"], "claim": "Workforce productivity / abandonment preparations / annual billions in AI value as concrete realized initiatives.", "reason": "Investor Day establishes broad AI ambition and value potential but does not quantify realized AI-only returns or provide a distinct named operational workflow for each legacy entry. APOLO is profiled separately.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "cohort": "majors_backfill"}, {"company": "Chevron", "existing_ids": ["AI-005"], "claim": "West Texas data-center power project classified as AI adoption.", "reason": "This is power infrastructure to supply AI demand; infrastructure for AI is not use of AI in Chevron operations.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "cohort": "majors_backfill"}, {"company": "Shell", "existing_ids": ["AI-025"], "claim": "Shell $5-7bn target as AI-generated savings.", "reason": "Shell lists AI among examples within a multicomponent $5-7bn structural-cost-reduction target from 2022–2028; no AI-only attribution.", "source_url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "cohort": "majors_backfill"}, {"company": "Shell", "existing_ids": ["AI-026"], "claim": "Immersion cooling fluids as oil-and-gas AI adoption.", "reason": "AI data center cooling product / infrastructure market, not AI use by Shell operating functions.", "source_url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "cohort": "majors_backfill"}, {"company": "ExxonMobil", "existing_ids": ["AI-023"], "claim": "Low-carbon data-center opportunity as internal AI workflow.", "reason": "Data-center power/carbon solution is infrastructure and commercial opportunity in the AI-demand ecosystem.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "cohort": "majors_backfill"}, {"company": "ExxonMobil", "existing_ids": ["AI-024"], "claim": "ERP simplification and marine logistics planning AI.", "reason": "Original reference is a third-party transcript; no Exxon primary source for a named AI workflow was located in this bounded pass. ERP simplification is not itself AI. This is not a disproval; it remains unverified.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "cohort": "majors_backfill"}, {"company": "ExxonMobil", "existing_ids": ["AI-073"], "claim": "AI identified four new exploration opportunities in Guyana.", "reason": "Only secondary reports/call reproductions surfaced; no primary Exxon release was located. Treat as secondary-only pending direct company confirmation. The claim is not disproved.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "cohort": "majors_backfill"}, {"company": "bp", "existing_ids": ["AI-060"], "claim": "2024 Palantir and Infosys refining digital implementation is established AI workflow.", "reason": "Official transcript states digital progress and an ontology, not an AI model or named deployed AI use case. Preserve as unverified digital transformation, not confirmed AI.", "source_url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/investors/bp-third-quarter-2024-results-qa-transcript.pdf", "cohort": "majors_backfill"}, {"company": "PETRONAS", "existing_ids": ["AI-079"], "claim": "Microsoft MoU proves agentic AI/Copilot deployment across PETRONAS value chain.", "reason": "Primary release says partners will explore initiatives and technologies; deployment scope/results are not described. AIngineer is separately profiled as a product/use-case collaboration.", "source_url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "cohort": "majors_backfill"}, {"company": "Santos", "existing_ids": ["AI-083"], "claim": "Five-year Xecta contract, Cooper Basin rollout and Kelvin.ai autonomous control are independently confirmed.", "reason": "Xecta official index supports only Santos IPS across Eastern Australia/PNG in accessible material. Original source body unavailable; other details not reverified. This is a gap, not disproval.", "source_url": "https://xecta.com/news-resources/", "cohort": "majors_backfill"}, {"company": "Emerson / AspenTech", "existing_ids": ["AI-130"], "claim": "Completion of Emerson acquisition of remaining AspenTech shares as an AI implementation.", "reason": "The transaction is verified, but the official acquisition notice establishes ownership and deal completion, not an AI workflow. Keep as corporate transaction context.", "source_url": "https://ir.emerson.com/news-events/press-releases/detail/11/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "cohort": "majors_backfill"}, {"company": "INPEX", "existing_ids": [], "claim": "Current official digital-strategy material describes an internal GenAI search application and an AI-enabled security operations center.", "reason": "Page is undated and accessible page does not establish when these capabilities launched or were deployed; therefore in-window adoption is unverified. This is no re-verification, not a disproved claim.", "source_url": "https://www.inpex.com/english/business/technology/dx/", "cohort": "regional"}, {"company": "Sonatrach", "existing_ids": [], "claim": "AI seminar and company R&D/digital initiatives.", "reason": "The located official AI event was outside the requested window; other material is strategy or generic R&D without a named deployed operational workflow.", "source_url": "https://sonatrach.com/wp-content/uploads/2025/06/SH-NEWS-VF-N%C2%B045.pdf", "cohort": "regional"}, {"company": "NNPC Limited", "existing_ids": [], "claim": "Digital-transformation and maintenance programs.", "reason": "Located company sources describe maintenance or digital modernization without establishing an AI model in a named workflow; excluded from AI adoption count.", "source_url": "https://cms1977.nnpcgroup.com/uploads/NNPC_GMP_2026_Report_Final_For_Upload_51bfeafda6.pdf", "cohort": "regional"}, {"company": "Sonangol", "existing_ids": [], "claim": "Digital systems modernization and general references to AI in strategic material.", "reason": "2025 annual report discusses systems/network modernization, while strategy language does not identify an operational AI workflow. No qualifying initiative found; this is not proof of no use.", "source_url": "https://www.sonangol.co.ao/wp-content/uploads/2026/05/Relatorio-Anual-e-Contas-2025.pdf", "cohort": "regional"}, {"company": "PTT Public Company Limited", "existing_ids": [], "claim": "A neighboring PTTEP partner record (AI-067) exists in the master register.", "reason": "AI-067 belongs to PTTEP / SLB and is an affiliate/service-provider pilot, not an exact PTT-parent match. PTTGC initiatives are reported separately.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "cohort": "regional"}, {"company": "Petrovietnam", "existing_ids": [], "claim": "Affiliate-level reports of AI for production data, maintenance and fuel-demand forecasting.", "reason": "A PetroTimes group-affiliate article surfaced as a secondary lead, including plans, but no primary company evidence was verified for an in-window named workflow. No-reverification gap, not a disproved claim.", "source_url": "https://petrovietnam.petrotimes.vn/tri-tue-nhan-tao-trong-linh-vuc-dau-khi-nang-luong-lo-trinh-va-hien-thuc-hoa-732289.html", "cohort": "regional"}, {"company": "Gazprom Neft", "existing_ids": [], "claim": "AI smart drilling/autopilot and Digital Well program.", "reason": "Officially surfaced references did not establish a dated in-window primary disclosure with sufficient named workflow evidence during this bounded screen; unverified, not disproved.", "source_url": "https://t.me/s/gazpromneft_official/503", "cohort": "regional"}, {"company": "Reliance Industries Limited", "existing_ids": [], "claim": "Broad AI-driven process optimization in O2C improves yield and lowers energy consumption.", "reason": "Retained as context for the named crude-blend tool, but not a separate dossier because workflow, scale and metrics are not disclosed.", "source_url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "cohort": "regional"}, {"company": "PTT Global Chemical", "existing_ids": [], "claim": "AI price forecasting, customer analytics and GenAI digital assistant.", "reason": "The July 2026 release names these applications but provides limited workflow/deployment detail; omitted from the 5–8 strongest dossiers to prioritize plant operations, maintenance and energy optimization.", "source_url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "cohort": "regional"}, {"company": "APA Corporation / Palantir", "existing_ids": [], "claim": "APA Palantir expansion included Palantir AIP across its operations and covered planning, supply chain, maintenance, production and contracts.", "reason": "Primary announcement date is 2024-09-25, one day before the window. 2025 10-K contains generic technology risk/evaluation language but did not verify an in-window named workflow.", "source_url": "https://www.nasdaq.com/press-release/apa-corporation-expands-partnership-palantir-leverage-ai-technology-across-their-oil", "cohort": "round2/additions-1"}, {"company": "SM Energy / Civitas Resources", "existing_ids": [], "claim": "Proposed combined company expected to optimize operations with machine learning and AI.", "reason": "Nov 3, 2025 merger-call statement is forward-looking synergy intent and does not define a workflow, implementation stage, or realized AI-specific value; not promoted to an initiative dossier.", "source_url": "https://www.sm-energy.com/investors/financial-information/all-sec-filings/content/0000893538-25-000149/smandcivitasmergerjointc.htm", "cohort": "round2/additions-1"}, {"company": "Range Resources / Antero Resources", "existing_ids": [], "claim": "AI references in investor communication relate to growing natural-gas demand from data centers.", "reason": "AI-demand infrastructure/market context, not internal AI adoption.", "source_url": "https://www.anteroresources.com/investors/press-releases/detail/247/antero-resources-announces-second-quarter-2025-financial", "cohort": "round2/additions-1"}, {"company": "Hess Corporation", "existing_ids": [], "claim": "AI mentions in Hess 2024 annual report.", "reason": "AI is discussed as a cybersecurity risk scenario, not an operational use. Hess is treated only as the standalone historical company; no in-window deployed workflow surfaced.", "source_url": "https://www.sec.gov/Archives/edgar/data/4447/000162828025008716/hes-20241231.htm", "cohort": "round2/additions-1"}, {"company": "Cheniere Energy", "existing_ids": [], "claim": "Potential AI adoption in LNG operations.", "reason": "The 2025 Form 10-K search did not surface a matching AI disclosure; no evidence identified in this pass. This is a research gap, not proof that no internal use exists.", "source_url": "https://lngir.cheniere.com/sec-filings/all-sec-filings/content/0000003570-26-000005/lng-20251231.htm", "cohort": "round2/additions-1"}, {"company": "Gulfport Energy", "existing_ids": [], "claim": "Potential AI adoption in upstream operations.", "reason": "Search did not identify a specific workflow in primary sources; stable filing sources were not reopened. Unverified, not disproved.", "source_url": "https://www.gulfportenergy.com/investors/financials/sec-filings/default.aspx", "cohort": "round2/additions-1"}, {"company": "KazMunayGas", "existing_ids": [], "claim": "ABAI Flood Control AI module shortens flood-analysis time and delivers KZT savings.", "reason": "Samruk-Kazyna company page search excerpt indicated these claims, but direct URL could not be reopened in the allotted pass; do not add as verified until source is accessible.", "source_url": "https://sk.kz/press-center/news/78145/", "cohort": "round2/additions-2"}, {"company": "Kuwait Petroleum Corporation", "existing_ids": [], "claim": "KPC AI models are used for seismic interpretation, drilling optimization and predictive maintenance.", "reason": "Only a search excerpt from a large PDF was available; no full text or named asset/workflow details verified.", "source_url": "https://www.kpc.com.kw/AnnualReports/KPC-Sustainability-Report-2023-25.pdf", "cohort": "round2/additions-2"}, {"company": "Bapco Energies", "existing_ids": [], "claim": "AI will optimize Bapco Modernization Program operations.", "reason": "Annual-report excerpt is a strategic aim, not evidence of a launched or operating AI workflow.", "source_url": "https://www.bapcoenergies.com/media-centre/reports/download/annual-report-2024", "cohort": "round2/additions-2"}, {"company": "OQ Group", "existing_ids": ["AI-013"], "claim": "Seven AI use cases executed and two agentic AI cases developed; >$10m return from digital portfolio.", "reason": "Search excerpt of OQ 2025 report exposes aggregate figures but not named cases or AI-attributable portion of overall digital economics; existing Neuro Oman is separately audited.", "source_url": "https://assets.oq.com/-/media/oq/files/sustainability/sustainability-at-oq/oq-sustainability-report--2025--english.pdf?rev=-1", "cohort": "round2/additions-2"}, {"company": "SOCAR Tech", "existing_ids": [], "claim": "SOCAR ML applications predict fuel quality and optimize refinery/energy operations.", "reason": "Undated current capability page does not establish first deployment or in-window operations.", "source_url": "https://socartech.az/en/services/data-management", "cohort": "round2/additions-2"}, {"company": "Sonatrach", "existing_ids": [], "claim": "Publicly disclosed AI operational program in window.", "reason": "No qualifying named primary-source workflow surfaced in targeted search. This is a search gap, not evidence of non-adoption.", "source_url": "", "cohort": "round2/additions-2"}, {"company": "Peloton Computer Enterprises", "existing_ids": [], "claim": "Peloton is an AI oil-and-gas workflow/product.", "reason": "Reviewed platform pages describe integrated data management, lifecycle workflows, automation, ETL and real-time data. No named AI method or customer AI deployment was evidenced. Carlyle’s 2026 investment announcement says AI is a forward growth driver only.", "source_url": "https://www.peloton.com/peloton-platform-overview", "cohort": "round2/additions-3"}, {"company": "W Energy Software", "existing_ids": [], "claim": "Stream+ is a deployed AI system for oil-and-gas operations.", "reason": "Official description covers integrated upstream/midstream ERP, accounting, production and business intelligence; reviewed page does not document a specific AI component or implemented AI workflow.", "source_url": "https://wenergysoftware.com/resources/insights/introducing-streamplus-the-future-of-oil-gas-software/", "cohort": "round2/additions-3"}, {"company": "IFS / Noble Corporation", "existing_ids": [], "claim": "IFS.ai has been implemented for predictive maintenance at Noble.", "reason": "IFS customer page frames predictive failure analysis, automated FMECA/work orders and maintenance scheduling as expected benefits of Noble’s planned IFS Cloud upgrade; it is not proof of operating deployment.", "source_url": "https://www.ifs.com/en/insights/customer-stories/noble-corporation", "cohort": "round2/additions-3"}, {"company": "ComboCurve", "existing_ids": [], "claim": "ComboCurve has a qualifying AI deployment/product.", "reason": "Official about page describes technical forecasting and economic modeling software but this review found no named AI technique or customer workflow.", "source_url": "https://combocurve.com/about-us/", "cohort": "round2/additions-3"}, {"company": "Verdane / Lime Rock / EnCap / NGP / Warburg Pincus", "existing_ids": [], "claim": "Investor ownership itself proves AI adoption in oil-and-gas portfolio companies.", "reason": "Investor mandate/portfolio pages are not operational evidence. The review found no dated primary evidence tying a named oil-and-gas AI workflow to these funds. NGP lists AI-enabled PhysicsX and analytics company Orennia, but portfolio descriptions do not establish oil-and-gas operating deployment.", "source_url": "https://ngpenergycapital.com/portfolio-energy-transition/", "cohort": "round2/additions-3"}], "cohorts": ["majors", "nocs", "north_america", "services", "software_capital", "midstream", "root_backfill", "majors_backfill", "regional", "round2/additions-1", "round2/additions-2", "round2/additions-3", "round2/additions-root"], "date": "2026-10-01", "method": "Targeted company and topic searches across a named screening universe. This is not a census or a document-by-document review of all earnings and investor materials.", "standouts": [{"key": "aramco-fadhili-agr", "group": "Control of physical operations", "why": "A bounded process-control application provides a clearer adoption test than a general AI strategy: control authority, stability and repeatability can be examined at the unit level.", "next": "Operating envelope, safety controls, sustained performance and results at a second plant."}, {"key": "slb-cnooc-terminal", "group": "Control of physical operations", "why": "Connecting a process model to plant control moves AI closer to operating decisions. That integration makes both technical ambition and validation requirements substantial.", "next": "Operator corroboration, control safeguards, baseline economics and measured operating results."}, {"key": "halliburton-coterra-autofrac", "group": "Control of physical operations", "why": "Automated stage execution reaches into the completion workflow itself. Repeatability across jobs matters more than a single efficiency headline.", "next": "Comparable job results, exception handling and the denominator behind the reported efficiency gain."}, {"key": "aker-yggdrasil-agents", "group": "Enterprise and asset transformation", "why": "Designing remote operations around agents from the outset is more extensive than adding an assistant to an existing process. The operational proof remains ahead.", "next": "Commissioning milestones, human accountability and results after the planned 2027 start."}, {"key": "total-cognite-industrial-ai", "group": "Enterprise and asset transformation", "why": "A shared industrial-data foundation can enable many workflows across assets. Its ambition lies in organizational scale and reusable integration.", "next": "Site-by-site adoption, data quality, deployed applications and attributable outcomes."}, {"key": "adnoc-energyai", "group": "Enterprise and asset transformation", "why": "Coordinating specialized agents across subsurface tasks is technically ambitious. Trial evidence and later contractual scope must remain distinct from completed field rollout.", "next": "Production deployments, benchmark definitions and independently inspectable performance."}, {"key": "total-mistral-frontier-models", "group": "New models and investment workflows", "why": "Developing specialized subsurface models could change how alternative exploration and reservoir decisions are evaluated. This is research ambition at an early stage.", "next": "Model evaluations, operating integrations and results against established engineering methods."}, {"key": "quantum-digital-mission", "group": "New models and investment workflows", "why": "Shared investment and operating data can connect asset screening, capital allocation and portfolio surveillance. The breadth is compelling, but public deployment detail is limited.", "next": "Named portfolio deployments, decision-control design and validated economic outcomes."}, {"key": "collide-battalion", "group": "New models and investment workflows", "why": "A named operator committing records and multiple business processes provides a concrete adoption pathway. The equity investment and proposed software rollout are separate milestones.", "next": "Delivery against the rollout sequence, user adoption and workflow-level measured results."}], "audit": {"edition": "Second research and audit pass", "date": "2026-09-27", "method": "Three Luna researchers reviewed assigned profiles and recorded source access, findings, corrections and residual gaps. A recorded outcome is not independent validation of company-reported performance. Sources marked inaccessible, excerpt-only or not re-reviewed retain those limitations.", "baseline_profiles": 187, "baseline_audited": 187, "missing_baseline_ids": [], "added_profiles": 12, "outcomes": {"supported": 27, "partially supported": 142, "corrected": 14, "unverified": 4}, "source_access_entries": {"excerpt reviewed": 24, "full text reviewed": 237, "inaccessible": 29}, "unique_sources_reopened": 248, "baseline_sources": 301, "baseline_sources_attempted": 301, "missing_baseline_source_urls": [], "record_audits": [{"id": "AI-011", "company": "ADNOC / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-012", "company": "bp / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-014", "company": "Kuwait Oil Company / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-022", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-023", "company": "ExxonMobil", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"], "findings": [{"severity": "medium", "field": "category / classification", "issue": "The disclosed data-center opportunity concerns energy supply for AI demand, not ExxonMobil adopting AI in its own workflow.", "correction": "Classify as AI-demand infrastructure/strategic opportunity and keep distinct from internal AI adoption.", "source_url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-024", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-027", "company": "Eni", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-028", "company": "Eni / Khazna", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-040", "company": "ADNOC / AIQ", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-045", "company": "Diamondback Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-056", "company": "Quorum / Streamba", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "https://www.franciscopartners.com/investments/quorum-software"], "findings": [{"severity": "low", "field": "classification / deployment", "issue": "Acquisitions and planned integration enlarge a vendor portfolio but do not alone establish an integrated AI workflow in production.", "correction": "Describe deal/integration stage explicitly; avoid implying realized customer outcomes.", "source_url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-060", "company": "bp / Palantir / Infosys", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "https://palantir2020ipo.q4web.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/default.aspx"], "findings": [{"severity": "high", "field": "Research_category / AI basis", "issue": "The in-window Q3 2024 bp call describes digital progress with Palantir and Infosys at the refining system and says the ontology was built; it does not establish an AI workflow. Palantir’s AI-capabilities relationship announcement is dated 2024-09-09, before the window.", "correction": "Classify as excluded/unverified AI adoption (digital transformation only) unless a later AI-specific operational disclosure is found.", "source_url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter"}], "residual_gaps": ["No qualifying in-window source in this assignment establishes AI use in bp refining. Existing profile already flags the issue; classification patch makes it explicit."], "partition": 1, "reviewer_outcome": "corrected"}, {"id": "AI-063", "company": "EQT / Homer City Redevelopment", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp"], "findings": [{"severity": "low", "field": "category / stage", "issue": "These records are infrastructure or supply arrangements responding to AI-load demand, not operational AI systems.", "correction": "Retain only in the AI-demand infrastructure category and do not describe gas infrastructure itself as AI adoption.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-064", "company": "Williams", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc"], "findings": [{"severity": "low", "field": "category / stage", "issue": "These records are infrastructure or supply arrangements responding to AI-load demand, not operational AI systems.", "correction": "Retain only in the AI-demand infrastructure category and do not describe gas infrastructure itself as AI adoption.", "source_url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-066", "company": "Kuwait Oil Company / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-069", "company": "ExxonMobil Trinidad and Tobago / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-071", "company": "ADNOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-072", "company": "ExxonMobil", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-073", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-081", "company": "PetroChina / CNPC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml"], "findings": [{"severity": "medium", "field": "stage / metrics", "issue": "CNPC’s official portal search-indexed article describes 152 Kunlun AI scenarios as scaled/deployed and reports selected speed/cost/accuracy results; the Chinese direct page could not be fetched, so this is excerpt-level verification. The separate 2025 annual report discusses earlier model/scenario scale.", "correction": "Retain reported scale/results as company-reported, but mark 2026 article excerpt-only and do not merge the 2025 model disclosures with later deployment dates.", "source_url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml"}], "residual_gaps": ["Direct 2026 article fetch failed (encoding/internal error); an official search-index excerpt and official 2025 annual report were reviewed. More operating-unit detail and independent validation remain undisclosed."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-083", "company": "Santos", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://xecta.com/news-resources/", "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-085", "company": "Kuwait Oil Company", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-090", "company": "ConocoPhillips", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf"], "findings": [{"severity": "medium", "field": "stage / workflow", "issue": "Enterprise AI intent is a strategy disclosure unless a specific workflow and deployment are documented.", "correction": "Do not imply operational deployment or measurable results beyond the disclosed strategy.", "source_url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-092", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-093", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-094", "company": "Ovintiv", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "low", "field": "duplicates / scope", "issue": "Portfolio-level Ovintiv AI capability and named DRIVE/frac workflows overlap but are different levels of disclosure.", "correction": "Keep the broad portfolio separate from concrete use cases; avoid adding component details to the portfolio as separately attributed outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-095", "company": "EQT", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-097", "company": "Chord Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-098", "company": "Canadian Natural Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-103", "company": "Williams", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-108", "company": "Microsoft", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-117", "company": "Patterson-UTI", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-122", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/"], "findings": [{"severity": "low", "field": "classification", "issue": "The $35m transaction is a growth investment in Novi Labs; it does not evidence operator deployment or customer impact.", "correction": "Keep as investment/acquisition event and separate from Novi product workflow dossiers.", "source_url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/"}], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-123", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "https://novilabs.com/news/novi-labs-acquires-rbn-energy/"], "findings": [{"severity": "low", "field": "classification / deployment", "issue": "Acquisitions and planned integration enlarge a vendor portfolio but do not alone establish an integrated AI workflow in production.", "correction": "Describe deal/integration stage explicitly; avoid implying realized customer outcomes.", "source_url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-127", "company": "Applied Computing", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/"], "findings": [{"severity": "medium", "field": "source / attribution", "issue": "The $20m Series A record relies on secondary TechCrunch in the legacy source ledger; an issuer/investor primary financing notice was not reopened in this audit.", "correction": "Keep financing amount and operational claims secondary-only until issuer/investor confirmation; separate from the earlier May 2025 seed raise.", "source_url": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-136", "company": "bp", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf"], "findings": [{"severity": "medium", "field": "disclosed_results", "issue": "Indexed official report text adds company-reported results: driver-distraction events down 89% and tailgating events down 75% between Dec 2023 and Dec 2024. Direct PDF fetch returned 403; evidence is excerpt-level.", "correction": "Add both metrics with the stated Dec 2023–Dec 2024 comparison period and company attribution; do not characterize as independently validated or as incident-rate reductions.", "source_url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf"}], "residual_gaps": ["Report does not disclose fleet denominator, sensor/model details, absolute event counts, or independent validation; PDF direct fetch remained blocked, so the report passage was reviewed via official search indexing only."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-137", "company": "Eni", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html"], "findings": [], "residual_gaps": ["No record-specific correction identified from the assignment/packet comparison; cited-source details not re-opened remain subject to source-level status."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-139", "company": "Santos", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "low", "field": "scope", "issue": "Portfolio or enterprise-level narratives aggregate multiple sub-workflows; specific named components should not be merged into a single deployment without attribution.", "correction": "Keep portfolio scope and concrete component/workforce initiatives distinct; retain only disclosed deployment status and metrics.", "source_url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf"}], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-144", "company": "Sinopec", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-152", "company": "ONGC", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207"], "findings": [{"severity": "low", "field": "stage", "issue": "The official tender specifies AI-driven optimization/predictive maintenance and gas-lift adjustment using SCADA/HMI with safe limits, but is an EOI/pilot solicitation rather than proof of awarded, commissioned deployment.", "correction": "Keep stage as proposed pilot / market exploration; do not represent as live operating AI.", "source_url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207"}], "residual_gaps": ["Tender source defines requirements and solicitation intent; no award, commissioning, or measured results located in this packet review."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-156", "company": "ONGC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-159", "company": "Precision Drilling", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, {"id": "AI-161", "company": "Transocean", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-162", "company": "Kuwait Oil Company", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1"], "findings": [{"severity": "medium", "field": "stage", "issue": "Procurement specifications/contracts establish solicitation or award, not necessarily operational commissioning.", "correction": "Preserve planned/contract stage unless evidence verifies a deployed asset workflow.", "source_url": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-163", "company": "Novi Labs", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-164", "company": "Applied Computing", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://appliedcomputing.com/press-release-28052025"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-170", "company": "TC Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-171", "company": "TC Energy", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "supported"}, {"id": "AI-172", "company": "Enbridge", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152"], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-178", "company": "ExxonMobil", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence"], "findings": [{"severity": "low", "field": "dates / workflow", "issue": "ExxonMobil Guyana/operating-site disclosures should be separated by workflow and date; an undated technology page is not a launch date.", "correction": "Treat the operating date as undated unless a dated disclosure confirms it; do not conflate AI drilling advisory, exploration opportunities, and bottleneck monitoring.", "source_url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence"}], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-181", "company": "Bharat Petroleum Corporation (BPCL)", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, {"id": "AI-182", "company": "Hindustan Petroleum Corporation (HPCL)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.hindustanpetroleum.com/NewsroomDetails/449"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-184", "company": "PTT Global Chemical (PTTGC / GC)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-185", "company": "PTT Global Chemical (PTTGC / GC)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness"], "findings": [], "residual_gaps": ["Some inherited packet URLs were inaccessible or excerpt-only on this pass; see the per-URL source_audits. This limits verification but does not itself disprove the claim."], "partition": 1, "reviewer_outcome": "partially supported"}, {"id": "AI-186", "company": "Rosneft", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, {"id": "AI-187", "company": "Rosneft", "outcome": "unverified", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["One or more cited source URLs failed or were not re-reviewed; see source_audits for exact source access state."], "partition": 1, "reviewer_outcome": "unverified"}, {"id": "AI-008", "company": "SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-009", "company": "SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f"], "findings": [{"severity": "medium", "field": "economics", "issue": "SLB Digital ARR is a broad digital business metric and does not isolate Lumi revenue or adoption.", "correction": "Keep platform positioning; do not assign corporate digital ARR to Lumi.", "source_url": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-013", "company": "OQ Exploration & Production / SLB", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-015", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [{"severity": "medium", "field": "classification", "issue": "Data-center equipment supply is an AI-demand infrastructure business, not SLB internal AI adoption.", "correction": "Classify as AI-demand infrastructure; results reflect capacity/portfolio rather than operational AI use.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-016", "company": "Halliburton", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "LOGIX source establishes automation/autonomy but does not consistently identify the operating workflow as AI/ML.", "correction": "Describe the autonomous drilling workflow; retain AI evidence as unproven absent an explicit model disclosure.", "source_url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"}], "residual_gaps": ["The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-017", "company": "Halliburton", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "Product label and cloud E&P scope do not prove individual customers use an AI feature.", "correction": "Separate DS365.ai product capability from deployment evidence.", "source_url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-018", "company": "Halliburton / Shape Digital", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management"], "findings": [{"severity": "medium", "field": "stage", "issue": "Strategic collaboration announcement describes planned joint development; no named customer deployment or realized outcome is supplied.", "correction": "Stage as collaboration/product development, not an operating deployment.", "source_url": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-037", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks"], "findings": [{"severity": "medium", "field": "metrics / economics", "issue": "Total 2025 TRV is $5.3bn and combines technology value, but the CEO media-call source explicitly says over half was enabled by AI. Earlier audit wording that no AI allocation was disclosed was too strong.", "correction": "Retain the source’s company-reported “over half enabled by AI” attribution; keep total TRV and component values clearly marked as aggregate, not project-level or independently validated.", "source_url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks"}], "residual_gaps": ["No initiative-level allocation, causal methodology or independent validation is disclosed; the AI share is high-level management attribution."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-038", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation"], "findings": [{"severity": "medium", "field": "workflow", "issue": "The source describes a portfolio spanning multiple AI applications rather than one bounded initiative.", "correction": "Retain as a broad portfolio record; avoid attributing specific deployment metrics without program-specific support.", "source_url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-039", "company": "Aramco / PIF / HUMAIN", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release"], "findings": [{"severity": "high", "field": "stage", "issue": "Primary HUMAIN source describes a non-binding term sheet, subject to definitive agreements/approvals.", "correction": "Stage as proposed investment/agreement; do not present as completed acquisition or capital deployed.", "source_url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-041", "company": "Woodside Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/what-we-do/innovation/artificial-intelligence"], "findings": [{"severity": "medium", "field": "dates", "issue": "Woodside’s current page lacks a dated first deployment for Lumina; last-update/access date is not the launch date.", "correction": "Keep timeline undated/current-page only and do not claim first deployment occurred in-window.", "source_url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence"}], "residual_gaps": ["Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-042", "company": "Woodside Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside"], "findings": [{"severity": "high", "field": "metrics / economics", "issue": "Prior packet included 15% yearly maintenance-hours reduction for five years, but the reopened January 2026 Woodside article reports only potential savings generically and does not provide that figure.", "correction": "Remove the 15% metric and related economics sentence unless supported by another traceable source; retain the five-day-to-under-two-hour model-processing-time result.", "source_url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside"}], "residual_gaps": ["Reopened Jan 2026 page did not substantiate packet’s 15% annual maintenance-hours estimate; no evidence from that source says estimate was achieved."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-043", "company": "Woodside Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.woodside.com/what-we-do/innovation/artificial-intelligence"], "findings": [{"severity": "medium", "field": "dates", "issue": "Startup Advisor page gives no publication or launch date.", "correction": "Treat as a current capability with unverified in-window timing, not a dated deployment.", "source_url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence"}], "residual_gaps": ["Current-page evidence lacks initial launch/deployment date; recency cannot be inferred from access/update date."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-046", "company": "SLB / Qualcomm", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm"], "findings": [{"severity": "medium", "field": "stage", "issue": "Agora/Qualcomm announcement describes a collaboration and edge-AI solution scope, not a named operating customer.", "correction": "Classify as commercial collaboration/product roadmap until a field deployment is evidenced.", "source_url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-047", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy"], "findings": [{"severity": "medium", "field": "stage", "issue": "Marketplace launch and listed products establish an offering, not resulting operator adoption.", "correction": "Describe marketplace launch; no customer deployment or outcome inferred.", "source_url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-048", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Production Studio page is product material; results in related SLB digital-solution case studies should not be attributed to IPS without explicit linkage.", "correction": "Separate product capability from customer case metrics.", "source_url": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-050", "company": "PETRONAS", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-051", "company": "Blackstone / Enverus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf"], "findings": [{"severity": "medium", "field": "stage", "issue": "Blackstone’s August 2025 source is an acquisition agreement; the earnings material does not establish Enverus-specific capital deployed.", "correction": "Keep agreement/closing status separate from capital deployment; use no unallocated Blackstone fund metrics.", "source_url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-059", "company": "ECP / KKR / CyrusOne / Calpine", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas"], "findings": [{"severity": "medium", "field": "classification", "issue": "Campus provides power/data-center capacity serving AI demand, not an ECP/KKR oil-and-gas operational AI workflow.", "correction": "Keep solely under AI-demand infrastructure.", "source_url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-061", "company": "Cenovus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Cenovus MD&A describes governed productivity pilots; no specific measured value is attributed exclusively to AI.", "correction": "Retain as pilot portfolio and avoid treating broader productivity/automation results as AI-specific.", "source_url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-065", "company": "Enverus / Blackstone", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Enverus ONE launch source supports platform availability/capabilities, not customer use or outcomes.", "correction": "Call it a commercial AI platform launch; deployments not named.", "source_url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-067", "company": "PTTEP / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [{"severity": "medium", "field": "attribution", "issue": "SLB Q2 2026 is supplier-side corroboration for PTTEP offshore pilot; no direct PTTEP corroboration located in this audit.", "correction": "Label vendor-reported and preserve operator-confirmation gap.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-070", "company": "Saudi Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation"], "findings": [{"severity": "medium", "field": "stage", "issue": "Aramco/Microsoft source is a non-binding MoU describing areas to explore.", "correction": "Stage as exploratory partnership, not deployed solution.", "source_url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-079", "company": "Petronas", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem"], "findings": [{"severity": "medium", "field": "stage", "issue": "PETRONAS/Microsoft source is an ecosystem collaboration/MoU; it does not itself prove Copilot or agent deployments at PETRONAS.", "correction": "Keep as partnership/strategy; do not call operational adoption.", "source_url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-080", "company": "Petronas", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep"], "findings": [{"severity": "medium", "field": "stage", "issue": "PETRONAS release supports agentic-AI enhancement to myPROdata, but deployment/production results and investment impact are not quantified.", "correction": "Report platform development/rollout language as disclosed; omit inferred economic return.", "source_url": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-084", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-087", "company": "Coterra Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html"], "findings": [{"severity": "high", "field": "source quality", "issue": "Cited Q2 2026 Coterra details are from a Yahoo-transcribed call and the URL returned 429; no issuer transcript corroboration was accessible.", "correction": "Mark claims unverified pending direct Coterra filing/transcript; this access failure is not proof the claim is false.", "source_url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html"}], "residual_gaps": ["Cited secondary transcript returned HTTP 429; claim is not disproved, but issuer-source verification remains outstanding."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-102", "company": "Valero Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "source quality", "issue": "Valero Q3 issuer event page only verifies that a call took place; the detailed AI-readiness statements are from a secondary transcript.", "correction": "Attribute detailed comments to secondary transcription until verified against an issuer recording/transcript.", "source_url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-105", "company": "SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/"], "findings": [{"severity": "medium", "field": "stage", "issue": "SLB/NVIDIA source describes industrialization/partnership and product infrastructure, not measured upstream operator outcomes.", "correction": "Present as commercial product/infrastructure collaboration; no AI adoption result inferred.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-106", "company": "SLB / Tachyus", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update"], "findings": [{"severity": "medium", "field": "stage", "issue": "Tachyus is described in an acquisition agreement; announcement does not disclose consideration or establish realized SLB integration outcomes.", "correction": "Keep as announced acquisition, distinguish subsequent closing if separately sourced.", "source_url": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-107", "company": "SLB / S&P Global Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition"], "findings": [{"severity": "medium", "field": "stage", "issue": "The April acquisition announcement and September completion are separate events; AI-model co-development is a stated ambition.", "correction": "Distinguish acquired portfolio from future joint AI-model work.", "source_url": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-109", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton primary release and World Oil account call the work automated/autonomous and explicitly describe intelligent automation; neither establishes an AI/ML method.", "correction": "Classify as automation evidence, not operational AI, unless another supporting source explicitly identifies AI/ML. Retain the 17% stage-efficiency result as vendor-reported automation result.", "source_url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program"}], "residual_gaps": ["No explicit AI method was found in the supporting primary or secondary sources checked."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-110", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton’s release describes closed-loop, feedback-driven autonomous completion control and a domain algorithm; no source checked calls it AI or machine learning.", "correction": "Classify as closed-loop automation, not AI, unless other explicit AI evidence is found.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"}], "residual_gaps": ["Chevron’s own announcement URL returned 403; Halliburton primary source was reviewed and does not identify an AI/ML method."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-111", "company": "Halliburton / Sekal", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/node/35181/pdf", "https://ir.halliburton.com/node/36266/pdf"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "Sekal acquisition source supports drilling-automation capability, not necessarily an AI model.", "correction": "Classify as drilling automation; do not infer AI from autonomy.", "source_url": "https://ir.halliburton.com/node/36266/pdf"}], "residual_gaps": ["The cited sources do not clearly establish an AI/ML method; automation alone is not sufficient."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-114", "company": "NOV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization"], "findings": [{"severity": "medium", "field": "stage", "issue": "NOV/AIQ source describes planned or pilot edge-AI drilling collaboration; no scaled commercial deployment or results shown.", "correction": "Retain pilot stage and avoid productized/deployed language.", "source_url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-118", "company": "Liberty Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289"], "findings": [{"severity": "medium", "field": "integration", "issue": "Liberty states an LLM interface/orchestration layer for StimCommander; evidence does not show the model directly bypassing safety checks or independently controlling pumps.", "correction": "Describe natural-language command/configuration workflow with human/control-system governance; do not claim autonomous pump control by the LLM.", "source_url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-119", "company": "Snowflake", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Snowflake Energy Solutions is a data/product ecosystem launch, not a named operator AI deployment.", "correction": "Separate platform offer and partners from actual adoption.", "source_url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-120", "company": "Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release"], "findings": [{"severity": "medium", "field": "stage", "issue": "Atlas AI launch describes a commercial agent workbench; no company-specific deployment or outcome is inherent in the product release.", "correction": "Keep capability claims separate from customer proof.", "source_url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-121", "company": "Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and"], "findings": [{"severity": "medium", "field": "classification / deployment scope", "issue": "The customer example is Celanese, a chemicals manufacturer; its time-series anomaly detection use case is process-industry adjacent to oil and gas.", "correction": "Keep the AI use case but label the customer and example as chemicals/adjacent process industry; exclude it from O&G-operator adoption comparisons.", "source_url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and"}], "residual_gaps": ["The release describes a product/customer collaboration and target use case; realized Celanese deployment/results are not established."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-128", "company": "Seeq", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/"], "findings": [{"severity": "medium", "field": "stage", "issue": "Seeq Intelligence release establishes product capabilities, not deployment at a named O&G customer.", "correction": "Keep as commercial product launch absent customer evidence.", "source_url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-129", "company": "AVEVA", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/"], "findings": [{"severity": "medium", "field": "stage", "issue": "AVEVA source includes future capabilities and planned releases; it is a product roadmap, not proof those functions were shipped/deployed in May 2026.", "correction": "Use future tense for roadmap items and distinguish available integrations from planned agentic features.", "source_url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-132", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-133", "company": "OMV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [{"severity": "medium", "field": "stage", "issue": "OMV annual report says low-carbon field-development agentic AI work is being tested; no oil-and-gas deployment or quantified outcome.", "correction": "Keep as test-stage adjacent low-carbon planning.", "source_url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-134", "company": "OMV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [{"severity": "low", "field": "technology_data", "issue": "OMV’s report places iDEA among examples illustrating AI across the value chain, but calls the tool a knowledge-based system and gives no model details.", "correction": "Retain as issuer-framed AI use; state that the AI technique is undisclosed rather than declassifying for lack of architecture detail.", "source_url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"}], "residual_gaps": ["The reviewed annual report gives use for drilling programs, analogues and lessons but no technical description, deployment count or quantified result."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-135", "company": "OMV", "outcome": "supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html"], "findings": [], "residual_gaps": [], "partition": 2, "reviewer_outcome": "supported"}, {"id": "AI-141", "company": "Aramco", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data"], "findings": [{"severity": "medium", "field": "dates", "issue": "Aramco AI and Big Data page is undated; 2026-09-27 is access date only.", "correction": "Use “Undated; accessed 2026-09-27”; keep pilot timing to separately stated source dates.", "source_url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-142", "company": "PETRONAS", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf"], "findings": [{"severity": "high", "field": "source access", "issue": "PETRONAS integrated report PDF could not be fully opened (viewer error); excerpt/search evidence is insufficient to certify every ERMAI technical detail.", "correction": "Mark excerpt-only/unverified where assertions exceed exposed text; do not call full report reviewed.", "source_url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-143", "company": "PETRONAS", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/bm/node/1688"], "findings": [{"severity": "medium", "field": "stage", "issue": "TriCipta release supports venture/product launch but does not establish named customer production deployment or measured subsurface outcome.", "correction": "Separate announced AI venture and intended product suite from operator use.", "source_url": "https://www.petronas.com/bm/node/1688"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-146", "company": "Petrobras", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://petrobras.com.br/inovacao-e-tecnologia"], "findings": [{"severity": "high", "field": "source support", "issue": "Petrobras generic innovation page did not expose the named Smart Tocha deployment details on reopen.", "correction": "Specific workflow/date/metrics remain unverified until a direct Petrobras technical or operations source is located.", "source_url": "https://petrobras.com.br/inovacao-e-tecnologia"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-147", "company": "Ecopetrol", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES"], "findings": [{"severity": "medium", "field": "scope", "issue": "Ecopetrol 2025 report supports group AI/analytics activity but not all profile details about a specific upstream planning workflow.", "correction": "Do not attribute general R&D, digital or group results to AI or to upstream planning without direct wording.", "source_url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-148", "company": "Ecopetrol", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/"], "findings": [{"severity": "high", "field": "source support", "issue": "CADE announcement URL failed with cache miss, and the profile title alone does not establish that the decision platform uses AI.", "correction": "Mark inaccessible/not re-verified, not disproved; keep AI status unverified.", "source_url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-150", "company": "Pertamina", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf"], "findings": [{"severity": "high", "field": "source access", "issue": "PINTAR primary PDF returned request rejected during review.", "correction": "Mark as unverified on this pass; access failure is not a contradiction.", "source_url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf"}], "residual_gaps": ["Primary URL was attempted, but full content was not accessible or did not expose the named technical claim; not verified does not mean disproved."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-151", "company": "Pertamina", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness"], "findings": [{"severity": "medium", "field": "classification", "issue": "PGE G-Bionic source is geothermal operations, an adjacent energy business rather than oil-and-gas production.", "correction": "Label geothermal scope and avoid combining program-level efficiency gains as AI-specific.", "source_url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-157", "company": "Halliburton", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "Halliburton/Sekal source establishes automated on-bottom drilling, control integration and autonomy, without AI/ML identification.", "correction": "Classify as drilling automation; remove it from AI adoption counts absent explicit AI evidence.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom"}], "residual_gaps": ["No AI/ML method identified in the primary deployment report or cited product material."], "partition": 2, "reviewer_outcome": "corrected"}, {"id": "AI-158", "company": "NOV", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center"], "findings": [{"severity": "medium", "field": "AI basis", "issue": "NOV Egypt source primarily supports remote operations and operational performance; the reopened announcement is vendor authored and AI method/effect must not be assumed.", "correction": "Separate remote-operation achievement from any unproven AI analytics attribution.", "source_url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-166", "company": "Encino Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps"], "findings": [{"severity": "medium", "field": "workflow / metric attribution", "issue": "Microsoft source explicitly identifies a Copilot Studio production-data agent and AI Builder invoice extraction with confidence review. However, “hundreds of hours” saved is attributed to the broader Power Platform invoice app, not isolated to AI extraction.", "correction": "Retain AI classification for the named Copilot/AI Builder components; do not attribute all Power Apps process savings to AI. Keep agent rollout/planning status distinct from the deployed invoice app.", "source_url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps"}], "residual_gaps": ["Microsoft’s partner case gives no independent audit and does not disaggregate the administrative-time saving by AI versus workflow automation."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-168", "company": "Encino Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy"], "findings": [{"severity": "medium", "field": "attribution", "issue": "Ambyint case is supplier-reported and offers no direct operator confirmation in this audit.", "correction": "Retain vendor attribution for deployment and benefits; avoid presenting as independent verification.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-169", "company": "Preem", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization"], "findings": [{"severity": "medium", "field": "stage", "issue": "Imubit/Preem announcement says Optimizing Brain “will be deployed”; claimed emissions/fuel benefits are anticipated.", "correction": "Keep deployment prospective and metrics as expected, not achieved.", "source_url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization"}], "residual_gaps": ["Partner/supplier primary evidence only; no operator-side corroboration found in this source pass."], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-176", "company": "Targa Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925"], "findings": [{"severity": "medium", "field": "stage", "issue": "Targa job description shows hiring intent and target responsibilities only.", "correction": "Do not infer a live operational AI system from the role.", "source_url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-177", "company": "Aramco", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025"], "findings": [{"severity": "medium", "field": "stage", "issue": "LEAP remarks use future-tense capability language for MetaBrain for Plants; no facility rollout/results established by this source.", "correction": "Stage as introduced/developing capability, not a verified plant deployment.", "source_url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-180", "company": "Indian Oil Corporation (IndianOil)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html"], "findings": [{"severity": "medium", "field": "detail", "issue": "IndianOil annual report supports AI/ML predictive maintenance across refineries/operating locations, but gives no site, asset, performance metric or spend.", "correction": "Keep description broad and do not infer a named tool, site or return.", "source_url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-183", "company": "Reliance Industries Limited (RIL)", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf"], "findings": [{"severity": "medium", "field": "economics", "issue": "RIL chairman statement names the AI feedstock-blend tool but discloses no method, scale or measured benefit.", "correction": "Treat workflow as company-reported and economics/outcome as undisclosed.", "source_url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf"}], "residual_gaps": [], "partition": 2, "reviewer_outcome": "partially supported"}, {"id": "AI-001", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-002", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-003", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-004", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-005", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Chevron’s cited Project Kilby page could not be reopened; the direct URL returned 403.", "correction": "Retain prior verified project information only as previously reviewed; mark this supplemental source not reverified, not disproved. Its category remains AI-demand infrastructure, not AI adoption.", "source_url": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-006", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf"], "findings": [{"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence type as a management earnings-call transcript and retain exact speaker-attributed claims.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf"}, {"severity": "low", "field": "Supplemental source review", "issue": "Devon management links AI investment/platforms to production projects beginning to scale in Q1/Q2 2026, but the remarks do not explicitly say the discussed gas-lift trials were AI-driven. Keep gas-lift optimization and AI-platform work as related but not proven identical workflows; the $1B business optimization target is not AI-attributed.", "correction": "", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/"}, {"severity": "low", "field": "Supplemental source review", "issue": "Devon’s Q1 2026 transcript adds management detail: Smart Gas Lift uses AI models/physics-based calculation in a closed loop; 2025 pilot reported 2–3% uplift, and full Delaware Basin implementation exceeded 850 wells. Current uplift above pilot was not quantified.", "correction": "No contradiction to “secondary-sourced”; ensure pilot uplift and company-wide $1B business-optimization savings remain separate, with only the pilot’s 2–3% attributed to Smart Gas Lift.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-007", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf"], "findings": [{"severity": "medium", "field": "Supplemental source review", "issue": "R043's IADC archive listing verifies only the article title/byline; the body could not be reopened, and the title concerns post-merger AI/synergies rather than the imported description of well-design/completions-sequencing agents.", "correction": "Do not use R043 as fresh verification of the named well-design workflow. Keep earlier presentation evidence as not reverified this pass rather than disproved; reconcile the R043 title/summary mismatch if retaining that citation.", "source_url": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-010", "company": "Shell / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The JPT/SPE report confirms a signed strategic collaboration aimed at developing agentic AI and unified upstream workflows; it does not establish an operating deployment.", "correction": "Keep the profile at collaboration/development stage and attribute it to secondary editorial corroboration; avoid wording that suggests Shell assets already use the joint agentic system.", "source_url": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-019", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-020", "company": "Baker Hughes / EPAM", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Baker Hughes Cordant announcement URL returned 403 during direct reopen.", "correction": "Mark this supplemental URL not reverified; do not treat access failure as contradiction. Preserve any earlier review status separately from this pass.", "source_url": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-021", "company": "Baker Hughes / C3 AI", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.c3.ai/node/10916/pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "C3 AI FY2025 Form 10-K was reopened. It verifies the April 2025 renewal, co-selling/co-investment language, and that Baker Hughes is no longer exclusive reseller; separate May 28 issuer release adds term through June 2028, named deployed customers and internal-use applications. Benefits and customer scope remain vendor-reported/aggregate.", "correction": "", "source_url": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm"}, {"severity": "low", "field": "Supplemental source review", "issue": "The C3 AI FY2026 Form 10-K says Baker Hughes is no longer C3 AI’s exclusive reseller after April 2025 amendment; the filing still describes a strategic relationship.", "correction": "Retain the multi-year relationship/renewal as reported, but update reseller status to non-exclusive from April 2025 onward. Do not describe the relationship as terminated.", "source_url": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-025", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-026", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-029", "company": "TotalEnergies / Mistral AI", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-030", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-031", "company": "Repsol / Accenture", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-032", "company": "Repsol", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Repsol’s page says Copilot for Web reached ~25,000 employees while M365 Copilot licenses are progressively scaling; a 550-person, four-month pilot reported 121 minutes/week per participant. The page describes >60 use cases in process, not 60+ in production.", "correction": "Keep product reach distinct from M365 licensing and pilot results. Do not attribute the 121 minutes to reduced labor cost or use this page to support a “60+ in production/22 scaled” count.", "source_url": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-033", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-034", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-035", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-036", "company": "Equinor", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-044", "company": "Occidental", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "stage / scope", "issue": "Oxy management described current AI use in Gulf of America supply chain, asset integrity and reservoir characterization in the Q4 2024 transcript, and AI assistance in Permian/onshore subsurface work in the Q2 2025 call. The complex Gulf subsurface project was still being developed, with execution expected in one to two years.", "correction": "Replace the Q1-2026-only “no specific workflow” reading: keep current onshore/Gulf operational use distinct from the forward-looking complex Gulf project; attribute transcript claims to management and note no AI-specific outcome was reported.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}, {"severity": "low", "field": "Supplemental source review", "issue": "The Q4 2024 call names current Gulf of America AI use in supply-chain management, asset integrity, and reservoir characterization and an enterprise AI center of excellence. Keep Algeria seismic dataset as future-use potential; no results or implementation detail were disclosed.", "correction": "", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/"}, {"severity": "low", "field": "Supplemental source review", "issue": "Oxy management describes AI already helping subsurface work in Permian/onshore, while the complex Gulf of America project is being developed with execution expected in the following one to two years.", "correction": "Separate existing onshore assistance from the forward-looking Gulf project; do not describe Gulf AI project as live/deployed. Retain source as secondary transcript.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}], "residual_gaps": ["Transcript sources are third-party renderings of management remarks; exact implementation, scale and AI-attributed results are not disclosed."], "partition": 3, "reviewer_outcome": "partially supported; current onshore use and planned Gulf work are distinct"}, {"id": "AI-049", "company": "Aker BP / Cognite", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "World Oil is secondary coverage and repeats Aker BP/Cognite claims, including the >70% RCA-time reduction; it is not independent validation.", "correction": "Keep the result labeled partner/company-reported and unbenchmarked; the existing Cognite primary partner release remains the preferred supporting citation.", "source_url": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-052", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-053", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-054", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-055", "company": "Quantum Capital Group", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.quantumcap.com/digital-mission/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-057", "company": "PakEnergy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "Ownership source confirms Charlesbank as PakEnergy private-equity partner and documents Petrosight well-lifecycle software acquisition; it does not support AI in PakEnergy accounting workflows. This is corroborating context only, not a factual AI correction.", "correction": "", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-058", "company": "PakEnergy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The reopened acquisition release is ownership/deal context only and does not substantiate the PakCAPTURE ticket/receipt extraction workflow; retain the existing AI workflow on its own cited evidence.", "correction": "", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-062", "company": "Energy Transfer / CloudBurst", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-068", "company": "CNOOC subsidiary / SLB", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-074", "company": "Chevron", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c"], "findings": [{"severity": "medium", "field": "date / metric", "issue": "Issuer article is dated 2025-11-12 and supports >1 million files, 4 TB and seconds-fast retrieval; it does not quantify pre-drill studies falling from weeks to minutes.", "correction": "Use the exact Nov. 12, 2025 date and remove the weeks-to-minutes result unless another source supports it.", "source_url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas"}, {"severity": "medium", "field": "Supplemental source review", "issue": "Chevron's own article is dated 2025-11-12, not the supplied December month. It states >1 million files, 4 TB, and seconds-fast retrieval; it does not say studies fell from weeks to minutes.", "correction": "Date REF-013 to 2025-11-12 and remove 'weeks to minutes' as an achieved metric unless another source directly supports it. Keep the corpus size and retrieval workflow.", "source_url": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "supported with metric/date correction"}, {"id": "AI-075", "company": "Chevron", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-076", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "C3 AI’s June 4, 2026 release confirms existing Shell predictive maintenance on 13,000+ pieces of equipment, deployed on Azure, and says the new multi-year agreement will introduce agent-based root cause analysis/remediation. Do not describe the new agentic component as already deployed; “hundreds of millions” is C3 AI executive-attributed, not independently validated.", "correction": "", "source_url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-077", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-078", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "TotalEnergies’ Sep 26, 2025 announcement plans three-year scale-out of Cognite data/AI across all operated upstream assets. It describes scope and intended value chain but is an expansion/scale-up announcement, not proof that worldwide rollout was already complete.", "correction": "", "source_url": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-082", "company": "CNOOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-086", "company": "Devon Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-088", "company": "EOG Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence type as a management earnings-call transcript and retain exact speaker-attributed claims.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-089", "company": "EOG Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence type as a management earnings-call transcript and retain exact speaker-attributed claims.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-091", "company": "Occidental Petroleum", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf"], "findings": [{"severity": "high", "field": "Supplemental source review", "issue": "R052 confirms general management statements about AI's growing role in workforce efficiency, but it does not substantiate a $4B target or connect that target to AI. The Motley Fool URL is a third-party transcript; the official Oxy-hosted transcript corroborates the remarks but provides no AI-specific economics.", "correction": "Remove the $4B figure from the AI profile's Reported_metric and from the R052 disclosure finding; state that no AI-specific quantified outcome was disclosed. Limit the record to management's stated strategic direction and its generic workforce/process-efficiency context.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-096", "company": "Matador Resources", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/"}, {"severity": "low", "field": "source provenance", "issue": "Support includes an earnings-call transcript rendered by a third party (Yahoo/Motley Fool); the statements are attributed to company management, but the transcript is not an issuer-authored presentation or filing.", "correction": "Label the evidence as a third-party rendering of management’s earnings-call remarks. Preserve speaker attribution and distinguish pilot/workflow disclosure from independently validated results.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/"}, {"severity": "medium", "field": "workflow / stage", "issue": "The Q1 2026 call names operational uses: alerts intended to reduce downtime, real-time frac monitoring of pressures/water/sand/logistics, and AI-assisted lateral targeting. The source describes use but does not independently validate performance or establish a measured AI uplift.", "correction": "Remove the stale strategy-only characterization; describe the call-specific workflows while keeping scale, validation and attribution limits explicit.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}, {"severity": "low", "field": "Supplemental source review", "issue": "Q1 2026 transcript gives more operational detail than the inherited “vendor-led cautious integration strategy” title: daily real-time data monitored by control-room/field teams, frac water/sand pressures/volumes and logistics, and AI-assisted preferred lateral targeting. Reported 40M data points/day is company-stated scale, not an outcome metric.", "correction": "", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "Q1 transcript is a third-party rendering; operational outcomes and AI-attributed metrics remain management-reported and unvalidated."], "partition": 3, "reviewer_outcome": "partially supported; operational workflows named, AI-attributed results not established"}, {"id": "AI-099", "company": "Suncor Energy", "outcome": "corrected", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"], "findings": [{"severity": "high", "field": "AI basis / classification", "issue": "The reopened Suncor source describes GPS, wireless/LTE, perceptive technologies and an Autonomous Haulage System; it contains no AI/artificial-intelligence reference. It confirms automation deployment but does not substantiate classification as AI.", "correction": "Classify as conventional autonomy/automation context or exclude from AI-adoption counts unless an additional technical source documents an AI component. Keep autonomous fleet scale and human control-room oversight as supported facts.", "source_url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"}, {"severity": "low", "field": "Supplemental source review", "issue": "The International Mining article supports almost 120 current autonomous trucks at Base Plant and 2025 expansion, but neither AI/ML basis nor the 140-truck planned target is stated there.", "correction": "Keep autonomous haulage distinct from AI. If retaining 140 as a plan, source it separately; this secondary article supports near-120 current fleet, not that target.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}, {"severity": "high", "field": "Supplemental source review", "issue": "The Suncor page describes autonomous trucks using GPS, wireless/LTE and perception technologies, with control-room oversight. It contains no AI/artificial-intelligence reference. The page is undated, so neither a 2025 nor 2026 publication date can be inferred; 140 trucks is a future target with no dated year on the page.", "correction": "Replace the AI framing with 'Automation / AI not established'; retain the operating haulage evidence, nearly-120 snapshot, human control-room role and future 140 target without assigning a publication/year to the page. Remove the claim that AI optimizes driving conditions and unverified workforce job-loss claims from the disclosure.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof.", "AI classification is not supported by the primary page; workflow and fleet facts are supported as autonomous automation."], "partition": 3, "reviewer_outcome": "corrected"}, {"id": "AI-100", "company": "Phillips 66", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-101", "company": "Phillips 66", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-104", "company": "Kinder Morgan", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx"], "findings": [{"severity": "medium", "field": "Supplemental source review", "issue": "The profile's cited Flyscan URL is dated 2025-01-14, while Source_date is 2025-04-16, the date of the separate KMI Q1 earnings release.", "correction": "Set the initiative Source_date to 2025-01-14 and keep the profile's caveat that investment/partnering does not establish KMI deployment.", "source_url": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-112", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-113", "company": "Weatherford", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "https://www.weatherford.com/documents/sustainability/sustainability-report/"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The user-supplied GlobeNewswire link failed, but Weatherford's issuer-hosted copy of the October 1, 2025 release was reviewed and supports the portfolio-launch details.", "correction": "Retain launch facts with the Weatherford investor-relations copy as primary evidence; describe components as offered capabilities, not customer deployment or outcomes.", "source_url": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one supporting long document was reviewed at relevant excerpts only.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-115", "company": "Nabors / Corva", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "https://www.nabors.com/drilling-automation/predictive-drilling/"], "findings": [{"severity": "low", "field": "Supplemental source review", "issue": "The Apr 11, 2025 article describes intended integration of Nabors RigCLOUD edge/cloud/rig instrumentation with Corva AI analytics. It does not document deployed customer count, basin/rig scope, or realized gains; retain as partner/vendor offering/rollout stage.", "correction": "", "source_url": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-116", "company": "Helmerich & Payne", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "https://www.hpinc.com/technologies/flexrobotics"], "findings": [{"severity": "medium", "field": "classification / scope", "issue": "The earnings release confirms a FlexRobotics rig-floor deployment, but does not say AI was deployed with it. The separate H&P product page calls the base product automation and lists AI-enabled zone monitoring as an add-on camera feature.", "correction": "Classify the base workflow as robotics/automation; use Commercial AI product for the add-on capability only and state that the customer deployment of the AI add-on is unconfirmed.", "source_url": "https://www.hpinc.com/technologies/flexrobotics"}, {"severity": "low", "field": "Supplemental source review", "issue": "Issuer earnings release confirms FlexRobotics deployment on a Permian rig for a supermajor, but does not establish AI in the deployed robot. H&P’s separate product page calls FlexRobotics automation and lists AI zone monitoring as an add-on camera feature.", "correction": "Keep AI only for the explicitly described camera-based monitoring capability and qualify it as add-on/product capability; do not imply it was part of the customer deployment absent evidence. The deployed base robotics workflow is automation.", "source_url": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported; deployment is automation, AI add-on deployment unconfirmed"}, {"id": "AI-124", "company": "Collide", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-125", "company": "Collide", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html"], "findings": [{"severity": "medium", "field": "verification / source type", "issue": "At least one assigned transcript or secondary source is not an issuer-authored primary disclosure.", "correction": "Retain only claims directly supported by the opened transcript/article and label source as secondary/transcript review; seek issuer earnings deck/transcript for primary confirmation.", "source_url": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html"}, {"severity": "low", "field": "Supplemental source review", "issue": "May 7, 2026 article adds a concrete Winn Resources deployment: monthly Texas RRC W-10/G-10 submissions, data-gap visibility, and human review; Collide/Winn-reported processing time reduction is >95%, with a quoted 50-well filing completed in 20 minutes. Other listed workflows are future targets. Evidence is secondary-only in this supplemental source.", "correction": "", "source_url": "https://energycapitalhtx.com/collide-riggs-ai-platform"}], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "At least one record basis is a secondary/transcript rendering, not an issuer-authored primary source.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-126", "company": "Collide", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "https://collide.io/product"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-130", "company": "Emerson / AspenTech", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-131", "company": "Aker BP", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-138", "company": "Aker BP", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-140", "company": "Shell", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-145", "company": "CNOOC", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-149", "company": "YPF", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": [], "findings": [{"severity": "medium", "field": "verification", "issue": "Packet marks primary verification although a linked source was inaccessible in this re-review.", "correction": "Record that this source was not reverified in the current pass. A failed reopen does not invalidate the earlier documented source review or disprove the claim; retain prior verification provenance and describe this pass as not reverified.", "source_url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf"}], "residual_gaps": ["The original register/packet reported source review, but its only linked source was inaccessible during this re-open; this pass cannot independently reproduce the claim. Treat as not reverified, not disproved."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-153", "company": "Occidental", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-154", "company": "Suncor Energy", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-155", "company": "Petoro", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-160", "company": "Saipem", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-165", "company": "Baker Hughes", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-167", "company": "TotalEnergies", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/25505-totalenergies-agents"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-173", "company": "ONEOK", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://sustainability.oneok.com/innovation-spotlight", "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-174", "company": "Energy Transfer", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-175", "company": "Energy Transfer", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}, {"id": "AI-179", "company": "PETRONAS / AIngineer / Microsoft", "outcome": "partially supported", "checks": ["identity", "AI basis", "dates", "stage", "metrics", "attribution", "classification", "duplicates"], "source_urls_reviewed": ["https://www.petronas.com/bm/node/1668"], "findings": [], "residual_gaps": ["This review confirms only the specific claims surfaced in the linked source(s); broader interpretation and performance remain unverified unless the source reports them.", "Some linked baseline URLs were inaccessible on re-open; that is not evidence of disproof."], "partition": 3, "reviewer_outcome": "partially supported"}], "source_audits": [{"url": "https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/sustainability/group-reports/bp-sustainability-report-2024.pdf", "access": "excerpt reviewed", "title": "bp Sustainability Report 2024", "publication_date": "Undated; bp Sustainability Report 2024 reporting period", "claims_checked": ["Official indexed report text says bpx implemented AI driving-risk alerts across Permian operations in 2024 and attributes 89% reduction in distraction events and 75% reduction in tailgating between Dec 2023 and Dec 2024."], "notes": "Direct open attempted twice this audit and returned 403. Official bp PDF is indexed by the web search tool and returned the substantive “Advancing safety through AI” passage; this is excerpt-level evidence, not direct PDF full-text review. Metrics are company-reported and period is Dec 2023–Dec 2024.", "partition": 1}, {"url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "access": "full text reviewed", "title": "Digital technologies — ExxonMobil", "publication_date": "Undated; see individual source review", "claims_checked": ["ExxonMobil describes a proprietary AI drilling advisory system in use in Guyana that determines drilling parameters and enables closed-loop control without human intervention."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: ; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.thenationalnews.com/business/energy/2025/11/02/adnoc-signs-microsoft-deal-to-co-develop-and-deploy-ai-agents-across-its-operations/", "access": "full text reviewed", "title": "40,000+ employees AI-trained; 70,000 productivity hours/month", "publication_date": "2025-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "access": "full text reviewed", "title": "Fourth-quarter and full-year 2025 results — Digital highlights", "publication_date": "2026-01-23", "claims_checked": ["engineers to diagnose issues and optimize wells in minutes instead of days"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.prnewswire.com/news-releases/aiq-announces-340-million-contract-for-large-scale-deployment-of-agentic-ai-across-adnoc-operations-302400274.html", "access": "full text reviewed", "title": "AIQ wins $340M contract to deploy ENERGYai agents", "publication_date": "2025-03-12", "claims_checked": ["Reviewed during consolidation; metrics remain company or vendor reported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "access": "full text reviewed", "title": "ADNOC and SLB launch AI-powered production optimization", "publication_date": "2025-11-03", "claims_checked": ["Production System Optimization platform launched with initial deployment across eight fields.", "Connects office and field workflows so engineers diagnose issues and optimize wells in minutes rather than days; planned deployment across all onshore/offshore fields by 2027.", "ADNOC says complex tasks can be completed up to ten times faster; this is a company claim, not independently validated."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-03; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "access": "full text reviewed", "title": "ENERGYai trial boosts seismic interpretation accuracy 70%", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}, {"url": "https://www.computerweekly.com/news/366646979/ADNOC-shifts-AI-strategy-from-isolated-pilots-to-enterprise-wide-operations", "access": "full text reviewed", "title": "ADNOC moves AI from pilots to enterprise-wide operations", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.eni.com/en-IT/media/press-release/2026/07/pr-eni-opens-supercomputing-capabilities-innovation-ecosystem.html", "access": "full text reviewed", "title": "Eni opens its supercomputing capabilities to the innovation ecosystem", "publication_date": "2026-07-28", "claims_checked": ["Eni opened HPC infrastructure, expertise and tools to companies, startups and research centres to develop AI and digital innovation; HPC6 and HPC7 together exceed exascale-class performance."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.slb.com/newsroom/press-release/2025/aiq-slb-energyai-pr", "access": "full text reviewed", "title": "AIQ and SLB co-develop ENERGYai for subsurface", "publication_date": "2025-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.eni.com/content/dam/enicom/documents/eng/sustainability/2025/eni-for-2025-just-transition-eng-print.pdf", "access": "inaccessible", "title": "Eni for 2025 — Just Transition", "publication_date": "2026-04", "claims_checked": ["Eni reports around 300 AI use cases (+40% vs. 2024), around 20 AI agents; digital plant target ranges of 2–3% production enhancement and emissions reduction and drilling automation target of 35% less drilling time."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.offshore-mag.com/drilling-completion/news/55380289/exxonmobil-achieves-milestone-automated-well-placement-offshore-guyana", "access": "inaccessible", "title": "First fully autonomous AI-drilled well offshore Guyana", "publication_date": "2026-03", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.worldoil.com/news/2026/8/13/exxonmobil-identifies-four-guyana-exploration-opportunities-using-ai/", "access": "full text reviewed", "title": "AI identifies four new Guyana exploration opportunities", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.constellationr.com/insights/news/exxonmobil-outline-process-automation-ai-efforts", "access": "full text reviewed", "title": "AI plus process mining behind $15B cost-cut push", "publication_date": "2024-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "access": "inaccessible", "title": "Santos 2026 Investor Briefing Day", "publication_date": "2026-05", "claims_checked": ["Santos says technology program applies digital and AI capability across integrated value chain, including cost reduction and production uplift across drilling, completions, projects and facilities; specific quantified AI-only impacts not isolated."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://stockanalysis.com/stocks/xom/transcripts/361820-q3-2025/", "access": "full text reviewed", "title": "Q3 2025 earnings call", "publication_date": "2025-10-31", "claims_checked": ["We’re deploying it downhole in the work that we’re doing in the Permian."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://stockanalysis.com/stocks/xom/transcripts/395149-q4-2025/", "access": "full text reviewed", "title": "Q4 2025 earnings call", "publication_date": "2026-01-30", "claims_checked": ["the ability to apply AI at scale much more easily"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution", "access": "full text reviewed", "title": "ADNOC and AIQ Successfully Complete Trial Phase of Agentic AI Solution", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}, {"url": "https://www.eni.com/content/dam/enicom/documents/eng/investor/presentations/2026/2026-capital-markets-update/2026-capital-markets-update.pdf", "access": "full text reviewed", "title": "2026 Capital Markets Update", "publication_date": "2026", "claims_checked": ["ENI AI | use cases: around 300"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://www.eni.com/en-IT/media/press-release/2025/07/pr-khazna-data-centers-eni-partner-develop-i-data-center-campus.html", "access": "full text reviewed", "title": "500 MW AI Data Center Campus near Milan", "publication_date": "2025-07-11", "claims_checked": ["total IT capacity of 500 MW"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://www.cnpc.com.cn/en/2024enbyfgrme/202508/0e28e14e761748e589be91257345869e/files/3d40e35a2b0f4eddbb6580003ad94579.pdf", "access": "inaccessible", "title": "2024 Annual Report", "publication_date": "2025-08-01", "claims_checked": ["CNPC reports launching a 70-billion-parameter Kunlun model in 2024 with 43 professional/general scenarios for petroleum.", "Report names a joint development group of CNPC, China Mobile, Huawei and iFLYTEK; platform included eight trained large models, 18 application scenarios, specialized datasets, AI middleware and computing environment."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-08-01; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://news.cnpc.com.cn/system/2026/05/29/030194338.shtml", "access": "excerpt reviewed", "title": "中国石油昆仑大模型迈入主动智能新阶段", "publication_date": "2026-05-29", "claims_checked": ["CNPC says Kunlun reached 152 deployed scenarios across the energy and chemical value chain in May 2026.", "Specific example: crude purchase evaluation with processing plans and risk advice in five minutes versus hours or days of manual analysis, as reported by a refinery manager."], "notes": "Exact direct URL failed due encoding/internal error. Official CNPC portal search-index result exposed the article text and was reviewed as an excerpt, not direct full-text. Use official 2025 annual report as additional accessible CNPC source for scale/context; preserve this URL’s direct-fetch limitation.", "partition": 1}, {"url": "https://jhof.sinopec.com/jhof/news/mtjj/2026/6/I1519033108454703105.shtml", "access": "inaccessible", "title": "〖中国石化报〗多场景推进人工智能技术应用", "publication_date": "2025-05-28", "claims_checked": ["Jianghan Oilfield describes AI for shale-gas digital-twin reservoir updates, drilling risk alerts, fracture optimization, video safety monitoring, field inspection and work assistance.", "Reported examples include fracture simulation accuracy above 80%, video safety inspection efficiency up over 50% and missed-violation detection below 10%.", "A liquid-loading warning use case had more than 800 well instances, parsed 150,000 records/day, and company-reported warning accuracy above 95%; treatment efficiency above 90%."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-and-aiq-successfully-complete-trial-phase-of-agentic-ai-solution/", "access": "full text reviewed", "title": "ENERGYai proof of concept", "publication_date": "2025-01-16", "claims_checked": ["90-day proof-of-concept combined a 70-billion-parameter LLM with ADNOC proprietary operational data; reported 70% improvement in major seismic interpretation aspects; next production version was planned for H1 2025, so not evidence by itself that scaled deployment occurred."], "notes": "Exact URL reopened directly on 2026-09-27; full primary press-release text reviewed. Distinguish the completed 90-day PoC from the subsequent H1 2025 operational/test-deployment plan.", "partition": 1}, {"url": "https://www.eni.com/en-IT/actions/energy-transition-technologies/supercomputing-artificial-intelligence.html", "access": "full text reviewed", "title": "HPC6 supercomputer and company-wide responsible-AI training", "publication_date": "2025-06", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "Q2 2026 earnings call transcript", "publication_date": "2026-08-10", "claims_checked": ["We’re tackling it first on artificial lift"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results — Digital highlights", "publication_date": "2026-07-24", "claims_checked": ["Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy).", "Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs.", "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-24; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://tenders.ongc.co.in/o/tender/dowlnload/ShowNewsTickerPdf?newsID=207", "access": "full text reviewed", "title": "ONGC tender for automated gas-lift system", "publication_date": "2025-11-03", "claims_checked": ["Tender technical scope requires AI/ML algorithms to analyze well data and optimize gas-lift injection in real time, with predictive maintenance based on valve response and pressure trends.", "SCADA and existing wellhead control panel integration are specified. Procurement requirements show intended scope, not completed deployment."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-03; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "access": "full text reviewed", "title": "ADNOC, Masdar, XRG and Microsoft to Advance AI for Energy and Energy for AI", "publication_date": "2025-11-02", "claims_checked": ["ADNOC reports enterprise Copilot rollout in 2023, >40,000 employees completing AI training, utilization above 90%, and >70,000 monthly productivity hours; no training causal attribution stated.", "ADNOC and Microsoft announced plans to co-develop and deploy AI agents across ADNOC’s value chain; no particular agent or deployment was named.", "The wider release separately includes Masdar/XRG energy and infrastructure solutions for Microsoft AI/data-center growth."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-02; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://reports.adnoc.ae/e/sustainability-report-2025/eJFjvFKd42", "access": "full text reviewed", "title": "ADNOC Sustainability Report 2025", "publication_date": "2026", "claims_checked": ["In 2025, ADNOC says its Introduction to Agentic AI program engaged >18,000 participants and resulted in >10,000 agents created across operations."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.quorumsoftware.com/about/press-releases/quorum-software-acquires-streamba-to-extend-leadership-as-the-ai-platform-for-the-energy-value-chain/", "access": "full text reviewed", "title": "Quorum acquires Streamba", "publication_date": "2026-06-25", "claims_checked": ["Streamba AI-native supply chain platform and Visibility, Orchestration and Response system unify logistics data, workflows and teams.", "Quorum says agentic capabilities can identify disruption and propose or execute plan changes with human oversight."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-06-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.cnpc.com.cn/en/nr2025/202506/e9b44b36040f452796711916fbf49135.shtml", "access": "inaccessible", "title": "CNPC releases 300B-parameter Kunlun energy model", "publication_date": "2025-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.kockw.com/sites/EN/Periodicals/Digest%20October%20-%20December%202025.pdf", "access": "full text reviewed", "title": "The Kuwait Digest 15 (October–December 2025)", "publication_date": "2025-Q4", "claims_checked": ["KOC says G Agent uses AI for drilling-rig resource management, systems integration and real-time data analysis; it also describes an agentic drilling-rig scheduling project.", "KOC names Microsoft and Halliburton partnership and says Ghaia AI manages technology programs/training at the Center."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-Q4; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.franciscopartners.com/investments/quorum-software", "access": "full text reviewed", "title": "Quorum Software investment", "publication_date": "Undated; accessed 2026-09-26", "claims_checked": ["Private Buy Out"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://www.energynewsbulletin.net/operations/news-articles/4516741/santos-expands-ai-driven-production-optimisation-xecta", "access": "excerpt reviewed", "title": "Xecta AI digital twin expanded to East Australia and PNG", "publication_date": "2025-07", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.cnrl.com/content/uploads/2026/03/CNQ_2025-AIF-March-25-2026.pdf", "access": "full text reviewed", "title": "Canadian Natural 2025 Annual Information Form", "publication_date": "2026-03-25", "claims_checked": ["AI is used to expand employee knowledge access, deliver data-driven insight and improve efficiency via incremental, cost-justified deployment; company will not rely on it for time-sensitive/production-critical processes until controls meet standards; employees remain involved/accountable; board risk oversight described."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-03-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://kuwaittimes.com/article/33109/kuwait/other-news/koc-delivers-first-ai-success-with-g-agent-drilling-project/", "access": "full text reviewed", "title": "AI Innovation Center delivers G Agent drilling win", "publication_date": "2025-09", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.cnrl.com/sustainability/sustainability-reporting/governance-and-risk-management/", "access": "full text reviewed", "title": "Governance and Risk Management", "publication_date": "Undated; see individual source review", "claims_checked": ["Corroborates program-level AI scope; refers readers to AIF for policy specifics."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: ; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter", "access": "full text reviewed", "title": "Q3 2024 earnings call", "publication_date": "2024-10-29", "claims_checked": ["digital now with Palantir and Infosys that will help the refining system over time"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://palantir2020ipo.q4web.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/default.aspx", "access": "full text reviewed", "title": "Five-year strategic relationship with new AI capabilities", "publication_date": "2024-09-09", "claims_checked": ["New AI Capabilities"], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date 2026-09-26 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/08/10/diamondback-energy-fang-q2-2026-earnings-call/", "access": "inaccessible", "title": "Diamondback Q2 2026 Earnings Call Transcript", "publication_date": "2026-08-10", "claims_checked": ["Existing record links discussion of lift optimization and downtime to transcript; use case and outcomes remain unverified by issuer source."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-08-10; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "access": "full text reviewed", "title": "Agreement in principle for Homer City gas supply", "publication_date": "2025-07", "claims_checked": ["Project announcement describes EQT agreement in principle for gas supply to redevelopment targeting AI/HPC campus; not an executed definitive supply contract."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-07; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investor.williams.com/static-files/f614f046-cccf-4f74-91e9-5202008eeafc", "access": "full text reviewed", "title": "Second-quarter 2026 earnings presentation", "publication_date": "2026-08-04", "claims_checked": ["Project details include gas supply, pipeline and generation; presented as power infrastructure for data-center load."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-08-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "access": "full text reviewed", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "publication_date": "2026-01-23", "claims_checked": ["Q4 release also records the bp partnership update.", "Q4 disclosure: bp remote drilling, OQ Neuro/PowerDrive results, or KOC AI seismic-processing contract, as stated in profile."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/05/05/diamondback-fang-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "ML and AI in field ops cut downtime post-Endeavor", "publication_date": "2026-05", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.slb.com/zh-cn/resource-library/case-study-with-navigation/di/slb-digital-solutions-save-1500-personnel-on-board-days", "access": "full text reviewed", "title": "SLB digital solutions save 1,500 personnel-on-board days and boost efficiency", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["Performance Live with DrillOps and Neuro reduced offshore well-placement crew size 66% and saved >1,500 POB days across several Western Asia platforms.", "Case attributes benefits to remote operations and autonomous directional drilling; exact operator not identified in the case page."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: Undated; accessed 2026-09-27; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://static.conocophillips.com/files/resources/conocophillips-annual-meeting-2026-transcript.pdf", "access": "full text reviewed", "title": "CEO: embedding AI into core processes at scale", "publication_date": "2026", "claims_checked": ["CEO described focus on embedding AI into core processes to reduce manual work, accelerate decisions and improve consistency under enterprise governance/security/responsible-use framework; no named application or outcome disclosed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2025/11/06/ovintiv-ovv-q3-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "In-house AI across production, drilling and completions", "publication_date": "2025-11-06", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-11-06; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/02/24/ovintiv-ovv-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "DRIVE Center AI optimizes rigs from day one", "publication_date": "2026-02-24", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed.", "Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-24; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/05/12/ovintiv-ovv-q1-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "AI tunes 70-odd frac design inputs in real time", "publication_date": "2026-05-12", "claims_checked": ["Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed.", "Third-party transcript supports disclosed Ovintiv AI use case detail; primary issuer recording/transcript not reviewed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-12; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/eqt-eqt-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "AI tools to target service rebids and procurement", "publication_date": "2026-02-18", "claims_checked": ["Management said AI tools would help target service rebids and procurement, expecting low-single-digit procurement cost reductions; not isolated to AI and no method, baseline or achieved value disclosed."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-18; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.sec.gov/Archives/edgar/data/889900/000088990026000010/a2025q4earningsreleaseex991.htm", "access": "full text reviewed", "title": "Patterson-UTI Energy Reports Financial Results for the Quarter Ended December 31, 2025", "publication_date": "2026-02-04", "claims_checked": ["Patterson-UTI said nearly all rigs were equipped with proprietary Cortex automation applications by year-end 2025.", "Management cited continued development of automation software and demand for performance-enhancing products; this is scale of automation suite, not AI-specific model usage."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-02-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/chord-energy-chrd-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "AI rod-pump optimization arrests base decline", "publication_date": "2026-05-06", "claims_checked": ["Transcript attributes improved legacy-well base production to AI pump tuning along with deeper pump placement and workovers; no AI-only effect is isolated."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-06; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.precisiondrilling.com/wp-content/uploads/2025/04/2024-Precision-Drilling-ESG-Report.pdf", "access": "inaccessible", "title": "Precision Drilling 2024 ESG Report", "publication_date": "2025-04", "claims_checked": ["Precision reports 78 rigs with Alpha digital technology at end-2024.", "Alpha suite combines digital automation and analytics with AI algorithms for drilling performance pattern recognition, real-time adjustments and predictive maintenance; report does not break out AI-specific results."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/08/12/chord-energy-chrd-q2-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "AI rod-pump program deployed field-wide", "publication_date": "2026-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.sec.gov/Archives/edgar/data/1017413/000101741326000018/cnq-20251231.htm", "access": "inaccessible", "title": "Formal AI policy: no AI in production-critical processes", "publication_date": "2026-03", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm", "access": "inaccessible", "title": "Transocean Ltd. 2025 Form 10-K", "publication_date": "2026-02-19", "claims_checked": ["Transocean explicitly says AI is not currently used in offshore drilling operations; it has introduced AI tools for high-volume office/administrative tasks and is exploring productivity use cases.", "Company describes AI governance committee; fleet automation/robotics systems are separate from AI."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-02-19; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.fool.com/earnings/call-transcripts/2025/08/05/williams-wmb-q2-2025-earnings-call-transcript/", "access": "inaccessible", "title": "New CIO; AI model out-trades 10-year gas trader", "publication_date": "2025-08", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-masdar-xrg-and-microsoft-to-advance-ai-for-energy-and-energy-for-ai", "access": "full text reviewed", "title": "ADNOC-Microsoft deal to co-develop autonomous ops agents", "publication_date": "2025-11", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/static-files/35c12412-e139-4d6d-b0fe-9e23b54028f1", "access": "full text reviewed", "title": "SLB Awarded Seven-Year Contract Under Kuwait Oil Company’s Ahmadi Innovation Valley Initiative", "publication_date": "2026-06-30", "claims_checked": ["KOC awarded SLB a seven-year AIV contract to evaluate/test/deploy technologies; scope spans nearly 100 projects across AI, IIoT, production optimization, reservoirs, water and energy transition.", "A dedicated facility is planned, with construction expected to start 2026 and opening in 2028; the announcement names no specific AI workflow or outcome."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-06-30; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://novilabs.com/news/novi-labs-unveils-novi-intelligence-a-breakthrough-in-energy-research-transparency-and-accuracy/", "access": "full text reviewed", "title": "Novi Labs unveils Novi Intelligence", "publication_date": "2025-09-10", "claims_checked": ["Novi Intelligence joins operator-sourced data, machine learning and forecasts with proprietary metrics such as rock quality, completion intensity, depletion and well density.", "Product exposes underlying data for client validation; accuracy claims are vendor claims."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-09-10; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://novilabs.com/news/novi-labs-acquires-turing-analytics-to-establish-a-unified-north-american-platform-for-upstream-investment-decisions/", "access": "full text reviewed", "title": "Novi Labs acquires Turing Analytics", "publication_date": "2026-04-29", "claims_checked": ["Novi acquired Turing Analytics/EVA to combine US and Canadian upstream datasets and analytics for investment evaluation.", "Novi says Turing clients represent over 50% of Canadian oil and gas production; company-reported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-04-29; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://appliedcomputing.com/press-release-28052025", "access": "full text reviewed", "title": "Applied Computing raises £9m seed to build Orbital AI", "publication_date": "2025-05-28", "claims_checked": ["Applied Computing raised £9m seed funding for Orbital, a process-industry model using language, time-series, physics and chemical engineering models.", "Company says it is on edge, can work on sensitive process data, and reports one large refinery live.", "Cost savings, energy reduction and payback cited are company claims/potential, not independently supported."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.sec.gov/Archives/edgar/data/889900/000088990025000053/a2025q2earningsreleaseex991.htm", "access": "full text reviewed", "title": "Cortex ML autodriller and REX monitoring drive rig margins", "publication_date": "2025-07-30", "claims_checked": ["Cortex platform enables a machine-learning autodriller; REX is a cloud-based early-alert field monitoring system.", "Company cites increased demand and revenue from drilling automation technologies, but no separate AI return metric."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-07-30; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investor.williams.com/static-files/a693248d-419a-44f7-852c-854899ab2787", "access": "full text reviewed", "title": "Williams Q2 2025 earnings call transcript", "publication_date": "2025-08-05", "claims_checked": ["Sequent gas-marketing AI anecdote; CEO describes one-month DFW test, model score 96% vs trader 94% against a 100% hindsight benchmark; narrow test only."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investor.williams.com/events/event-details/williams-second-quarter-earnings-2025", "access": "full text reviewed", "title": "Williams Q2 2025 earnings event page", "publication_date": "2025-08-05", "claims_checked": ["Confirms issuer-hosted earnings materials including transcript and presentation."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.tcenergy.com/siteassets/pdfs/sustainability/sustainability-report/2025/tce-2025-ros.pdf", "access": "inaccessible", "title": "TC Energy 2025 Report on Sustainability", "publication_date": "2025", "claims_checked": ["Autonomous Pipeline analytics and machine learning for commercial daily/weekly asset strategies and transport/PAL opportunities; RAMP described as advanced algorithms only."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.tcenergy.com/siteassets/pdfs/investors/notice-and-access/2026/tce-2026-management-information-circular.pdf", "access": "full text reviewed", "title": "TC Energy 2026 Management Information Circular", "publication_date": "2026-03-05", "claims_checked": ["AI governance, training, board oversight; governance does not evidence operating deployment."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-03-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://novilabs.com/news/novi-labs-raises-35-million-from-invictus-growth-partners/", "access": "full text reviewed", "title": "Novi Labs raises $35M for AI-first energy analytics", "publication_date": "2025-06-25", "claims_checked": ["Invictus Growth Partners invested $35 million to support Novi product innovation, expansion and acquisitions.", "Novi describes platform combining machine learning, proprietary datasets and reservoir science for upstream capital decisions."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-06-25; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://aws.amazon.com/solutions/case-studies/tcenergy-bedrock/", "access": "full text reviewed", "title": "TC Energy Pipeline IQ on AWS", "publication_date": "undated", "claims_checked": ["PIPER reviews engineering documentation/material test reports against weld maps and pressure-test docs; SCOUT discovery; >100 users, 85% review-time reduction, expected $18.7m five-year maintenance-capital avoidance. Page is undated, date/start of deployment not established."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: undated; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://novilabs.com/news/novi-labs-acquires-rbn-energy/", "access": "full text reviewed", "title": "Novi acquires Turing Analytics and RBN Energy", "publication_date": "2026-05-28", "claims_checked": ["RBN acquisition combines Novi upstream AI forecasts with commodity, midstream and demand intelligence.", "Decision context includes basis, transport capacity, commodity flows and end-market pricing."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-05-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.enbridge.com/~/media/Enb/Documents/Reports/Sustainability-Report-2025/Enbridge_SR_2025.pdf?hash=56D227E5C9F3A8358E11182DC558E83B&rev=12caefdeac3a4219be884c2318636c31", "access": "inaccessible", "title": "Enbridge 2025 Sustainability Report", "publication_date": "2025", "claims_checked": ["Describes developing machine-learning prioritization of potentially injurious deformation features from in-line inspection and field data; engineers assess and mitigate priority features."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.enbridge.com/reports/2026-liquids-pipelines-customer-handbook/operational-safety-and-reliability", "access": "full text reviewed", "title": "Enbridge 2026 Liquids Pipelines Customer Handbook — Operational Safety and Reliability", "publication_date": "2026", "claims_checked": ["States ML predictive analysis in complex pipeline defect assessment to prioritize repairs and allocate resources; no quantified outcome."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.enbridge.com/~/media/Enb/Documents/Work-with-Enbridge/ENB_Artificial_Intelligence_Policy.pdf?hash=3CBE0387A12B003EF26811A7DE7B1A40&rev=f55179092d224cbc94b484c1215de152", "access": "full text reviewed", "title": "Enbridge 2026 AI Policy", "publication_date": "2026", "claims_checked": ["Company policy describes AI governance and general AI uses; does not by itself date or quantify named operating workflow."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://techcrunch.com/2026/07/15/applied-computing-wants-to-give-oil-and-gas-operators-an-ai-model-for-the-entire-plant/", "access": "full text reviewed", "title": "$20M Series A for Orbital O&G foundation model", "publication_date": "2026-07", "claims_checked": ["User-supplied reference preserved. Claims below require source review."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: none recorded; baseline accessed date Not independently accessed is preserved separately.", "partition": 1}, {"url": "https://investor.williams.com/news-releases/news-release-details/williams-delivers-strong-second-quarter-2025-results-and-raises", "access": "full text reviewed", "title": "Williams Q2 2025 results", "publication_date": "2025-08-04", "claims_checked": ["Describes $1.6bn Socrates power innovation project to serve growing AI demand; infrastructure, not internal AI."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-08-04; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://corporate.exxonmobil.com/locations/india/newsroom/2025/turning-artificial-intelligence-into-business-intelligence", "access": "full text reviewed", "title": "ExxonMobil: A to B, Turning Artificial Intelligence into Business Intelligence", "publication_date": "2025-10-23", "claims_checked": ["ExxonMobil India describes using AI with production staff to monitor equipment in real time and identify current or prospective bottlenecks affecting production.", "Engineers and asset operators use domain knowledge to investigate and remove bottlenecks; no basin/site counts or metrics."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2025-10-23; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://xecta.com/news-resources/", "access": "excerpt reviewed", "title": "Xecta news and resources index", "publication_date": "June/July 2025 listing; exact publication date unavailable", "claims_checked": ["Xecta index identifies Santos integrated production-system model/digital oilfield across Eastern Australia and PNG.", "Underlying dated technical release did not open; current confirmation limited to vendor page index."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: June/July 2025 listing; exact publication date unavailable; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.bharatpetroleum.in/bharat-petroleum/pdf/annual-report-2024-25.pdf", "access": "inaccessible", "title": "Annual Report 2024-25", "publication_date": "2025 (FY2024-25)", "claims_checked": ["AI predictive analytics for asset reliability/performance was initiated across Mumbai, Kochi and Bina refineries, extending a successful Kochi trial; cloud predictive alerts target failures/degradation in identified critical equipment.", "AI video algorithms detect non-compliance and safety hazards from field video; Ultra Critical Video Analytics was implemented across all three refineries and integrated with the in-house Work Permit System."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025 (FY2024-25); baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.hindustanpetroleum.com/NewsroomDetails/449", "access": "full text reviewed", "title": "HPCL Commissions Residue Upgradation Facility with World’s First LC-Max Unit at Visakh Refinery", "publication_date": "2026-01-05", "claims_checked": ["At commissioning of the Visakh residue-upgradation facility, HPCL said its LC-Max Digital Suite integrates real-time monitoring, predictive analytics using proprietary thermodynamic models and AI-driven optimization.", "The source reports RUF physical-process conversion/yield figures separately; it does not attribute those outcomes to AI."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-01-05; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.pttgcgroup.com/en/newsroom/news/1730/gc-advances-holistic-optimization-through-digital-technology-enhancing-end-to-end-operational-efficiency-and-strengthening-organizational-competitiveness", "access": "full text reviewed", "title": "GC Advances Holistic Optimization Through Digital Technology", "publication_date": "2026-07-20", "claims_checked": ["GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale.", "GC described AI and digital plant models for real-time complex-wide optimization; AI/ML condition analysis and AI-vision/robotics inspection; HERO applies AI plus process simulation to heat integration, claimed at industrial scale."], "notes": "Direct URL reopened this audit; specific AI/use-case terms were searched when relevant. This-run page content is accessible, but source claims remain company/vendor-reported where applicable. Prior source review dates: 2026-07-20; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://rosneft.ru/press/news/item/222577", "access": "inaccessible", "title": "Rosneft improves well research methods (EchoTools)", "publication_date": "2025-07-17", "claims_checked": ["AI interprets fluid level and sound speed in the well annulus; the derived values support bottomhole-pressure calculation and selection of an operating mode.", "Rosneft says the software processes over 10,000 well indicators in about two minutes versus over three days manually, and recommendations yield an average additional 1 tonne/day per well."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-07-17; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://www.rosneft.ru/press/news/item/224545/", "access": "inaccessible", "title": "Rosneft increases production through innovative software (EchoTools update)", "publication_date": "2026-08-20", "claims_checked": ["Rosneft reports that in 1H 2026 EchoTools analyzed information from more than 2,000 wells and generated a ranking of over 500 candidate wells for operating-mode optimization.", "This is a later rollout update to the July 2025 named workflow, not its launch date."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2026-08-20; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://rosneft.ru/press/news/item/222604/", "access": "inaccessible", "title": "Rosneft introduces new technologies using artificial intelligence", "publication_date": "2025-07-28", "claims_checked": ["Company-developed LLM-based virtual assistant is used for text search, verification and formatting tasks, with applications from geology to field-development design.", "Rosneft claims 70-80% lower routine workload for these tasks; no independent validation or financial return is disclosed."], "notes": "Direct open attempted this audit and failed (fetch error, timeout, access denied or decoding/size issue); a primary alternate was sought when available. Do not treat legacy excerpt/full-text review as this-run access. Prior source review dates: 2025-07-28; baseline accessed date 2026-09-27 is preserved separately.", "partition": 1}, {"url": "https://investorcenter.slb.com/static-files/377e2aaf-38e0-451d-a3c3-84818ac2dc5f", "access": "full text reviewed", "title": "Q4 2025 earnings call", "publication_date": "2026-01-23", "claims_checked": ["AI-008 source claim(s) checked against cited profile", "AI-009 source claim(s) checked against cited profile"], "notes": "SLB Q4 2025 earnings call transcript opened; partner/product references and broad digital ARR checked.", "partition": 2}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "access": "full text reviewed", "title": "Fourth-quarter and full-year 2025 results — Digital highlights", "publication_date": "2026-01-23", "claims_checked": ["AI-013 source claim(s) checked against cited profile", "AI-015 source claim(s) checked against cited profile"], "notes": "SLB Q4 results page opened; broad digital business results distinguishable from AI-specific economics.", "partition": 2}, {"url": "https://ir.halliburton.com/static-files/38b04c8d-e373-49ab-b3e5-f1d13297bd80", "access": "full text reviewed", "title": "Q1 2025 investor presentation", "publication_date": "2025-Q1", "claims_checked": ["AI-016 source claim(s) checked against cited profile", "AI-017 source claim(s) checked against cited profile"], "notes": "Halliburton Q1 2025 investor presentation opened; LOGIX and DecisionSpace product portfolio statements reviewed.", "partition": 2}, {"url": "https://www.halliburton.com/en/about-us/press-release/halliburton-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results", "publication_date": "2026-07-21", "claims_checked": ["AI-018 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks", "access": "full text reviewed", "title": "CEO opening remarks: FY2025 results", "publication_date": "2026-03-10", "claims_checked": ["AI-037: $5.3bn total 2025 TRV; CEO says over half enabled by AI; mixed-technology components are not an AI-only breakdown."], "notes": "Primary Aramco FY2025 remarks/presentation reviewed. CEO remarks explicitly state total TRV was $5.3bn, with over half enabled by AI; this is high-level company attribution, not independently validated project-level value.", "partition": 2}, {"url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "access": "full text reviewed", "title": "AI and supercomputing digital transformation", "publication_date": "2025-11-24", "claims_checked": ["AI-038 source claim(s) checked against cited profile", "AI-039 source claim(s) checked against cited profile"], "notes": "Aramco AI/supercomputing page opened; broad AI portfolio claims reviewed; no single workflow results certified from it.", "partition": 2}, {"url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "access": "full text reviewed", "title": "Artificial Intelligence at Woodside", "publication_date": "updated 2026-05-27; page undated", "claims_checked": ["AI-041 source claim(s) checked against cited profile", "AI-042 source claim(s) checked against cited profile", "AI-043 source claim(s) checked against cited profile"], "notes": "Woodside current AI page opened; Lumina and Startup Advisor claims reviewed, but date/first-deployment absent.", "partition": 2}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "access": "full text reviewed", "title": "Second-quarter 2026 results — Digital highlights", "publication_date": "2026-07-24", "claims_checked": ["AI-046 source claim(s) checked against cited profile", "AI-047 source claim(s) checked against cited profile", "AI-048 source claim(s) checked against cited profile", "AI-067 source claim(s) checked against cited profile"], "notes": "SLB Q2 2026 results opened; PTTEP/Cognite/OptiFlow/Tela offshore pilot, ONGC DrillOps and CNOOC terminal automation statements checked as partner claims.", "partition": 2}, {"url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "access": "full text reviewed", "title": "TIRAI myPROdata chatbot launch", "publication_date": "2025-11-03", "claims_checked": ["AI-050 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.blackstone.com/news/press/blackstone-announces-agreement-to-acquire-enverus/", "access": "full text reviewed", "title": "Agreement to acquire Enverus", "publication_date": "2025-08-06", "claims_checked": ["AI-051 source claim(s) checked against cited profile"], "notes": "Blackstone agreement announcement opened; Aug 2025 definitive agreement and expected transaction timing checked.", "partition": 2}, {"url": "https://media.kkr.com/news-details?news_id=05131f1d-9cac-4cc4-afb2-19a83bd4cee3", "access": "full text reviewed", "title": "Texas hyperscale data-center campus", "publication_date": "2025-07-30", "claims_checked": ["AI-059 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.sec.gov/Archives/edgar/data/1475260/000147526026000008/a2025managementsdiscussion.htm", "access": "full text reviewed", "title": "2025 Management’s Discussion and Analysis", "publication_date": "2026", "claims_checked": ["AI-061 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1103-slb-tela-ai", "access": "full text reviewed", "title": "Tela agentic AI launch", "publication_date": "2025-11-03", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "SLB Tela launch release opened; launch date and product claims checked.", "partition": 2}, {"url": "https://www.blackstone.com/wp-content/uploads/sites/2/2026/04/Blackstone1Q26EarningsPressRelease.pdf", "access": "full text reviewed", "title": "First-quarter 2026 earnings results", "publication_date": "2026-04-23", "claims_checked": ["AI-051 source claim(s) checked against cited profile"], "notes": "Blackstone Q1 2026 earnings release opened; transaction and reported capital context checked; no Enverus-specific fund deployment inferred.", "partition": 2}, {"url": "https://www.enverus.com/newsroom/enverus-one-is-live-the-governed-ai-platform-built-for-the-energy-industry/", "access": "full text reviewed", "title": "Enverus ONE launch", "publication_date": "2026-04-07", "claims_checked": ["AI-065 source claim(s) checked against cited profile"], "notes": "Enverus ONE launch announcement opened; product launch, capabilities and rollout wording checked, not customer use evidence.", "partition": 2}, {"url": "https://www.aramco.com/en/news-media/news/2025/pif-and-aramco-agree-for-aramco-to-acquire-a-significant-minority-stake-in-humain", "access": "full text reviewed", "title": "Aramco takes significant minority stake in HUMAIN", "publication_date": "2025-10", "claims_checked": ["AI-039 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.aramco.com/en/news-media/news/2026/aramco-signs-mou-with-microsoft-to-help-advance-industrial-ai-and-digital-talent-transformation", "access": "full text reviewed", "title": "Aramco-Microsoft MoU on industrial AI and digital talent", "publication_date": "2026-02-12", "claims_checked": ["AI-070 source claim(s) checked against cited profile"], "notes": "Aramco/Microsoft release opened; non-binding MoU and exploration language confirmed, not deployment.", "partition": 2}, {"url": "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai", "access": "full text reviewed", "title": "AI reprocesses legacy seismic; 6,000+ AI developers in training", "publication_date": "2025-06", "claims_checked": ["AI-038 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.petronas.com/media/media-releases/petronas-and-microsoft-expand-collaboration-support-malaysias-ai-ecosystem", "access": "full text reviewed", "title": "Petronas-Microsoft MoU on agentic AI and Copilot", "publication_date": "2025-07-31", "claims_checked": ["AI-079 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.oedigital.com/news/542771-petronas-turns-to-ai-to-accelerate-upstream-investment", "access": "full text reviewed", "title": "Agentic AI added to myPROdata upstream platform", "publication_date": "2026-09", "claims_checked": ["AI-080 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://reports.omv.com/en/annual-report/2025/directors-report-management-review/company/digitalization.html", "access": "full text reviewed", "title": "OMV AI Hub: 2,500 monthly GPT users, 10,000 trained", "publication_date": "Undated; see individual source review", "claims_checked": ["AI-084 source claim(s) checked against cited profile", "AI-132 source claim(s) checked against cited profile", "AI-133 source claim(s) checked against cited profile", "AI-134 source claim(s) checked against cited profile", "AI-135 source claim(s) checked against cited profile"], "notes": "OMV 2025 annual-report Digitalization section reviewed in full; AI Hub counts, Norwegian Companion, field-agent test, iDEA wording, and Romanian pump-ML pilot checked.", "partition": 2}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/08/11/devon-energy-dvn-q2-2026-earnings-call-transcript/", "access": "excerpt reviewed", "title": "AI speeds Coterra integration; in-house plus vendor model", "publication_date": "2026-08", "claims_checked": ["AI-087 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.fool.com/earnings/call-transcripts/2025/10/23/valero-energy-vlo-q3-2025-earnings-call-transcript/", "access": "excerpt reviewed", "title": "'Cautiously optimistic' on AI; data quality first", "publication_date": "2025-10-23", "claims_checked": ["AI-102 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.slb.com/news-and-insights/newsroom/updates/2025/slb%E2%80%99s-flagship-digital-platform-to-be-available-on-new-mistral-compute-ai-infrastructure", "access": "full text reviewed", "title": "Lumi platform runs on Mistral Compute, sovereign EU hosting", "publication_date": "2025-06", "claims_checked": ["AI-009 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.businesswire.com/news/home/20251103326704/en/SLB-Unveils-Groundbreaking-New-Agentic-AI-Technology-for-the-Energy-Industry", "access": "full text reviewed", "title": "Tela agentic AI assistant launched for upstream", "publication_date": "2025-11", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.energy-pedia.com/news/general/slb-industrializes-ai-for-the-energy-industry-with-nvidia-203316", "access": "full text reviewed", "title": "SLB and NVIDIA build 'AI Factory for Energy'", "publication_date": "2026-03", "claims_checked": ["AI-105 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.businesswire.com/news/home/20260615078195/en/SLB-Launches-Digital-Marketplace-to-Scale-AI-and-Digital-Innovation-Across-Energy", "access": "full text reviewed", "title": "SLB Digital Marketplace lists ~200 AI and digital products", "publication_date": "2026-06", "claims_checked": ["AI-047 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://jpt.spe.org/slb-expands-digital-portfolio-with-acquisition-of-ai-firm-tachyus", "access": "full text reviewed", "title": "SLB acquires Tachyus physics-ML reservoir startup", "publication_date": "2026-05", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.orphan-idle-wells.com/news/slb-acquires-the-software-running-us-oil-fields", "access": "full text reviewed", "title": "SLB buys Kingdom, Petra upstream software for agentic AI", "publication_date": "2026-04", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://worldoil.com/news/2025/1/6/halliburton-coterra-launch-first-fully-automated-hydraulic-fracturing-program/", "access": "full text reviewed", "title": "Octiv Auto Frac: first fully automated frac with Coterra", "publication_date": "2025-01", "claims_checked": ["AI-109 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.chevron.com/newsroom/2025/q2/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "access": "full text reviewed", "title": "ZEUS IQ closed-loop intelligent frac with Chevron", "publication_date": "2025-06", "claims_checked": ["AI-110 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://finance.yahoo.com/sectors/energy/articles/halliburton-strengthens-automation-strategic-sekal-133400448.html", "access": "full text reviewed", "title": "Halliburton acquires Sekal drilling-automation software", "publication_date": "2026-04", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.nov.com/news/nov-and-aiq-to-advance-ai-enabled-drilling-optimization", "access": "full text reviewed", "title": "NOV and AIQ pilot AI drilling on Max platform", "publication_date": "2026-01-12", "claims_checked": ["AI-114 source claim(s) checked against cited profile"], "notes": "NOV/AIQ AI-enabled drilling pilot release opened; pilot/partner scope checked, no commercial deployment claim inferred.", "partition": 2}, {"url": "https://investors.libertyenergy.com/news-and-events/press-releases/2025/10-16-2025-221920289", "access": "full text reviewed", "title": "Liberty launches Forge LLM for frac asset orchestration", "publication_date": "2025-10-16", "claims_checked": ["AI-118 source claim(s) checked against cited profile"], "notes": "Liberty Q3 2025 release opened; company-wide StimCommander outcomes are not isolated to LLM contribution.", "partition": 2}, {"url": "https://www.snowflake.com/en/news/press-releases/snowflake-launches-energy-solutions-for-the-ai-data-cloud-to-accelerate-shift-to-a-lower-carbon-future/", "access": "full text reviewed", "title": "Snowflake launches Energy Solutions with 30+ partner apps", "publication_date": "2026-01-27", "claims_checked": ["AI-119 source claim(s) checked against cited profile"], "notes": "Snowflake Energy Solutions announcement opened; ecosystem product/partner offering, not operator adoption.", "partition": 2}, {"url": "https://www.businesswire.com/news/home/20250909488860/en/Cognite-Atlas-AI-Drives-Customer-Momentum-with-New-Major-Release-Accelerates-Adoption-of-Agentic-AI-to-Unlock-More-Industrial-Value", "access": "full text reviewed", "title": "Cognite Atlas AI major release, low-code agent workbench", "publication_date": "2025-09", "claims_checked": ["AI-120 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and", "access": "full text reviewed", "title": "Cognite integrates NVIDIA accelerated computing for agents", "publication_date": "2025-10-14", "claims_checked": ["AI-121 source claim(s) checked against cited profile"], "notes": "Cognite/NVIDIA Celanese release opened; chemical-plant anomaly detection is prospective/product collaboration; not an O&G deployment.", "partition": 2}, {"url": "https://www.enverus.com/newsroom/enverus-ai-powering-the-next-era-of-energy-intelligence/", "access": "full text reviewed", "title": "Enverus AI launched as 'AI operating system for energy'", "publication_date": "2025-12", "claims_checked": ["AI-065 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.prnewswire.com/news-releases/introducing-seeq-intelligence-bridging-industrial-ai-and-human-expertise-for-smarter-operational-decisions-302702464.html", "access": "full text reviewed", "title": "Seeq Intelligence with Agent Q and custom agents", "publication_date": "2026-03", "claims_checked": ["AI-128 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.aveva.com/en/about/news/press-releases/2026/aveva-announces-new-capabilities-to-embed-ai-across-industrial-organizations-and-data-infrastructure-at-aveva-world-2026/", "access": "full text reviewed", "title": "AVEVA embeds agentic AI, MCP across CONNECT", "publication_date": "2026-05-20", "claims_checked": ["AI-129 source claim(s) checked against cited profile"], "notes": "AVEVA AI roadmap announcement opened; future Q1 2027 capabilities identified as planned, not shipped at source date.", "partition": 2}, {"url": "https://www.middleeastainews.com/p/aramco-ai--drives-4-billion-value", "access": "full text reviewed", "title": "Aramco: $4B tech value, 6,000+ AI-enabled staff", "publication_date": "2025-01", "claims_checked": ["AI-037 source claim(s) checked against cited profile"], "notes": "Direct URL was attempted as part of this partition’s unique-source sweep; claims reviewed against the profile title, source locator and returned page text where available.", "partition": 2}, {"url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "access": "full text reviewed", "title": "Right fix at right time – AI delivers for Woodside", "publication_date": "2026-01-05", "claims_checked": ["AI-042 source claim(s) checked against cited profile"], "notes": "Woodside page reviewed in full: Maint Intel workflow, Angel testing, North West Shelf launch, planned 2026 rollout and processing-time metric; the packet 15% annual hours estimate does not appear in the reopened page.", "partition": 2}, {"url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data", "access": "full text reviewed", "title": "Aramco AI and Big Data", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-141 source claim(s) checked against cited profile"], "notes": "Aramco AI and Big Data page directly opened; current page is undated, so accessed date is not publication date. Fadhili claim must be scoped to pilot and use separate dates from any source evidence.", "partition": 2}, {"url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "access": "inaccessible", "title": "PETRONAS Integrated Report 2025", "publication_date": "2026-02-27", "claims_checked": ["AI-142 source claim(s) checked against cited profile", "AI-143 source claim(s) checked against cited profile"], "notes": "PETRONAS 2025 Integrated Report PDF directly attempted twice; web viewer returned payload/size error. Search excerpts only; not full-text verified.", "partition": 2}, {"url": "https://www.petronas.com/bm/node/1688", "access": "full text reviewed", "title": "PETRONAS Carigali Unveils Pioneering AI Venture - TriCipta AI", "publication_date": "2025-06-18", "claims_checked": ["AI-143 source claim(s) checked against cited profile"], "notes": "PETRONAS TriCipta company release opened; product/venture scope checked, but no operator deployment or performance result established.", "partition": 2}, {"url": "https://petrobras.com.br/inovacao-e-tecnologia", "access": "full text reviewed", "title": "Inovação e tecnologia: nossa fonte de energia", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-146 source claim(s) checked against cited profile"], "notes": "Petrobras innovation and technology page opened; umbrella page does not itself prove the specific Smart Tocha workflow in sufficient detail.", "partition": 2}, {"url": "https://www.ecopetrol.com.co/wps/wcm/connect/4fec4771-90bd-446d-a107-cbe140583b88/Circular-031-esp-anexo-vf.pdf?CVID=pP5vUHj&MOD=AJPERES", "access": "full text reviewed", "title": "Informe de gestión y resultados 2025", "publication_date": "2026-03-31", "claims_checked": ["AI-147 source claim(s) checked against cited profile"], "notes": "Ecopetrol 2025 management report PDF opened; report text checked for AI/analytics distinctions. Group-science benefits are not automatically AI results.", "partition": 2}, {"url": "https://www.ecopetrol.com.co/wps/portal/Home/en/news/detail/%21ut/p/z0/pZHLTsMwEEV_BRZdWnZClJplFaI-1FciFVpvoqlljCGxXdsplK_HkRAgdcluRnNn7tEdzPAeMw1nJSEoo6GN_YHlzWYyownNyIbuioxU65oWi3ox3sxzvMAsCtZpXiTzLNmSfPdAqrKkVb1cJtNpNlxI3apYScwshBek9LPB-8I4axwEUQtvjfbqqFoVLnhf6rNyRndCh6ERTl6-PX4hUkpJdb8txnflfPw4SwYP9Xo6sQlm3OggPuKu4MaK4EzbvPOuadXRgbuMCPhG-BHRJiiu4E91k5I0GebQHVsBSIpoDi0CzmMUygfwKErSK5qrSP5DI11vDfpRx7Sgl70DxGMgziALsTY2qE59gkO-j5x--BUacKWIdwDbN3bYdk-IHSa3X8GQ_jA%21/", "access": "inaccessible", "title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "publication_date": "2026-05-04", "claims_checked": ["AI-148 source claim(s) checked against cited profile"], "notes": "Ecopetrol CADE announcement URL directly attempted; web cache miss. No claim-level verification from this URL.", "partition": 2}, {"url": "https://phe.pertamina.com/uploads/2025/09/f7c71ecf-01e5-4b33-b506-81282125ca2c.pdf", "access": "inaccessible", "title": "PINTAR Powered by ChatGPT: HSE Q&A Platform for PHR Rokan Zone", "publication_date": "2025-09-01", "claims_checked": ["AI-150 source claim(s) checked against cited profile"], "notes": "PHE PINTAR PDF URL directly attempted; request rejected. No full-text verification.", "partition": 2}, {"url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness", "access": "full text reviewed", "title": "PGE Demonstrates Digital Transformation Boosts National Geothermal Efficiency and Competitiveness", "publication_date": "2025-10-27", "claims_checked": ["AI-151 source claim(s) checked against cited profile"], "notes": "PGE release opened; its digital/geothermal scope is adjacent energy, not oil-and-gas operations. Bundled efficiency results cannot be isolated as AI outcomes.", "partition": 2}, {"url": "https://www.petronas.com/media/media-releases/petronas-advances-malaysias-upstream-data-platform-agentic-ai-accelerate-ep", "access": "full text reviewed", "title": "PETRONAS Advances Malaysia’s Upstream Data Platform with Agentic AI to Accelerate E&P Investment", "publication_date": "2026-09-03", "claims_checked": ["AI-080 source claim(s) checked against cited profile"], "notes": "PETRONAS agentic myPROdata release opened; reviewed stated scope and roadmap language. It establishes product/platform enhancement, not quantified asset outcomes.", "partition": 2}, {"url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "access": "inaccessible", "title": "Devon Energy Q2 2026 Earnings Call Transcript", "publication_date": "2026-08", "claims_checked": ["AI-087 source claim(s) checked against cited profile"], "notes": "Yahoo-hosted Devon transcript directly attempted and returned 429; secondary source not accessible. No Coterra claim certified from it.", "partition": 2}, {"url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/software/tela", "access": "full text reviewed", "title": "Tela agentic-AI assistant catalog", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-008 source claim(s) checked against cited profile"], "notes": "SLB Tela live product catalog opened; 33 is a mutable catalog/capability count, not customer deployments.", "partition": 2}, {"url": "https://www.slb.com/newsroom/press-release/2026/slb-launches-digital-marketplace-to-scale-ai-and-digital-innovation-across-energy", "access": "full text reviewed", "title": "SLB Launches Digital Marketplace to Scale AI and Digital Innovation Across Energy", "publication_date": "2026-06-15", "claims_checked": ["AI-047 source claim(s) checked against cited profile"], "notes": "SLB Digital Marketplace launch opened; catalog/marketplace scope checked, not proof of deployments.", "partition": 2}, {"url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "access": "full text reviewed", "title": "SLB Announces Fourth-Quarter and Full-Year 2025 Results", "publication_date": "2026-01-23", "claims_checked": ["AI-013 source claim(s) checked against cited profile"], "notes": "SLB Q4 2025 earnings PDF fully reviewed; includes OQ Neuro Oman, KOC AI work and broad digital results. Vendor evidence; no operator corroboration assumed.", "partition": 2}, {"url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program", "access": "full text reviewed", "title": "Coterra Energy and Halliburton launch first fully automated hydraulic fracturing program", "publication_date": "2025-01-06", "claims_checked": ["AI-109 source claim(s) checked against cited profile"], "notes": "Halliburton/Coterra announcement reviewed; source establishes automated fracturing, not an AI method.", "partition": 2}, {"url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing", "access": "full text reviewed", "title": "Chevron and Halliburton Enable Intelligent Hydraulic Fracturing", "publication_date": "2025-04-12", "claims_checked": ["AI-110 source claim(s) checked against cited profile"], "notes": "Halliburton/Chevron release reviewed; intelligent/closed-loop control is described, but source does not establish AI/ML.", "partition": 2}, {"url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom", "access": "full text reviewed", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "publication_date": "2025-02-26", "claims_checked": ["AI-157 source claim(s) checked against cited profile"], "notes": "Halliburton/Sekal announcement reviewed; on-bottom autonomous drilling outcome and control workflow checked; no explicit AI method.", "partition": 2}, {"url": "https://www.nov.com/news/nov-supports-record-bit-run-and-egypts-first-remote-operations-center", "access": "full text reviewed", "title": "NOV supports record bit run and Egypt’s first remote operations center", "publication_date": "2025-06-27", "claims_checked": ["AI-158 source claim(s) checked against cited profile"], "notes": "NOV Egypt/remote operations announcement opened; treated as supplier-authored deployment report; AI basis kept separate from remote automation.", "partition": 2}, {"url": "https://www.halliburton.com/en/resources/autonomous-drilling-technology-enhances-well-placement-and-rop", "access": "full text reviewed", "title": "Enhanced well delivery with LOGIX automation and remote operations", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-016 source claim(s) checked against cited profile"], "notes": "Halliburton LOGIX source opened; automation, ROP and well placement content checked; it describes autonomous/automated drilling, not a demonstrated AI model.", "partition": 2}, {"url": "https://www.halliburton.com/en/about-us/press-release/halliburton-shape-digital-strategic-collaboration-advance-digital-asset-management", "access": "full text reviewed", "title": "Halliburton and Shape Digital form strategic collaboration to advance digital asset performance management", "publication_date": "2026-07-21", "claims_checked": ["AI-018 source claim(s) checked against cited profile"], "notes": "Halliburton/Shape collaboration release opened; future strategic collaboration, no customer deployment or achieved result.", "partition": 2}, {"url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "access": "full text reviewed", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "publication_date": "2026-06-09", "claims_checked": ["AI-046 source claim(s) checked against cited profile"], "notes": "SLB Agora/Qualcomm release opened; edge-AI product/collaboration scope checked, no named deployed customer.", "partition": 2}, {"url": "https://www.slb.com/zh-cn/newsroom/features/2026/2026-0618-intelligent-production-studio", "access": "full text reviewed", "title": "Inside the Future of Production Operations", "publication_date": "2026-06-18", "claims_checked": ["AI-048 source claim(s) checked against cited profile"], "notes": "SLB Intelligent Production Studio feature opened; product workflow descriptions checked; case results require careful attribution to SLB digital suite.", "partition": 2}, {"url": "https://www.slb.com/products-and-services/delivering-digital-at-scale/production/intelligent-production-studio", "access": "full text reviewed", "title": "Intelligent Production Studio", "publication_date": "Undated; accessed 2026-09-27", "claims_checked": ["AI-048 source claim(s) checked against cited profile"], "notes": "SLB product page opened; current capability is vendor offer, page is undated.", "partition": 2}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-industrializes-ai-energy-industry-nvidia/", "access": "full text reviewed", "title": "SLB Industrializes AI for the Energy Industry with NVIDIA", "publication_date": "2026-03-25", "claims_checked": ["AI-105 source claim(s) checked against cited profile"], "notes": "SLB/NVIDIA AI Factory announcement opened; commercialization partnership/product roadmap, not upstream AI deployment.", "partition": 2}, {"url": "https://www.slb.com/newsroom/updates/2026/2026-0528-slb-tachyus-update", "access": "full text reviewed", "title": "SLB to Acquire Tachyus", "publication_date": "2026-05-29", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "SLB Tachyus acquisition update opened; agreement/transaction timing checked, no purchase price or close date inferred.", "partition": 2}, {"url": "https://investorcenter.slb.com/static-files/446a7319-9ca3-495a-83fb-c8afed1f7ebc", "access": "full text reviewed", "title": "SLB Announces Second-Quarter 2026 Results", "publication_date": "2026-07-24", "claims_checked": ["AI-106 source claim(s) checked against cited profile"], "notes": "SLB Q2 2026 earnings Form 8-K opened; portfolio and transaction references checked.", "partition": 2}, {"url": "https://www.slb.com/newsroom/updates/2026/2026-0423-sp-global-acquisition", "access": "full text reviewed", "title": "SLB Advances Digital Strategy with Acquisition of S&P Global Energy’s Upstream Software Portfolio", "publication_date": "2026-04-24", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "SLB S&P acquisition release opened; agreement distinguished from completion.", "partition": 2}, {"url": "https://press.spglobal.com/2026-09-01-S-P-Global-Completes-Divestiture-of-Upstream-Energy-Software-Portfolio-to-SLB", "access": "full text reviewed", "title": "S&P Global Completes Divestiture of Upstream Energy Software Portfolio to SLB", "publication_date": "2026-09-01", "claims_checked": ["AI-107 source claim(s) checked against cited profile"], "notes": "S&P Global completion release opened; completion date checked.", "partition": 2}, {"url": "https://ir.halliburton.com/node/36266/pdf", "access": "full text reviewed", "title": "Halliburton Acquires Sekal AS From Sumitomo Corporation to Accelerate Drilling Automation Services", "publication_date": "2026-04-01", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Halliburton Sekal acquisition release opened; acquisition closed/disclosed date distinct from product automation claims.", "partition": 2}, {"url": "https://ir.halliburton.com/node/35181/pdf", "access": "full text reviewed", "title": "Halliburton and Sekal Deliver World’s First Automated On-bottom Drilling System for Equinor in the North Sea", "publication_date": "2025-02-26", "claims_checked": ["AI-111 source claim(s) checked against cited profile"], "notes": "Halliburton/Sekal Equinor case PDF opened in full; Feb 2025 on-bottom system, integration, autonomy and safety limits checked; AI itself is not explicit.", "partition": 2}, {"url": "https://www.cognite.com/en/company/newsroom/cognite-atlas-ai-drives-customer-momentum-with-new-major-release", "access": "full text reviewed", "title": "Cognite Atlas AI major release accelerates industrial AI", "publication_date": "2025-09-09", "claims_checked": ["AI-120 source claim(s) checked against cited profile"], "notes": "Cognite Atlas AI product release opened; product architecture/workbench claims checked; not proof of customer deployment.", "partition": 2}, {"url": "https://www.seeq.com/resources/press-releases/introducing-seeq-intelligence/", "access": "full text reviewed", "title": "Introducing Seeq Intelligence", "publication_date": "2026-03-03", "claims_checked": ["AI-128 source claim(s) checked against cited profile"], "notes": "Seeq Intelligence release opened; product/Agent Q scope checked; no named O&G deployment or outcome.", "partition": 2}, {"url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "access": "full text reviewed", "title": "Encino Energy streamlines operations with Microsoft Power Platform", "publication_date": "2025-04-29", "claims_checked": ["AI-166 source claim(s) checked against cited profile"], "notes": "Microsoft Encino customer case opened; Power Platform process workflows reviewed; does not establish AI as the enabling technology.", "partition": 2}, {"url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "access": "full text reviewed", "title": "Infinity PL drives efficiency gains for Encino Energy", "publication_date": "2025-08-05", "claims_checked": ["AI-168 source claim(s) checked against cited profile"], "notes": "Ambyint vendor case opened; vendor-reported InfinityPL lift optimization/results, no independent operator corroboration.", "partition": 2}, {"url": "https://imubit.com/blog/preem-partners-with-imubit-to-drive-sustainability-and-lower-emissions-through-advanced-closed-loop-ai-optimization", "access": "full text reviewed", "title": "Preem partners with Imubit for closed-loop AI optimization", "publication_date": "2025-02-19", "claims_checked": ["AI-169 source claim(s) checked against cited profile"], "notes": "Imubit/Preem announcement opened; says Optimizing Brain will be deployed, so outcomes are expected, not achieved.", "partition": 2}, {"url": "https://investorvalero.com/events-and-presentations/events/event-details/2025/Q3-2025-Valero-Energy-Corp-Earnings-Conference-Call-2025-qoWmz9-RZb/default.aspx", "access": "full text reviewed", "title": "Valero Q3 2025 earnings call event", "publication_date": "2025-10-23", "claims_checked": ["AI-102 source claim(s) checked against cited profile"], "notes": "Valero issuer earnings event page opened; event metadata only, not transcript content.", "partition": 2}, {"url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas", "access": "full text reviewed", "title": "ECP and KKR announce Texas hyperscale data-center campus", "publication_date": "2025-07-30", "claims_checked": ["AI-059 source claim(s) checked against cited profile"], "notes": "ECP/KKR campus announcement opened; AI data-center demand infrastructure, not oil/gas company AI adoption.", "partition": 2}, {"url": "https://targaresources.referrals.selectminds.com/jobs/principal-engineer-ai-emerging-technology-3925", "access": "full text reviewed", "title": "Targa Principal Engineer, AI & Emerging Technology job posting", "publication_date": "2026-05-07", "claims_checked": ["AI-176 source claim(s) checked against cited profile"], "notes": "Targa job posting opened; establishes hiring intent/role, not deployed operational AI.", "partition": 2}, {"url": "https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/saudi-aramco-fy-2025-webcast-presentation-script-english.pdf", "access": "full text reviewed", "title": "FY 2025 results presentation and script", "publication_date": "2026-03-10", "claims_checked": ["AI-037: $5.3bn total 2025 TRV; CEO says over half enabled by AI; mixed-technology components are not an AI-only breakdown."], "notes": "Primary Aramco FY2025 remarks/presentation reviewed. CEO remarks explicitly state total TRV was $5.3bn, with over half enabled by AI; this is high-level company attribution, not independently validated project-level value.", "partition": 2}, {"url": "https://www.aramco.com/en/news-media/news/2026/fourth-quarter-and-full-year-press-release", "access": "full text reviewed", "title": "Aramco announces fourth quarter and full-year 2025 results", "publication_date": "2026-03-10", "claims_checked": ["AI-039 source claim(s) checked against cited profile"], "notes": "Aramco FY2025 result release opened; company financial/AI statements checked.", "partition": 2}, {"url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025", "access": "full text reviewed", "title": "Remarks by Ahmad O. Al-Khowaiter at LEAP 2025", "publication_date": "2025-02-09", "claims_checked": ["AI-177 source claim(s) checked against cited profile"], "notes": "Aramco LEAP remarks opened; MetaBrain for Plants described with future tense; workflow not treated as achieved plant operation.", "partition": 2}, {"url": "https://www.iocl.com/uploads/AnnualReport2024-25/intellectual-cap.html", "access": "full text reviewed", "title": "IndianOil – Intellectual Capital, Integrated Annual Report 2024-25", "publication_date": "2025 (FY2024-25)", "claims_checked": ["AI-180 source claim(s) checked against cited profile"], "notes": "IndianOil FY2024-25 Intellectual Capital page opened; broad AI/ML predictive-maintenance statement checked; no sites, asset types or outcomes disclosed.", "partition": 2}, {"url": "https://www.ril.com/sites/default/files/2026-06/SE_Chairmans_Statement_signed.pdf", "access": "full text reviewed", "title": "Chairman’s Statement, 49th Annual General Meeting", "publication_date": "2026-06", "claims_checked": ["AI-183 source claim(s) checked against cited profile"], "notes": "RIL chairman statement PDF opened; crude-feedstock AI optimizer statement checked; no quantified result, scale or method disclosed.", "partition": 2}, {"url": "https://chevroncorp.gcs-web.com/static-files/f8964224-9b9a-4be6-8644-9f96bd1e1fe4", "title": "APOLO prospectivity and development optimization", "publication_date": "2025-11-12", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://finance.yahoo.com/markets/stocks/articles/devon-energy-dvn-q2-2026-234337276.html", "title": "Closed-loop AI adjusts gas-lift settings at operating wells", "publication_date": "2026-08", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf", "title": "AI agents for drilling and completions monitoring", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing profile says models and agents identify unproductive time and repeat better drilling/completion practices; claim is not reverified in this pass."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.slb.com/newsroom/press-release/2025/pr-2025-1211-slb-shell-collab", "title": "Upstream data/AI infrastructure and agentic-solutions collaboration", "publication_date": "2026-01-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["SLB and Shell agreed to develop upstream digital and AI solutions; roadmap aims for open data/AI infrastructure across subsurface, well construction and production on SLB Lumi.", "Announcement states ambition/aim, not a deployed customer workflow or realized results."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://investors.bakerhughes.com/files/doc_financials/2025/Q4/Fourth-Quarter-2025_Earnings-Conference-Call-Presentation.pdf", "title": "Data-center orders", "publication_date": "2026-01-26", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.epam.com/insights/webinars/speed-to-value-in-asset-performance-with-azure-ai-a-case-study-on-epam-and-baker-hughes", "title": "Cordant Asset Strategy AI with EPAM", "publication_date": "2025-04-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["EPAM describes generative AI and unstructured text processing for Cordant Asset Strategy, including engineering document analysis and Azure integration."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.c3.ai/node/10916/pdf", "title": "Baker Hughes–C3 AI enterprise AI joint venture renewal", "publication_date": "2025-05-28", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["C3 AI and Baker Hughes renewed and expanded their joint venture agreement for multiple years through June 2028."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.shell.com/investors/investor-presentations/capital-markets-day-2025/_jcr_content/root/main/section/promo_1962010312_cop/links/item3.stream/1752057184780/7babf10c6fc8e1415eadc37050252bb4a7c879a2/shell-capital-markets-day-2025-accessibility-sides.pdf", "title": "Technology and AI cost reduction", "publication_date": "2025-03", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://totalenergies.com/newsroom/totalenergies-collaborate-mistral-ai-increase-application-artificial/?lang=eng", "title": "Joint AI innovation lab for researcher and industrial decision support", "publication_date": "2025-06-12", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Joint lab use cases include an assistant for 1,000 researchers, decision-support for industrial performance and CO2 emissions, and customer energy-saving support.", "TotalEnergies reported Digital Factory had developed over 100 solutions, 60 using ML through generative AI; these are aggregate portfolio counts, not outcome metrics."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-to-collaborate-with-mistral-ai-to-increase_2025_en.pdf", "title": "Digital Factory portfolio of AI and digital solutions", "publication_date": "2025-06-12", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["TotalEnergies says its Digital Factory had 300 digital experts and developed >100 solutions, 60 using technologies from ML to GenAI.", "The announcement creates a distinct Mistral joint innovation lab for researcher assistant, industrial asset decision support, customer and renewable use cases; it says those use cases will be designed/tested, not deployed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "title": "Multi-agent workflow platform for employee productivity", "publication_date": "2026-02-19", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Accenture collaboration produced an agent-construction platform and pilot in Digitalization and Services department.", "Three teams with 22 agents supported processes for over four months; over 50 people supervised the agent teams.", "Repsol reports 60 production GenAI use cases and 22 scaled as of Oct 2025; these are company-reported portfolio measures.", "Repsol says its GenAI Competence Center had produced 60+ use cases in production and 22 scaled; 5,000+ employees trained.", "Platform was a pilot in Digitalization and Services with 3 teams and 22 agents supporting processes for more than four months; 50+ people supervised agents.", "A separate 2025 consolidated management report later says 34 agents across three workflows; dates/scope differ, so counts should not be blended."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "title": "AI-enabled maintenance, field planning and seismic portfolio", "publication_date": "2026-01-07", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Company says AI generated USD 130m value for Equinor and partners in calendar 2025, with >USD 330m cumulative since 2020.", "Three applications: 700+ rotating machines/24,000 sensors predictive maintenance (USD 120m since 2020); well and field planning, including USD12m Johan Sverdrup phase-3 saving; seismic interpretation of 2m km2 in 2025 with tenfold capacity."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "title": "AI-enhanced subsurface development", "publication_date": "2026-05-06", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing register cites AI-enhanced subsurface development; excerpt reviewed from same transcript only supports a broader increase in AI use and does not name this workflow.", "COO said AI is taking a larger role across company disciplines as part of workforce efficiency; no workflow, deployed scale or AI-only savings quantified."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.cognite.com/en/company/newsroom/aker-bp-leverages-cognite-atlas-ai-tm-to-pioneer-an-ai-first-future-in-exploration-and", "title": "Atlas AI agents for root-cause analysis and E&P workflows", "publication_date": "2025-09-16", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Aker BP expanded Atlas AI adoption. Described workflows include document processing for supply chain, barrier management, automatic equipment registration, data foundation for predictive maintenance and production chemistry.", "Root-cause agents identify relevant info, build cause maps and suggest steps; Cognite/Aker BP report >70% reduction in engineer RCA time."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.quantumcap.com/digital-mission/", "title": "Quantum Energy Cloud portfolio data foundation, with related QIQ and AI-at-Work workflows", "publication_date": "2025-05-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Quantum Energy Cloud unifies portfolio financial, operational and market data using Microsoft and Databricks.", "QIQ is a proprietary generative AI interface that answers questions, automates analyses and initiates workflows over portfolio data.", "Named workflows include single-well economics, geospatial capital allocation, completion design, artificial lift, anomaly detection and downtime prediction."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://pakenergy.com/blog/pakenergy-launches-embedded-ai-powered-lease-provision-capability", "title": "Embedded AI lease provision processing", "publication_date": "2025-09-25", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Embedded AI processes lease provision research and data inside PakEnergy Land Management.", "Company says work previously taking hours can be done in seconds; no accuracy or independent outcome data disclosed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://pakenergy.com/blog/pakenergy-launches-ai-powered-ticket-n-receipt-management-to-pakenergy-transportation?hs_amp=true", "title": "PakCAPTURE ticket and receipt extraction", "publication_date": "2026-05-20", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Drivers photograph receipts or paper tickets; PakAI extracts data, populates pickup/drop forms and associates information with a load.", "Feature embedded in PakTransportation and available to customers."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-and-cloudburst-sign-agreement-natural-gas-supply", "title": "CloudBurst San Marcos natural-gas supply for AI campus", "publication_date": "2025-02-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Up to 450,000 MMBtu/day firm gas supply for campus near San Marcos, supporting up to ~1.2 GW behind-the-meter power, contingent on customer FID; AI-demand infrastructure."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "title": "Closed-loop process optimization at CNOOC Bohai offshore terminal", "publication_date": "2026-07-24", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Q2 disclosure: four KOC Jurassic-gas AI awards (rig scheduling, GenAI, log QC and academy).", "Q2 disclosure: three-year DrillOps award for up to 120 ONGC rigs.", "Q2 disclosure: PTTEP one-asset offshore pilot using Cognite, OptiFlow and Tela agents.", "Q2 disclosure: CNOOC subsidiary terminal connects Symmetry twins to APC/DCS.", "ExxonMobil Trinidad and Tobago Deepwater awarded SLB a two-year seismic data-processing project for sparse 2D ultra-deepwater block.", "Omega workflow includes full-waveform inversion, velocity modeling, source separation, multicomponent streamer processing and AI-enabled horizon interpretation.", "Coverage is nearly 6,000 km2 and intended to deliver first 3D image; this is project scope, not discovery."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas", "title": "A smarter way to prospect for oil and gas", "publication_date": "2025-11-12", "notes": "Official Chevron newsroom article reopened successfully; replaces the earlier stale inaccessible status.", "claims_checked": ["Chevron says ApEX launched Aug. 2024 and uses specialized agents to search more than one million files; the issuer page reports a 4 TB corpus and seconds-fast retrieval.", "The article does not state that pre-drill studies fell from weeks to minutes."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.datarobot.com/newsroom/press/datarobot-and-chevron-collaborate-to-advance-agentic-ai-for-autonomous-inspections/", "title": "Edge agentic assessment for robotic inspections", "publication_date": "2026-06-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Chevron and DataRobot collaboration applies edge agents to support autonomous inspection at Chevron facilities.", "Safe Start assesses conditions before/during robotic missions using gas sensors, vision and locally running models; integrates with existing operating systems and NVIDIA NIM.", "Vendor and Chevron program manager describe reduced manual point-in-time checks and human oversight/control; no measured outcomes."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://c3ai.gcs-web.com/news-releases/news-release-details/c3-ai-and-shell-expand-collaboration-scaling-reliability-ai", "title": "Agentic root-cause analysis in global reliability program", "publication_date": "2026-06-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["C3 AI says it has worked with Shell since 2018 on predictive maintenance monitoring over 13,000 equipment items.", "A new multi-year agreement announced June 2026 extends reliability AI and introduces agent-based root cause analysis and remediation; the agreement is prospective.", "Program uses C3 AI Reliability and C3 Agentic AI Platform on Microsoft Azure.", "Hundreds of millions of dollars value is a C3 AI executive claim, not independent or Shell-attributed validation."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://totalenergies.com/newsroom/totalenergies-annonce-un-partenariat-avec-mistral-en-vue-de-developper-des-modeles-de-frontiere-dintelligence-artificielle-dedies-a-lexploration-et-a-lingenierie-des-reservoirs-498514/?lang=eng", "title": "Frontier AI models for reservoir exploration and engineering", "publication_date": "2026-09-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three-year joint program represents over €100 million investment to develop frontier AI models supporting exploration, characterization and development of oil and gas reservoirs.", "Program will use agentic AI and combine almost 10 petaflops of data with geoscience/reservoir expertise; develop multiple scenarios for exploration opportunities and optimizing existing projects.", "Scientific lab will combine TotalEnergies subsurface experts with Mistral AI science and technology."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://totalenergies.com/system/files/documents/totalenergies_pr-totalenergies-cognite-reinforce-partnership-accelerate-digital-transformation-scale-industrial-AI_2025_en.pdf", "title": "Industrial data and AI foundation across operated upstream", "publication_date": "2026-02", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three-year rollout of Cognite industrial data and AI platform across all TotalEnergies-operated upstream assets worldwide, drilling through production.", "The initiative contextualizes industrial data, supports dynamic asset visualization and monitoring of critical equipment, and aims to accelerate applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://global.chinadaily.com.cn/a/202604/01/WS69cc7f5aa310d6866eb411c3.html", "title": "Offshore predictive maintenance cuts failures ~20%", "publication_date": "2026-04", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Specific claim in the assigned record was checked against the opened secondary/editorial/transcript page, not an issuer-original page."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.aogr.com/magazine/cover-story/devon-energy-stays-ahead-of-technology-curve", "title": "ChatDVN and AI-101 workforce enablement", "publication_date": "2025-07", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["ChatDVN internal AI platform; leaders trained via AI-101; engineers, software developers and geoscientists report 15%-30% productivity improvement; humans review/approve AI recommendations; video analytics plans for 2026."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.fool.com/earnings/call-transcripts/2025/11/27/eog-resources-eog-q3-2025-earnings-call-transcript/", "title": "Generative AI for geology, drilling and safety", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Third-party transcript reproduces EOG comments on generative AI, geology, drilling, safety, predictive maintenance and in-house applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/eog-resources-eog-q4-2025-earnings-transcript/", "title": "ML production optimizers and remote Utica operations", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Third-party transcript reproduces EOG claims on ML production optimizers and remote Utica operations."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.matadorresources.com/news-releases/news-release-details/matador-resources-company-reports-fourth-quarter-and-full-13", "title": "Vendor-led, cautious AI integration strategy", "publication_date": "2026-02-25", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Matador lists AI integration among focus areas, without named deployment, vendor, use case, budget or outcome; distinct from natural-gas marketing to AI data centres."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous", "title": "Autonomous haulage fleet expansion at Base Plant", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["AHS fleet expansion reached almost 120 trucks across two integrated mines; operators manage flow from control room with real-time vehicle visibility; Suncor planned 140 equipped trucks by year end. Suncor says the system uses AI to optimize for driving conditions and reduce wear/maintenance."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://s22.q4cdn.com/128149789/files/doc_events/2026/Feb/04/PSX-4Q-2025-Earnings-Call-Transcript.pdf", "title": "Machine learning in refinery turnaround planning", "publication_date": "2026-02-04", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes machine learning tools in turnaround planning, moving work outside turnaround to shorten duration/financial impact; company says forecast/spend improved, does not isolate AI causation or savings."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://s22.q4cdn.com/128149789/files/doc_financials/2026/q2/Transcript-PHILLIPS-66-FQ2-2026-EARNINGS-CALL.pdf", "title": "Company AI program and frontline solution replication", "publication_date": "2026-08", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes broad AI program, copying use cases, frontline engineers extracting data and deploying solutions more quickly; leadership says ML long used for maintenance/turnarounds."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/", "title": "Strategic investment in Flyscan AI-assisted aerial inspection", "publication_date": "2025-04-16", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["KMI joins strategic investor group; vendor describes remote sensing and AI visual patrol/leak detection. This confirms investment/partnership, not KMI rollout."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.bakerhughes.com/company/news/baker-hughes-repsol-launching-new-aipowered-functionality-leucipatm", "title": "Leucipa GenAI assistant and lift optimizer for Repsol", "publication_date": "2025-06-10", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126", "title": "Industrial Intelligence portfolio for AI and edge operations", "publication_date": "2026-04-01", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Portfolio includes MetaFlow AI/physics-based virtual flow meter, Engineer’s Briefcase AI document transcription/recognition/translation tools and StrataCore production foundational model.", "It also includes unified data model, edge and SaaS software; portfolio launch does not itself verify customer deployment."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.nabors.com/nabors-and-corva-expand-strategic-alliance-to-accelerate-digital-innovation/", "title": "RigCLOUD Powered by Corva drilling intelligence", "publication_date": "2025-04-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Expanded alliance integrates RigCLOUD edge/cloud platform with Corva AI analytics and apps for real-time drilling data, predictive insights and optimization.", "Operators and contractors can use the tools at rigsite and remotely; SmartROS controls are optional. Announcement describes solution capability, not specific named customer rollout."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HELMERICH-PAYNE-INC--ANNOUNCES-FISCAL-FIRST-QUARTER-RESULTS-02-04-2026/default.aspx", "title": "FlexRobotics rig-floor handling with AI-enabled zone monitoring", "publication_date": "2026-02-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["H&P reports FlexRobotics deployed on a Permian rig for a supermajor across multiple pads; growing customer interest.", "Rig-floor system is collaborative robotics for tubular handling. H&P product page says AI-enabled zone monitoring is an ancillary feature; core handling automation is not AI."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.businesswire.com/news/home/20250415281159/en/Collide-Secures-$5M-to-Scale-First-GenAI-Platform-for-Energy-Sector", "title": "Collide $5m seed and energy-specific GenAI knowledge platform", "publication_date": "2025-04-15", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Collide announced a $5m seed led by Mercury Fund; proceeds support platform and engineering expansion.", "Platform uses retrieval augmented generation and LLMs over curated technical papers, internal processes and Q&A to provide cited, domain-specific energy answers.", "Collide reported 6,000+ community users and estimated 25%-33% time savings for petroleum engineers; these are company estimates and community users are not equivalent to deployed enterprise users."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.streetinsider.com/PRNewswire/Winn%2BResources%2BPartners%2Bwith%2BCollide%2Bto%2BTurn%2BOperational%2BData%2BInto%2Ba%2BCompetitive%2BAdvantage/26116586.html", "title": "Winn Resources Texas RRC W-10/G-10 filing workflow", "publication_date": "2026-03-05", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["Winn says Collide automates monthly Texas RRC W-10/G-10 filings using existing well test/production data; task moves from multi-hour effort to under 30 minutes.", "Winn says workflow surfaces missing-data gaps earlier and reports >95% time reduction. This is customer-reported; no independent audit."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.globenewswire.com/news-release/2026/09/08/3358243/18320/en/battalion-oil-corporation-announces-strategic-investment-in-collide-the-ai-native-operations-system-for-oil-and-gas.html", "title": "Battalion strategic investment and planned Collide Operations System rollout", "publication_date": "2026-09-08", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Battalion made strategic equity investment and will deploy Collide Operations System under subscription.", "Initial plan unites more than 100 TB of records, contracts and production history; planned use cases include gas reconciliation, field invoices, power, and failure diagnosis.", "Riggs is expected to be available to production engineers around day 90; human review/control retained."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.emerson.com/en/corporate/news/2025/emerson-completes-acquisition-of-remaining-outstanding-shares-of-aspentech", "title": "Emerson completes AspenTech buyout for industrial AI", "publication_date": "2025-03", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Company/workflow identity and AI-vs-conventional-automation basis where stated", "Specific workflow, rollout stage/date and disclosed metrics relevant to assigned record"], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.microsoft.com/en/customers/story/24604-aker-bp-microsoft-365-copilot", "title": "Copilot rollout, employee agents and Knowledge Agents", "publication_date": "2025-06-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Aker BP started with 300 Copilot licenses and planned all-staff licensing in 2025; trained employees, set up help desk and champion/accountability network.", "Employee/small-team-built agents, including Knowledge Agents, answer repetitive documentation questions and route harder queries; several agents live.", "Microsoft case quotes 96% active users in prior four weeks; vendor-published customer metric, not independent audit."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/", "title": "AI agents for Yggdrasil remote operations", "publication_date": "2026-02-11", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Yggdrasil operations from onshore centre planned from 2027; AI agents planned from day one for condition-based maintenance and planning, with scaling to other workflows as mature.", "Aker BP says industrial data platform Cognite Data Fusion structures thousands of sensor streams and data is to be shared with people and AI agents."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.shell.com/news-and-insights/our-stories/putting-safety-first.html", "title": "Real-time machine vision and equipment health AI at Pearl GTL", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Shell says Pearl GTL uses machine vision to flag unsafe behaviors and hazardous conditions in real time and alert supervisors.", "Pearl uses AI for near-real-time equipment health insights to reduce manual inspections and support earlier intervention.", "The plant was recognized by the World Economic Forum Global Lighthouse Network in September 2025."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ltd.cnooc.com.cn/xwzx/gsxw/2025/202508/t20250827_117385.html", "title": "AI+ smart field, intelligent drilling and typhoon response", "publication_date": "2025-08-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["CNOOC reported implementing priority AI+ scenarios, scaling intelligent injection-production technology, and advancing intelligent drilling/completion construction.", "Company reported demonstration project drilling speed improvement of 26%; the release does not allocate this result exclusively to AI.", "Deep Sea No.1 smart gas field named as a national excellent-level smart factory; satellite remote sensing, unmanned equipment and AI algorithms were combined for typhoon response."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://sustentabilidad.ypf.com/assets/docs/es/YPF-Reporte-de-sustentabilidad-2024.pdf", "title": "Generative AI for contract cataloging, customer service and refinery inspection", "publication_date": "Undated source; see chronology", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy4q24transcript.pdf", "title": "AI Center of Excellence and cross-functional operating use cases", "publication_date": "2025-02-19", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Oxy named AI use in Gulf of America supply-chain management, asset integrity and reservoir characterization; established an AI Center of Excellence to align cross-company initiatives."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/dispatching-innovation-suncors-new-dispatch-system-improves-mine-operations", "title": "Dynamic AI dispatch assigns mine haul trucks", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Dynamic dispatch automates dump/refueling assignments and slowdown rerouting; human dispatchers remain active."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.slb.com/resource-library/article/2026/petoro-and-slb-pioneering-ai-driven-well-planning", "title": "Petoro AI well-data extraction and field-development screening", "publication_date": "2026-06-05", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Petoro and SLB built Lumi/Innovation Factori workflows to extract data from well schematics/reports and rank field-development/sidetrack opportunities.", "Well Schematic Extractor and Well History use LLMs to extract casing, completion and formation information, with deterministic checks and user visual QC.", "SLB reports threefold productivity improvement; preliminary QC rose from two schematics/day to six or seven. This is vendor case-study reporting."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.saipem.com/sites/default/files/2026-04/Saipem_Sustainability_Overview_2025.pdf", "title": "AI Project Schedule Onshore for EPC risk forecasting", "publication_date": "2026-04", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Saipem’s AI for Engineering is a two-year cross-functional effort launched 2024; 13 use cases explored by 2025.", "AI Project Schedule Onshore analyzes historical project data with predictive models for EPC construction risks/delays and bottlenecks; 12 additional apps were planned for 2026."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://c3.ai/customers/baker-hughes-achieves-bottom-line-savings-with-ai-sourcing-optimization", "title": "AI sourcing optimization across procurement data", "publication_date": "Undated source; see chronology", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Case study says C3 AI Sourcing Optimization supported 900,000 SKUs across 800+ sites and realized millions in economic benefit in CY2025.", "Workflow addresses fragmented procurement data and manual spreadsheet reconciliation.", "Vendor-published customer story; not independent assurance."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "title": "MARGE maintenance report classification and BuyerCompanion procurement agent", "publication_date": "2025-10-22", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["MARGE uses Copilot Studio to classify intervention reports; company reports 25-45% improvement in fault-cause identification.", "BuyerCompanion streamlines purchases up to €50,000, operational in France, with company target of 10% savings on certain purchases.", "30,000 Copilot licenses, training and internal agent-validation committee described."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://sustainability.oneok.com/innovation-spotlight", "title": "AI virtual river gauges for extreme-weather response", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reports AI-powered virtual river gauges and 48,872 deployed; weather-response risk triage. Undated page does not establish launch date or confirm in-window deployment."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.energytransfer.com/news-releases/news-release-details/entergy-louisiana-and-energy-transfer-sign-agreement-supports", "title": "Long-term gas transportation for Meta Louisiana data center", "publication_date": "2025-11-04", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Long-term gas transportation infrastructure for future Meta site; separate AI-demand category, not internal AI."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.energytransfer.com/news-releases/news-release-details/energy-transfer-reports-first-quarter-2026-results-and-updates", "title": "Nexus Hubbard Campus gas transport for behind-the-meter AI campus", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reports agreements to transport gas to Nexus Hubbard Campus, an AI data center campus with on-site gas generation; infrastructure supporting AI demand."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.petronas.com/bm/node/1668", "title": "Industrial AI copilot integrated with engineering technical references", "publication_date": "2025-06-10", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["AIngineer is a PETRONAS-incubated industrial AI copilot using LLMs and ML and integrating technical databases and standards.", "PETRONAS says it has technical use cases and prior deployments in energy/government; this release does not name a customer asset or quantify results.", "Microsoft collaboration includes cloud scaling and prospective marketplace commercialization."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://totalenergies.com/sites/g/files/nytnzq121/files/documents/totalenergies_2025-results-and-2026-objectives-presentation_2026_en.pdf", "title": "2025 Results and 2026 Objectives Presentation", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Reported data platform footprint: 40 upstream, 16 refining and chemicals, and 500 integrated power sites; platform captures 10x real-time data points and contextualizes 35 million points.", "The deck names three AI strategic programs and a new global competency center in India, with target production and availability uplift but no quantified realized uplift."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/informes-anuales/2025/consolidated-management-report-2025.pdf", "title": "Consolidated Management Report 2025 — Repsol", "publication_date": "2026-02-19", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["Repsol reports Automated Well Data Ingestion automates daily drilling and geology reports and integrates data into corporate systems to speed exploration decisions.", "Annual report later lists 34 agents in three workflows, distinct from the Oct 2025 pilot snapshot of 22 agents.", "2025 report says Competence Center promoted >75 initiatives after two years, and the multi-agent platform had 34 agents in three workflows.", "Report names automated well-data ingestion for daily drilling/geology reports, integration into corporate systems and faster exploration decision-making."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf", "title": "Aker BP Q4 2025 transcript", "publication_date": "2026-02-11", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["CEO describes AI-enabled earlier exploration decisions and AI agents reducing operational troubleshooting time and improving uptime.", "The CEO describes high-quality structured real-time data connection across the company; 10 years of digital investment is contextual foundation.", "Microsoft speaker (partner attribution) says Copilot adoption nearly 100%; not independently audited."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "title": "Devon Q4 2025 Earnings Q&A Webcast Commentary Transcript", "publication_date": "2026-02", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Existing profile claims second-half 2025 lift trials and Q1–Q2 2026 broader rollout; those details remain unverified in this review."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/02/25/matador-mtdr-q4-2025-earnings-call-transcript/", "title": "Matador Q4 2025 Earnings Call", "publication_date": "2026-02-25", "notes": "Reopened and checked the visible record-specific claim. This is a secondary article or third-party transcript rendering, not issuer-authored primary evidence; it cannot by itself substantiate an issuer-verified label.", "claims_checked": ["CEO describes vendor-supported, cautious “fast follower” approach, without naming vendors or applications."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://investorcenter.slb.com/static-files/44014844-6083-4e5a-8492-f8f89a1ef609", "title": "SLB Announces Third-Quarter 2025 Results", "publication_date": "2025-10-17", "notes": "Long PDF/transcript was opened and only relevant pages/passages were reviewed; claims are limited to surfaced excerpts.", "claims_checked": ["Q3 earnings: Petoro awarded a generative-AI solution to optimize drilling resources using FDPlan and Lumi."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.weatherford.com/documents/sustainability/sustainability-report/", "title": "Weatherford 2025 Sustainability Report: Artificial Intelligence", "publication_date": "2026-04-01", "notes": "The linked PDF/report was opened; relevant excerpt only was reviewed. The assigned packet’s claim-level support statements were checked against the surfaced passage. No full-document review is claimed.", "claims_checked": ["Weatherford reports an AI Impact Hub and agents embedded in internal workflows in 2025.", "Named agents review material test reports, cross-reference parts/inventory, classify spend and simplify bidding; training and human involvement stated."], "access": "excerpt reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.nabors.com/drilling-automation/predictive-drilling/", "title": "Predictive Drilling - Nabors", "publication_date": "Undated; accessed 2026-09-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Nabors reports the Predictive Drilling ROP optimizer combines AI/ML with SmartROS/RigCLOUD for remote autodriller setpoint control.", "First deployment reported 36% higher average ROP, 9.7% lower vibration and 80% lateral footage utilization; historic case source (2023) is outside window."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.hpinc.com/technologies/flexrobotics", "title": "FlexRobotics Hands-Free Rig Floor", "publication_date": "Undated; accessed 2026-09-27", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Three robot arms handle pipe tailing, cleaning/alignment and racking. H&P lists 1:40 slip-to-slip connection times and red-zone personnel exposure reduction.", "System was developed with NOV and uses Yaskawa robot manipulators. Cameras support AI-enabled zone monitoring."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.quantumcap.com/wp-content/uploads/2025/05/QCG_2024_ESG_Report_vF_05.02.2025.pdf", "title": "2024 ESG Report", "publication_date": "2025-05-02", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://collide.io/product", "title": "Collide product: Operations System", "publication_date": "undated", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Collide describes petroleum-specific connected asset model and grounded answers with source citations.", "Company describes deterministic tested calculations and deployment in customer cloud/data platform; page undated."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://collide.io/case-studies/winn-resources-ai-regulatory-filing-oil-gas", "title": "How Winn Resources Slashed Regulatory Filing Time by 95%", "publication_date": "undated", "notes": "URL attempted during this audit and was inaccessible/internal error; do not treat its claims as reverified. Earlier inherited record remains unverified by this pass unless another source supports it.", "claims_checked": [], "access": "inaccessible", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.oneok.com/news-and-events/press-releases/2026/02-23-2026-211624638", "title": "ONEOK Q4 2025 earnings release", "publication_date": "2026", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reviewed alongside company materials; no specific AI workflow found in release."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://ir.kindermorgan.com/news/news-details/2025/Kinder-Morgan-Reports-First-Quarter-2025-Financial-Results/default.aspx", "title": "Kinder Morgan Q1 2025 earnings release", "publication_date": "2025-04-16", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Reviewed as in-window operating/financial source; no Flyscan or internal AI deployment disclosed in this release."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "title": "SLB Q4 and full-year 2025 results", "publication_date": "2026-01-23", "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "claims_checked": ["Q4 report repeats Shell collaboration aims and scope; this is an in-window disclosure/update, not evidence the collaboration was launched in January or deployed."], "access": "full text reviewed", "accessed": "2026-09-27", "partition": 3}, {"url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction", "publication_date": "2025-05-09", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Ambyint says $4.1m from ERA supports a next-generation InfinityPL platform upgrade.", "Company says the existing platform is deployed by 17 producers, and the project is expected to reduce methane venting by up to 90%.", "Tourmaline is named as key planned deployment partner at plunger-lift sites in Alberta."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector", "publication_date": "2025-04-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["ERA confirms $4.1m funding for Ambyint’s upgrade to the InfinityPL AI platform using advanced ML and reinforcement learning.", "ERA describes prediction of inefficiencies, complex-anomaly detection and reduced manual intervention; up to-90% methane-venting reduction remains expected.", "Tourmaline plans deployment across a substantial number of Alberta plunger-lift well sites."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "title": "Infinity PL Drives Efficiency Gains for Encino Energy", "publication_date": "2025-08-05", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Page identifies Encino, Appalachian Basin, 344 platform wells and project timeframe 2023 to present.", "Page describes real-time surveillance, anomaly detection, autonomous setpoint control, dashboard monitoring and operator intervention.", "Page reports decline, setpoint and venting metrics, but does not show cohort sizes or validation method; outcomes remain vendor-reported."], "notes": "Direct page open succeeded in this pass; full case text and outcome section reviewed. This is vendor case evidence, not independent validation.", "partition": 3}, {"url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "title": "Navigating Risk & Harnessing AI Briefing", "publication_date": "2025-10-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Kpler described AI-powered parsing of emails, PDFs and spreadsheets into structured cargo and position lists.", "Kpler previewed an AI chartering agent to recommend vessel candidates based on freight economics, risk and custom filters, and said it was launching Q4."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.kpler.com/solutions/ai", "title": "Kpler AI", "publication_date": null, "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Current Kpler page describes natural-language questions, AI-assisted market analysis, and chartering/cargo/compliance agent categories.", "Page describes sources including proprietary history, AIS, satellite tracking, customs/trade flows and refinery information.", "Page contains vendor-stated capabilities and accuracy/time benefits; publication date and independent results are not provided."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI", "publication_date": "2025-11-27", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["IFS reported Kodiak rolled out a Material Replenisher agent for technicians to find and order parts conversationally.", "CIO’s $3m and 90,000-hour figures are conditional on half the workforce engaging once a day; they are modeled potential, not achieved results."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "title": "IFS Loops launches Agent Studio", "publication_date": "2026-04-23", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["IFS again describes Kodiak Material Replenisher as rolled out; technicians find and order parts through conversation.", "Release repeats potential $3m annual ROI and 90,000 hours; not treated as realized outcomes."], "notes": "Reopened the linked page/document and checked the specific claim statements listed here against the visible source content. This does not imply exhaustive review of other company disclosures.", "partition": 3}, {"source_id": "S11", "url": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm", "access": "full text reviewed", "title": "C3 AI FY2025 Form 10-K", "publisher": "C3 AI", "publication_date": "2025-06-26", "source_type": "SEC regulatory filing", "locator": "FY2025 Form 10-K, strategic partnership discussion", "claims_checked": ["C3 AI says Baker Hughes renewed and expanded the agreement in April 2025; the filing was submitted in June.", "The agreement continues co-selling, co-investment, and scaling joint solutions; C3 AI states Baker Hughes is no longer its exclusive oil-and-gas reseller."], "notes": "Direct SEC filing opened and relevant strategic-partnership and exclusivity text searched. Renewal date is the operating/corporate event; filing publication month is June 2025.", "accessed": "2026-09-27"}, {"source_id": "S36", "url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition", "access": "full text reviewed", "title": "PakEnergy Expands Solutions for Energy Customers With Petrosight Acquisition", "publisher": "PakEnergy", "publication_date": "2025-01-07", "source_type": "Company announcement", "locator": "Acquisition and Charlesbank ownership/support paragraph", "claims_checked": ["PakEnergy announced the Petrosight acquisition on Jan. 7, 2025.", "The source identifies Charlesbank as PakEnergy’s private-equity partner; Petrosight’s wellsite software organizes lifecycle data across drilling, completions, facilities/pipelines, and environmental/reclamation."], "notes": "Full official PakEnergy announcement opened. This supports ownership and deal context only; it is not evidence of AI.", "accessed": "2026-09-27"}, {"source_id": "R016", "url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations", "access": "full text reviewed", "title": "C3 AI and Shell Expand Collaboration, Scaling Reliability AI Deployment Across Global Asset Operations", "publisher": "C3 AI / Business Wire", "publication_date": "2026-06-04", "source_type": "Company announcement distributed by Business Wire", "locator": "Release body, deployment and platform description", "claims_checked": ["C3 AI says Shell’s existing predictive-maintenance program monitors more than 13,000 pieces of equipment.", "The new multi-year agreement extends Reliability and plans to introduce agent-based root-cause analysis and remediation; current predictive maintenance runs on C3 AI Reliability/Agentic AI Platform deployed on Microsoft Azure.", "Hundreds of millions of dollars in value is a C3 AI executive’s attributed statement, not independently quantified in this release."], "notes": "Full Business Wire release opened and key passage reviewed. Distinguish existing program from new planned agentic capabilities; no independent Shell validation located in this source.", "accessed": "2026-09-27"}, {"source_id": "R024", "url": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale", "access": "full text reviewed", "title": "Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI", "publisher": "TotalEnergies", "publication_date": "2025-09-26", "source_type": "Company announcement", "locator": "Partnership deployment scope and objectives", "claims_checked": ["The dated release says TotalEnergies will scale Cognite’s industrial data and AI platform over three years across all operated upstream assets worldwide, drilling through production.", "The intended foundation makes industrial data AI-ready and supports dynamic asset visualization, decision support, critical-equipment monitoring, and later AI application adoption."], "notes": "Full official TotalEnergies release opened after redirect to the newsroom page. It describes a three-year scale-up plan, not completed global deployment at announcement.", "accessed": "2026-09-27"}, {"source_id": "R040", "url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "Devon Q4 2025 Earnings Call Transcript", "publisher": "Devon Energy / The Motley Fool transcription", "publication_date": "2026-02-18", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "Management Q&A, production and AI investment remarks", "claims_checked": ["Management says Devon trialed production-space gas-lift optimization and other projects, while investments in AI and platforms made during the year were beginning to come to fruition in production; projects trialed in 2025 were expected to scale during Q1/Q2 2026.", "The source does not expressly establish that the gas-lift trials themselves were AI-driven, and does not attribute the company’s broader optimization value targets exclusively to AI."], "notes": "Full transcript opened and direct AI passages checked. Secondary transcript; use to narrow language, not to merge gas-lift trials with AI without an explicit linkage.", "accessed": "2026-09-27"}, {"source_id": "R050", "url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/", "access": "full text reviewed", "title": "Occidental Petroleum Q4 2024 Earnings Call Transcript", "publisher": "Occidental Petroleum / The Motley Fool transcription", "publication_date": "2025-02-19", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "CEO remarks on Gulf of America and AI center of excellence", "claims_checked": ["CEO says Gulf of America operations are using AI for supply-chain management, asset integrity, and reservoir characterization, and that Oxy created an AI center of excellence to align intercompany initiatives.", "Oxy says its large Algeria seismic acquisition would provide a rich dataset for future AI applications; this is future use, not a reported current AI result."], "notes": "Full transcript opened; specific AI paragraph reviewed. The transcript supports named use areas but provides no implementation design or AI-attributed outcome.", "accessed": "2026-09-27"}, {"source_id": "R058", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "access": "full text reviewed", "title": "Matador Q1 2026 Earnings Call Transcript", "publisher": "Matador Resources / The Motley Fool transcription", "publication_date": "2026-05-07", "source_type": "Earnings-call transcript (secondary transcription)", "locator": "Management Q&A on AI operations, production control room, completions, and lateral placement", "claims_checked": ["Management describes AI-driven analytics across operations; a control room and field staff monitor more than 40 million daily data points in real time to identify downtime and inefficiencies.", "Management says real-time frac monitoring covers water/sand pressures, volumes, and logistics; drilling AI helps target the preferred portion of the lateral target.", "The same answer describes a committee-led cautious adoption approach. Operational gains cited elsewhere in the call are not assigned exclusively to AI."], "notes": "Full transcript opened; Q&A detail reviewed. This materially expands generic “cautious integration” wording; no model vendor, model architecture, or AI-only economic result is disclosed.", "accessed": "2026-09-27"}, {"source_id": "R079", "url": "https://c3.ai/c3-ai-and-baker-hughes-renew-and-expand-joint-venture-agreement/", "access": "full text reviewed", "title": "C3 AI and Baker Hughes Renew and Expand Joint Venture Agreement", "publisher": "C3 AI", "publication_date": "2025-05-28", "source_type": "Company announcement", "locator": "Renewal term, deployed customers, Baker Hughes internal-use applications", "claims_checked": ["C3 AI and Baker Hughes renewed and expanded the JV through June 2028 to develop, deliver, and market AI solutions to oil/gas and chemicals.", "C3 AI lists deployed solutions at Shell, Eni, QatarEnergy LNG, Petronas, ExxonMobil, LyondellBasell, and Flint Hills Resources; it describes customer benefits generally, without customer-specific metrics.", "Baker Hughes plans to expand internal use of sourcing optimization, inventory optimization, and sustainability-suite applications."], "notes": "Full issuer announcement opened. Deployment/customer and economic claims are vendor-reported and aggregate; do not overstate them as independently confirmed site-level outcomes.", "accessed": "2026-09-27"}, {"source_id": "R084", "url": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/", "access": "full text reviewed", "title": "Nabors, Corva AI Join to Advance Digital Transformation in Drilling Industry", "publisher": "World Oil (Nabors/Corva announcement coverage)", "publication_date": "2025-04-11", "source_type": "Industry trade publication", "locator": "RigCLOUD Powered by Corva product and integration description", "claims_checked": ["The announcement says RigCLOUD Powered by Corva would combine Nabors edge/cloud computing and rig instrumentation/apps with Corva AI analytics for real-time data processing, predictive insight, and drilling performance optimization.", "The release language describes a solution set to be integrated and capabilities intended for customers; this story does not establish fleetwide deployment or measured field results."], "notes": "Full article opened. Trade source republishes/relays vendor announcements; treat deployment and expected-benefit wording as partner-reported and planned.", "accessed": "2026-09-27"}, {"source_id": "R102", "url": "https://energycapitalhtx.com/collide-riggs-ai-platform", "access": "full text reviewed", "title": "Houston AI Startup Rolls Out Platform to Reshape Oil and Gas Workflows", "publisher": "Energy Capital HTX", "publication_date": "2026-05-07", "source_type": "Secondary interview/article", "locator": "RIGGS product, Winn Resources RRC filings workflow", "claims_checked": ["Article reports that Collide and Winn Resources collaborated on monthly Texas Railroad Commission W-10/G-10 filings; Collide reported processing over 95% faster, with an example of 50 wells completed in 20 minutes.", "Winn’s operations director is quoted describing a human-in-the-loop workflow; article says reservoir analysis, production optimization, and workover report generation were future/near-term targets, not yet documented deployments."], "notes": "Full article opened. Implementation/result is attributed to Collide and Winn through the article; no primary Winn release was part of this 10-source supplemental assignment.", "accessed": "2026-09-27"}, {"url": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497", "access": "full text reviewed", "title": "FY2026 Form 10-K", "publication_date": "2026-06-24 (Form 10-K; FY ended 2026-04-30)", "claims_checked": ["Page 35 says Baker Hughes ceased being C3 AI’s exclusive oil-and-gas channel/reseller after amendments in January 2023 and April 2025.", "Filing still describes a strategic relationship; it does not say the partnership ended."], "notes": "Primary C3 AI filing directly opened and claim passage located."}, {"url": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center", "access": "inaccessible", "title": "Microsoft power agreement for Project Kilby", "publication_date": "2026-06-22", "claims_checked": ["Attempted Chevron’s direct announcement URL; web fetch returned HTTP 403."], "notes": "Not reverified from the cited page; this does not disprove the agreement. No claim promoted to verified from access metadata."}, {"url": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts", "access": "full text reviewed", "title": "SLB and Shell co-develop agentic AI for upstream", "publication_date": "2025-12-18", "claims_checked": ["SLB and Shell signed a strategic collaboration to develop AI/agentic AI for technical experts; planned objective is to unify data/workflows across subsurface, well construction and production, using Lumi."], "notes": "JPT/SPE report is secondary editorial confirmation; it says aim/develop/deploy, not that a production system is already deployed."}, {"url": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml", "access": "full text reviewed", "title": "Copilot to 25,000 staff; pilot saved 121 min/week", "publication_date": "Undated current Repsol webpage; includes historic 2023 launch and 4-month pilot results", "claims_checked": ["Copilot for Web extended to around 25,000 employees; Microsoft 365 Copilot licenses are being progressively scaled.", "Four-month M365 Copilot pilot involved 550 professionals and reported average 121 minutes saved weekly per participant, 16% higher work quality and >60% preference to continue.", "More than 60 use cases are “in process”; this page does not support interpreting that count as 60+ in-production cases or 22 scaled cases."], "notes": "Primary Repsol webpage opened. Metrics are company-reported pilot results; time savings are not a realized labor-cost reduction."}, {"url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/", "access": "full text reviewed", "title": "Smart Gas Lift closed-loop AI scaled to 850+ wells", "publication_date": "2026-05-06", "claims_checked": ["Transcript says Smart Gas Lift uses AI models and physics-based calculations to optimize gas injection on a closed loop directly to wells.", "Management says pilot in 2025 yielded about 2–3% uplift; by Q1 2026 full implementation in Delaware Basin covered >850 wells; current uplift was said to exceed pilot but not quantified; path to 1,500 wells described."], "notes": "Motley Fool earnings transcript is a secondary transcription of management remarks, not an issuer-published transcript. All operational metrics remain management-reported and secondary-transcribed."}, {"url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "access": "full text reviewed", "title": "AI subsurface project targets complex Gulf reservoirs", "publication_date": "2025-08-07", "claims_checked": ["Oxy CEO said focused AI teams were working in Gulf of America, operations, and logistics/supply chain.", "Management said the complex Gulf subsurface project was still being developed and expected to be ready to start execution in the following one to two years; AI was already helping subsurface work in the Permian and other onshore areas."], "notes": "Motley Fool transcript is secondary. Keep Gulf project future/planning stage distinct from existing Permian/onshore assistance."}, {"url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/", "access": "full text reviewed", "title": "Autonomous haul fleet nears 140 trucks", "publication_date": "2025-07-04", "claims_checked": ["Article says nearly 120 autonomous haul trucks at Suncor Base Plant, with all ore moved by driverless autonomous-haulage systems; operation began in 2018.", "Describes 15 trucks added at Millennium in May 2024, reaching 91 by year-end; further 2025 deployment, control-room oversight, obstacle/collision detection and workflow benefits.", "The article does not establish that AI/ML powers the autonomous haulage system and does not support a 140-truck planned target."], "notes": "International Mining is secondary. Autonomy is confirmed; AI basis and any 140-truck target require separate primary evidence."}, {"url": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial", "access": "inaccessible", "title": "Cordant built on Azure AI Foundry with Microsoft", "publication_date": "2025-02 (per source metadata)", "claims_checked": ["Attempted Baker Hughes announcement URL; web fetch returned HTTP 403."], "notes": "Direct vendor source not accessible during this pass; claims remain unverified here, not disproved."}, {"url": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results", "access": "full text reviewed", "title": "FlexRobotics rig-floor robotics deployed for supermajor", "publication_date": "2026-02-04 (quarter ended 2025-12-31)", "claims_checked": ["H&P reported FlexRobotics deployed on a Permian rig for a supermajor and supporting multiple pads; described reliable rig-floor performance and customer interest in additional units.", "The earnings release does not say that the deployed robotic equipment itself uses AI. H&P’s product page separately describes a camera-based AI zone-monitoring add-on; the cited earnings release does not establish that this add-on was deployed with the customer."], "notes": "Business Wire is the issuer-distributed earnings release. Separately reopened H&P product page: it calls FlexRobotics rig-floor automation and lists AI-enabled zone monitoring under an add-on camera-support feature."}, {"baseline_id": "AI-058", "source_id": "S35", "url": "https://www.prnewswire.com/news-releases/pakenergy-launches-pakcapture-delivering-ai-powered-ticket--receipt-management-to-pakenergy-transportation-302777729.html", "title": "PakCAPTURE launch", "publication_date": "2026-05-20", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["PakEnergy-provided release says drivers photograph tickets/receipts at pickup or delivery; PakAI extracts data, populates forms, and associates the record with a load for office processing.", "The announcement calls PakCAPTURE immediately available as part of PakEnergy Transportation; it reports no measured accuracy, time savings, or cash-flow results."], "notes": "Issuer release distributed by PR Newswire. Consistent with AI-058; supports launch capability, not realized outcomes. The baseline also links PakEnergy's own copy and another source, so this does not require a new profile."}, {"baseline_id": "AI-074", "source_id": "R013", "url": "https://www.marketscreener.com/news/chevron-a-smarter-way-to-prospect-for-oil-and-gas-ce7d5fdfdd81f02c", "title": "Chevron: A smarter way to prospect for oil and gas", "publication_date": "2025-11-12", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Chevron's article says ApEX launched in August 2024 and was first used in Gulf of America deepwater exploration; users query a chat interface and specialized agents search more than one million files.", "The article reports a 4 TB corpus and describes seconds-fast retrieval. It does not quantify a reduction from weeks of pre-drill studies to minutes."], "notes": "The supplied MarketScreener page is Chevron-authored copy redistributed via Publicnow. The same article was located and reviewed on Chevron's own newsroom page, which is the preferred primary citation. The supplied reference's December month is not the publication date; the original article is dated November 12, 2025."}, {"baseline_id": "AI-029", "source_id": "R022", "url": "https://totalenergies.com/news/press-releases/totalenergies-collaborate-mistral-ai-increase-application-artificial", "title": "TotalEnergies to Collaborate with Mistral AI to Increase the Application of Artificial Intelligence", "publication_date": "2025-06-12", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["TotalEnergies announced a shared AI lab and named an assistant for its 1,000 researchers, industrial-asset decision support, customer energy support, and renewable-production use cases.", "The release describes work to design/develop use cases; it does not report production rollout or measured performance."], "notes": "Official TotalEnergies press release. Supports the existing AI-029 planned-use description and the stated researcher cohort; it does not substantiate realized impacts."}, {"baseline_id": "AI-031", "source_id": "R029", "url": "https://newsroom.accenture.com/news/2025/repsol-boosts-its-digital-strategy-by-collaborating-with-accenture-to-adopt-ai-agents", "title": "Repsol Boosts its Digital Strategy by Collaborating with Accenture to Adopt AI Agents", "publication_date": "2025-02-06", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Accenture's release describes Repsol's planned co-innovation and deployment of custom AI agents using Accenture and NVIDIA platforms, across processes including planning/forecasting, application maintenance, and incident resolution.", "It also says Repsol would explore agents with digital twins and robotics for industrial/logistics maintenance and expand staff training; these are forward-looking plans, not measured operating outcomes."], "notes": "Partner-issued primary announcement; Repsol also published its own release. This is an earlier collaboration/update and is consistent with AI-031's later, separately sourced deployment counts."}, {"baseline_id": "AI-007", "source_id": "R043", "url": "https://drillingcontractor.org/post-merger-devon-sees-big-role-for-ai-as-it-tackles-integration-seeks-1-billion-in-synergies-78873", "title": "Post-merger Devon sees big role for AI as it tackles integration, seeks $1 billion in synergies", "publication_date": "2026-07/08 issue (article page does not expose a day in the archive listing)", "access": "excerpt reviewed", "accessed": "2026-09-27", "claims_checked": ["The IADC Drilling Contractor July/August 2026 archive lists the article title and author Jessica Whiteside.", "The specific article URL returned a fetch/cache error; the listing does not expose the article body or substantiate the separate imported description 'AI agents for well design and completions sequencing'."], "notes": "Treat the cited source as not reverified for the detailed workflow, not as disproving the earlier Q4 earnings-presentation account. The archive article title concerns post-merger integration/synergies, so the source title/summary mismatch merits reconciliation before using it as evidence for well-design agents."}, {"baseline_id": "AI-091", "source_id": "R052", "url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/occidental-oxy-q1-2026-earnings-transcript/", "title": "Occidental (OXY) Q1 2026 Earnings Transcript", "publication_date": "2026-05-06", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["The transcript attributes to COO Richard Jackson a statement that Oxy was reviewing workforce processes and priorities, including technologies such as AI.", "The transcript attributes to CEO Vicki Hollub the statement that AI was taking a larger role across company disciplines as part of workforce efficiency.", "The Q1 transcript does not substantiate the profile's $4 billion incremental cash-flow target by 2030; no AI-specific productivity or savings result is quantified."], "notes": "Motley Fool is a third-party earnings-call transcript, not an issuer-authored source. The corresponding Occidental-hosted transcript PDF was also located and reviewed at https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf (excerpt/full document text available); the management statements are consistent. Keep the source type as transcript and distinguish the speaker's remarks from independently measured outcomes."}, {"baseline_id": "AI-104", "source_id": "R068", "url": "https://www.prnewswire.com/news-releases/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-302350209.html", "title": "Flyscan Systems Announces that Kinder Morgan Joins its Strategic Investor Group alongside Marathon Petroleum, Enbridge and Hatch", "publication_date": "2025-01-14", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["Flyscan says Kinder Morgan joined its strategic investor group; the funding was to accelerate commercialization and entry into natural-gas detection/quantification.", "Flyscan describes remote sensing, AI, liquid-hydrocarbon detection, and automated right-of-way visual inspection from routine patrol aircraft. KMI's COO endorses the technology and says the parties will partner.", "The announcement does not establish a KMI operational rollout or quantified leak-detection result."], "notes": "Issuer Flyscan announcement distributed by PR Newswire; Flyscan's official newsroom copy was also reviewed. The date in the baseline initiative's Source_date (2025-04-16) is the date of KMI's Q1 earnings release, not the cited Flyscan announcement; correct the initiative source date to 2025-01-14."}, {"baseline_id": "AI-113", "source_id": "R082", "url": "https://www.globenewswire.com/news-release/2025/10/01/3159510/0/en/weatherford-launches-industrial-intelligence-at-fwrd-2025.html", "title": "Weatherford Launches Industrial Intelligence at FWRD 2025", "publication_date": "2025-10-01", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["The supplied GlobeNewswire URL could not be fetched in this pass, but Weatherford's investor-relations copy of the same release was opened and reviewed.", "The issuer release describes a product portfolio and names AI/physics-based MetaFlow flow estimation, AI tools for document digitization/transcription/image recognition/translation in Engineer's Briefcase, and other data/edge components.", "The launch announcement describes offered products, not a customer's deployment or measured results."], "notes": "A broken syndication URL was replaced with the issuer-hosted copy at https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18126. Do not treat the release date as a deployment date."}, {"baseline_id": "AI-049", "source_id": "R092", "url": "https://worldoil.com/news/2025/9/16/aker-bp-expands-ai-first-strategy-with-cognite-to-transform-e-p-efficiency/", "title": "Aker BP expands AI-first strategy with Cognite Atlas AI to transform E&P efficiency", "publication_date": "2025-09-16", "access": "full text reviewed", "accessed": "2026-09-27", "claims_checked": ["World Oil describes expansion of Aker BP's Cognite Atlas AI use, prebuilt agents/low-code tooling, document processing, barrier management and equipment registration.", "It attributes to Aker BP a claim that AI-agent root-cause analysis reduced engineering time by more than 70%; the article supplies no sample/baseline, and that result remains company/partner reported.", "The baseline also links Cognite's partner announcement, which is the more direct source for the initiative."], "notes": "Secondary trade-publication coverage, not issuer-primary. Keep the 70% result explicitly attributed and unbenchmarked; the added source does not independently validate it."}, {"url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "access": "full text reviewed", "claims_checked": ["Leucipa AI-powered production management, field optimization and forecasting; Lucy described as a planned pilot with GenAI chat interface and real-time production insights; deployment beginning at Operations Support Center.", "Snowflake data platform adoption; ML model forecasting; real-time drilling data/models hosted; SCADA signals used to detect issues and alert Operations Support Center."], "notes": "The issuer’s announcement supports both packets but distinguishes an operating platform/deployment from the planned Lucy pilot.", "title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "publication_date": "2026-03-03"}, {"url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "access": "full text reviewed", "claims_checked": ["SLB/Qualcomm edge-AI collaboration and product/workflow development scope."], "notes": "Same exact initiative and announcement as assigned existing AI-046; a second addition would duplicate the existing record.", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "publication_date": "2026-06-09"}, {"url": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "access": "full text reviewed", "claims_checked": ["Evergreen Data Solution product launch and AI-assisted classification, curation, contextualization and validation of subsurface data."], "notes": "Commercial product capability, not customer deployment proof.", "title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "publication_date": "2026-06-10"}, {"url": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "access": "full text reviewed", "claims_checked": ["TGS multi-year AI seismic foundation-model contract award to an unnamed international operator; scope includes structural and stratigraphic interpretation and TGS multi-client data."], "notes": "Customer remains unnamed; no contract value or field result disclosed.", "title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "publication_date": "2026-08-18"}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "access": "excerpt reviewed", "claims_checked": ["SLB Q1 2025 earnings release lines 285–286: Woodside award for integrated delivery of 18 Trion ultra-deepwater wells; AI-enabled drilling capabilities included; service start scheduled early 2026 and delivery via Performance Live."], "notes": "The cited release supports the packet’s timing as a plan and the contracted scope; it does not prove services commenced or outcomes were achieved.", "title": "SLB Announces First-Quarter 2025 Results", "publication_date": "2025-04-25"}, {"url": "https://investors.technipenergies.com/shareholder-information", "access": "excerpt reviewed", "claims_checked": ["2025 shareholder/management messages say DAP leverages advanced analytics and large language models; DAP launched in 2025; target €100m cost savings and incremental revenues by 2028; AI for All implemented."], "notes": "The €100m is a 2028 target, not realized savings. The source supplies no named AI workflow or AI-specific attribution for savings.", "title": "Technip Energies shareholder information — 2025 chair and CEO messages", "publication_date": "Undated; see individual source review"}, {"url": "https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "access": "full text reviewed", "publication_date": "2026-05-22", "notes": "Official KMG Russian-language release. Confirms 2025 pilots; 2026 rollout plans, Shymkent phase, and polypropylene plant 10 units.", "title": "КМГ внедряет предиктивную аналитику на НПЗ"}, {"url": "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/", "access": "full text reviewed", "publication_date": "2025-12-25", "notes": "Official Atyrau refinery Russian-language release. October pilot start, Honeywell, equipment examples, operational anomaly cases, and 28-unit 2026 target reviewed.", "title": "AI control of equipment at Atyrau refinery: predictive analytics in action"}, {"url": "https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf", "access": "full text reviewed", "publication_date": "2025-07 (Issue 167)", "notes": "Primary company magazine, complete PDF opened; page 5 strategy initiative reviewed.", "title": "The Pioneer 167 — QatarEnergy LNG"}, {"url": "https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "access": "full text reviewed", "publication_date": "2025-05-09", "notes": "Official Ambyint release says $4.1m, 17 existing producer deployments, up-to-90% expected venting reduction, Tourmaline planned deployment, and new capabilities.", "title": "Ambyint Secures $4.1M to Advance AI for Methane Emission Reduction"}, {"url": "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/", "access": "full text reviewed", "publication_date": "2025-04-16", "notes": "Funder primary release confirms $4.1m, advanced ML/reinforcement learning, expected up-to-90% reduction, and planned Tourmaline deployment. Does not state exact $4.08m or setpoint-management detail.", "title": "Alberta Invests $7 Million to Tackle Methane Emissions in the Oil and Gas Sector"}, {"url": "https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "access": "full text reviewed", "publication_date": "2025-10-16 (event page)", "notes": "Official Kpler event content directly states structured email/PDF/spreadsheet parsing and chartering-agent Q4 plan.", "title": "Navigating Risk & Harnessing AI Briefing"}, {"url": "https://www.kpler.com/solutions/ai", "access": "full text reviewed", "publication_date": "undated current page", "notes": "Official current product page reviewed and searched for accuracy and 99%; general AI/data features appear, but no email-parsing accuracy metric was located.", "title": "Kpler AI"}, {"url": "https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "access": "full text reviewed", "publication_date": "2025-11-27", "notes": "IFS release directly quotes the conditional half-workforce daily-use scenario; does not establish deployment by this date.", "title": "IFS Cloud 25R2: Unlocking 10X Industrial Workforce Capacity with Digital Workers and Applied Industrial AI"}, {"url": "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio", "access": "full text reviewed", "publication_date": "2026-04-23", "notes": "IFS says Kodiak rolled out Material Replenisher and describes parts requests conversationally; calls $3m and 90k hours the result. Vendor claim, not independently validated.", "title": "IFS Loops launches Agent Studio"}, {"url": "https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf", "access": "full text reviewed", "title": "Q3 2025 Earnings Presentation", "publication_date": "2025-11-05", "claims_checked": ["AI-driven subsurface insight and precision wellbore placement on printed slide 5; model and scale not disclosed; Haley results separately presented."], "notes": "Independent lead-agent source check; public company-reported description, not measured AI return."}], "profile_patches": [{"id": "AI-060", "fields": {"Research_category": "Excluded / unverified AI claim", "Research_verification": "unable to verify", "Audit_flags": [{"claim": "2024 Palantir and Infosys refining digital implementation is established AI workflow.", "reason": "The in-window Q3 2024 call establishes digital/ontology progress but names no AI workflow; the Palantir AI relationship announcement is pre-window. Exclude from confirmed AI adoption absent a qualifying AI-specific source.", "source_url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter"}]}, "reason": "The source supports digital transformation, not the inherited AI-adoption classification.", "source_url": "https://www.roic.ai/quote/BP-B.L/transcripts/2024-year/3-quarter"}, {"id": "AI-039", "fields": {"Stage_at_disclosure": "Non-binding term sheet announced; completion subject to definitive agreement and approvals."}, "reason": "Primary Aramco source makes status conditional.", "source_url": "https://www.aramco.com/en/news-media/news/2025/pif-and-aramco-agree-for-aramco-to-acquire-a-significant-minority-stake-in-humain"}, {"id": "AI-059", "fields": {"Theme": "AI-demand infrastructure"}, "reason": "Hyperscale campus supplies infrastructure serving AI demand; no operator AI workflow.", "source_url": "https://www.bridgepointgroup.com/about-us/news-and-insights/press-releases/2025/energy-capital-partners-ecp-and-kkr-announce-development-of-hyperscale-data-centre-campus-in-texas"}, {"id": "AI-141", "fields": {"Source_date": "Undated; accessed 2026-09-27"}, "reason": "Do not present page access date as disclosure date.", "source_url": "https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data"}, {"id": "AI-151", "fields": {"Segment": "Geothermal energy (adjacent energy; not oil and gas)"}, "reason": "PGE source is for geothermal operations.", "source_url": "https://pge.pertamina.com/en/press-release/pge-demonstrates-digital-transformation-boosts-national-geothermal-efficiency-and-competitiveness"}, {"id": "AI-177", "fields": {"Stage_at_disclosure": "Announced/development-stage process-modeling system; future-tense benefits, facility deployment not established"}, "reason": "LEAP remarks do not establish achieved plant deployment.", "source_url": "https://singapore.aramco.com/en/news-media/news/2025/remarks-by-ahmad-al-khowaiter-at-leap-2025"}, {"id": "AI-037", "fields": {"Reported_metric": "2025 Technology Realized Value: $5.3bn total; Aramco CEO said over half was AI-enabled (company-reported aggregate attribution; no project-level allocation disclosed)."}, "reason": "CEO remarks explicitly attribute over half of the total TRV to AI; preserve the company-reported proportion with qualification.", "source_url": "https://www.aramco.com/en/investors/ceo-fy-2025-media-call-remarks"}, {"id": "AI-109", "fields": {"Research_category": "Automation / AI not established", "Theme": "Autonomous hydraulic fracturing automation; AI method not established"}, "reason": "Halliburton and World Oil describe autonomous/intelligent automation but not AI/ML.", "source_url": "https://www.halliburton.com/en/about-us/press-release/coterra-energy-halliburton-launch-first-fully-automated-hydraulic-fracturing-program"}, {"id": "AI-110", "fields": {"Research_category": "Automation / AI not established", "Theme": "Closed-loop hydraulic fracturing automation; AI method not established"}, "reason": "Halliburton release describes feedback-driven autonomous workflows and a domain algorithm without identifying AI/ML.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/chevron-and-halliburton-enable-intelligent-hydraulic-fracturing"}, {"id": "AI-121", "fields": {"Research_category": "Adjacent process industry (chemicals)", "Segment": "Industrial AI — chemicals (adjacent to oil and gas)"}, "reason": "The use case is at Celanese, a chemical manufacturer, rather than an oil-and-gas operator.", "source_url": "https://www.cognite.com/en/company/newsroom/cognite-to-unleash-next-gen-industrial-ai-to-unlock-new-levels-of-intelligence-innovation-and"}, {"id": "AI-157", "fields": {"Research_category": "Automation / AI not established", "Theme": "Drilling automation; AI method not established"}, "reason": "The report supports automated drilling controls but not an AI method.", "source_url": "https://ir.halliburton.com/news-releases/news-release-details/halliburton-and-sekal-deliver-worlds-first-automated-bottom"}, {"id": "AI-099", "fields": {"Research_category": "Automation / AI not established", "Theme": "Autonomous haulage automation; AI method not established"}, "reason": "The supplemental Suncor source documents autonomous trucks and control-room oversight but does not establish AI/ML; a separate Suncor dynamic-dispatch workflow should remain separate from exact AI-099.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}, {"id": "AI-044", "fields": {"Theme": "Operational AI", "Initiative": "AI-assisted subsurface work and Gulf of America operational use cases", "Extracted_finding": "Oxy management described current AI use in Gulf of America supply chain, asset integrity and reservoir characterization, and AI assistance in Permian/onshore subsurface work. A separate complex Gulf subsurface project was still under development; no AI-specific outcome was reported.", "Stage_at_disclosure": "Management-reported current use in selected Gulf and onshore work; complex Gulf subsurface project in development in 2025, with execution expected in 1–2 years.", "Reported_metric": "No AI-specific quantified outcome disclosed", "Interpretation_limit": "Evidence is management statements in third-party earnings-call transcripts. Do not merge current onshore/Gulf uses with the future complex Gulf project or attribute operating reliability/results solely to AI.", "Context": "Oxy described AI use in selected operating functions and subsurface work, with a coordinated AI center of excellence; complex Gulf subsurface application remained a forward-looking project.", "Implementation": "Management named AI in Gulf of America supply-chain management, asset integrity and reservoir characterization, and said AI was already helping subsurface work in the Permian and other onshore areas. The CEO described the complex Gulf subsurface project as under development, with execution expected in the following one to two years. Model, data, human-review and implementation details were not disclosed.", "Assessment": "The transcripts support current management-reported use cases plus a separate planned Gulf subsurface project. They do not provide implementation detail, deployment scale, or AI-specific performance/economics.", "Source_IDs": "R050; R051; S25", "Source_date": "2025-02-19; 2025-08-07; 2026-05-06 (management statements)", "Source_type": "Earnings-call transcripts (third-party transcriptions of management remarks)", "Source_locator": "Q4 2024 call: AI CoE and Gulf use cases; Q2 2025 call: onshore AI assistance and future Gulf project; Q1 2026 call: broader company AI role", "Source_URL": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/", "Evidence_status": "Management-reported transcript evidence reviewed; operational details and outcomes not independently validated", "Review_tier": "Reviewed transcripts; secondary renderings", "Time_scope": "Q4 2024 and Q2 2025 operating claims; Q1 2026 strategic update. The planned Gulf project is not reported as deployed."}, "reason": "Supplemental Q4 2024 and Q2 2025 transcript review supplies specific current use cases and separates them from the future complex Gulf project.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}, {"id": "AI-116", "fields": {"Theme": "Commercial AI product capability / operational automation", "Initiative": "FlexRobotics rig-floor automation with an AI-enabled monitoring add-on", "Extracted_finding": "H&P reports a customer deployment of FlexRobotics rig-floor automation. H&P separately offers AI-enabled zone monitoring as an add-on camera feature; the reviewed sources do not confirm that the deployed customer rig used that AI add-on.", "Research_category": "Commercial AI product", "Stage_at_disclosure": "Base FlexRobotics automation deployed on a Permian rig; AI-enabled zone-monitoring add-on is described in product documentation, but customer deployment is unconfirmed.", "Implementation": "Three robotic arms automate pipe tailing, cleaning/dope/alignment and racking. The separate product page lists AI-enabled zone monitoring as an add-on camera capability; deployment of that feature on the reported customer rig is not established.", "Reported_metric": "No quantified customer result disclosed. Product page lists 1:40 slip-to-slip connection time as a product capability, not a verified result for the reported deployment.", "Interpretation_limit": "Do not attribute AI to the deployed pipe-handling workflow or imply the AI camera feature was installed at the reported customer rig. The documented customer deployment is automation; AI is an available ancillary product capability.", "Assessment": "The deployment evidence is for industrial robotics/automation. The AI claim belongs to an ancillary commercial camera add-on and must remain separate from verified customer deployment.", "Disclosures": [{"ID": "REF-085", "Initiative_ID": "AI-116", "Title": "FlexRobotics deployment and separate AI-enabled monitoring add-on", "Date": "2026-02-04", "Finding": "H&P reported FlexRobotics rig-floor automation deployed on a Permian rig for a supermajor across multiple pads. Separately, H&P product documentation lists AI-enabled zone monitoring as an add-on camera feature; it does not establish that this add-on was used on the reported customer deployment.", "Metric": "No quantified customer result disclosed; 1:40 connection time is a product-page capability statement, not deployment-specific", "Source_IDs": "RS056; RS057; R085", "Evidence": "Primary issuer earnings release and product documentation reviewed; AI add-on deployment unconfirmed", "Origin": "Issuer earnings release and H&P product page", "Partners": "NOV; Yaskawa", "Related_IDs": []}]}, "reason": "Separates confirmed base robotics deployment from the AI-enabled add-on capability that is not confirmed at the deployed site.", "source_url": "https://www.hpinc.com/technologies/flexrobotics"}, {"id": "AI-104", "fields": {"Source_date": "2025-01-14", "Source_type": "partner announcement", "Time_scope": "Dated source: 2025-01-14; strategic investment announced, KMI operating rollout not established."}, "reason": "The cited Flyscan announcement is dated January 14, 2025; April 16 was the date of a separate KMI earnings release.", "source_url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/"}, {"id": "AI-096", "fields": {"Theme": "Operational AI", "Initiative": "Cross-functional AI program for operations analytics, frac monitoring and geosteering", "Extracted_finding": "The Q1 2026 call names a cross-functional AI committee and practical operational applications: analytics alerts intended to reduce downtime, real-time frac monitoring and AI-assisted lateral targeting. The call supplies operating context and data scale but does not isolate AI-attributed performance.", "Context": "Matador describes using operational data and AI in drilling/completions and lateral targeting, with a cross-functional committee to coordinate adoption.", "Implementation": "Management described AI analytics that surface actionable alerts to reduce downtime; real-time frac monitoring of pressures, water and sand volumes and logistics as simultaneous/trimul-frac operations expand; and AI-assisted geosteering to remain in the target zone. The transcript cites about 40 million data points per day and more than 36 drilling records in Q1 2026, but neither figure is an AI outcome.", "Assessment": "This is more than a vendor-only strategy statement: management names operating workflows and coordination. Deployment counts, model details and measured AI-specific improvement remain undisclosed.", "Reported_metric": "About 40 million operational data points per day; 36+ drilling records in Q1 2026 (company-reported operating context, not an AI-attributed result)", "Stage_at_disclosure": "Q1 2026 call describes workflows in use/underway and a cross-functional AI committee; deployment scale and individual workflow maturity are not quantified.", "Interpretation_limit": "Transcript is a third-party rendering of management remarks. Data/record counts do not establish AI contribution or performance; no controlled result is reported.", "Source_IDs": "RS039; R057; R058", "Source_date": "2026-05-07", "Source_type": "earnings-call transcript (third-party transcription)", "Source_URL": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/", "Source_locator": "Q1 2026 call: AI committee, real-time operations data, frac monitoring and geosteering", "Evidence_status": "Management-reported transcript reviewed; workflows named, outcomes not independently validated", "Review_tier": "Reviewed source; third-party transcript", "Time_scope": "Q1 2026 management remarks; exact individual workflow start dates not disclosed.", "Disclosures": [{"ID": "REF-058", "Initiative_ID": "AI-096", "Title": "AI committee and operating workflows in frac monitoring and geosteering", "Date": "2026-05-07", "Finding": "Management described a cross-functional committee for controlled AI adoption, analytics alerts intended to reduce downtime, real-time frac monitoring of pressure/water/sand/logistics, and AI-assisted lateral targeting. The call cited about 40 million data points per day and 36+ drilling records in Q1; those are context/operating-scale figures, not AI-attributed results.", "Metric": "About 40 million data points/day; more than 36 drilling records in Q1 2026 (company-reported, not AI-specific performance)", "Source_IDs": "R058", "Evidence": "Third-party transcript of company management remarks; reviewed", "Origin": "Q1 2026 earnings-call transcript", "Supplied_interpretation": "Operational applications are named, but contribution and realized performance are not isolated.", "Partners": "MaxCom (real-time operations center)", "Related_IDs": []}]}, "reason": "Fresh Q1 2026 transcript review identifies named workflows, superseding the earlier stale strategy-only characterization. Metrics remain company-reported context, not AI outcomes.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}, {"id": "AI-099", "fields": {"Theme": "Operational automation", "Research_category": "Automation / AI not established", "Initiative": "Autonomous haulage fleet expansion at Base Plant (AI not established)", "Extracted_finding": "Suncor describes a large operating autonomous haul fleet and control-room oversight. The source supports autonomy/automation but does not establish that AI is used.", "Context": "Suncor moves oil-sands material across Base Plant mine routes using an Autonomous Haulage System.", "Implementation": "Suncor says the driverless haul trucks use GPS, wireless/LTE communications and perceptive technologies. AHS software gives control-room operators real-time visibility so they can manage traffic flow and adjust to conditions. The page does not identify an AI model or say AI optimizes driving conditions.", "Assessment": "The page supports a substantial autonomy deployment and human oversight, not an AI adoption claim. Nearly 120 is the reported snapshot; 140 is a future target whose year cannot be established because the page is undated.", "Reported_metric": "Autonomous haul trucks: almost 120 across two integrated mines at the undated page snapshot; 140 equipped trucks planned by an unspecified year-end (Company-reported)", "Stage_at_disclosure": "Autonomous haulage operating; page snapshot is undated; 140 is prospective", "Interpretation_limit": "The page does not mention AI. It gives no publication date, so the snapshot and prospective target cannot be assigned to 2025 or 2026. Do not treat the target as deployed. Keep dynamic dispatch as a separate record.", "Source_type": "undated company feature", "Source_date": "Undated; accessed 2026-09-27", "Source_locator": "Base Plant fleet, command centre and end-of-page target statement", "Evidence_status": "Verified primary for autonomous automation; AI not established", "Time_scope": "Undated page accessed 2026-09-27; an in-window publication date is not established.", "Disclosures": [{"ID": "REF-062", "Initiative_ID": "AI-099", "Title": "Autonomous haulage fleet snapshot and prospective target", "Date": null, "Finding": "Suncor describes an autonomous haul fleet operating at Base Plant, with almost 120 trucks across two integrated mines in the current page snapshot and a plan for 140 equipped trucks by an unspecified year-end. It identifies GPS, wireless/LTE and perceptive technologies and central control-room oversight. The page does not establish an AI component or support an AI-specific optimization claim.", "Metric": "Almost 120 operating autonomous haul trucks in the undated page snapshot; 140 planned by an unspecified year-end", "Source_IDs": "R062", "Evidence": "Primary source reviewed; automation supported, AI not established", "Origin": "User-supplied reference, corrected against source", "Supplied_interpretation": "Autonomous fleet scale and workforce implications require separate evidence; this page does not quantify job impacts.", "Partners": "Komatsu trucks; source does not establish a named AHS software vendor", "Related_IDs": []}]}, "reason": "Direct source review establishes autonomous haulage and control-room operations but contains no AI claim. Corrects the site-facing classification, removes unsupported AI language and date/job-impact inference, and preserves the supported scale and prospective target.", "source_url": "https://www.suncor.com/en-ca/news-and-stories/our-stories/the-future-of-canadian-mining-is-autonomous"}, {"id": "AI-074", "fields": {"Source_date": "2025-11-12", "Source_type": "Chevron newsroom article", "Disclosures": [{"ID": "REF-013", "Initiative_ID": "AI-074", "Title": "ApEX multi-agent GenAI searches more than one million exploration files", "Date": "2025-11-12", "Finding": "Chevron says its ApEX generative-AI platform launched in August 2024 and was first used in Gulf of America deepwater exploration. Explorationists query a chat interface; multiple specialized search agents search more than one million files. Chevron reports a 4 TB corpus and says results surface in seconds.", "Metric": "More than 1 million files; 4 TB corpus; search responses in seconds (company-reported)", "Source_IDs": "R013; RS001", "Evidence": "Verified primary; company-reported capabilities", "Origin": "Chevron newsroom article and redistributed copy", "Supplied_interpretation": "Analyst inference: specialist agents can make heterogeneous exploration archives easier to search; no independently measured drilling or study-time result disclosed.", "Partners": "In-house", "Related_IDs": []}]}, "reason": "The issuer article gives the correct date and corpus facts, but does not support a weeks-to-minutes study-time outcome.", "source_url": "https://www.chevron.com/newsroom/2025/q4/a-smarter-way-to-prospect-for-oil-and-gas"}, {"id": "AI-091", "fields": {"Context": "On the Q1 2026 earnings call, Oxy leaders connected AI with workforce efficiency and process review. They gave no named AI workflow, deployment scale, or AI-specific productivity result.", "Reported_metric": "No AI-specific quantified productivity, savings, or financial result disclosed", "Stage_at_disclosure": "Management described AI as taking a larger role across disciplines; specific deployments were not identified in this call.", "Interpretation_limit": "This source supports a general workforce-efficiency strategy statement, not an operating workflow or an AI-attributed cash-flow target.", "Source_URL": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf", "Source_type": "issuer-hosted earnings-call transcript (with third-party transcript rendition reviewed)", "Evidence_status": "Primary earnings-call transcript excerpt reviewed", "Disclosures": [{"ID": "REF-052", "Initiative_ID": "AI-091", "Title": "AI identified as part of workforce-efficiency and process review", "Date": "2026-05-06", "Finding": "Oxy management said AI was taking a larger role across disciplines and discussed workforce efficiency, including reviewing processes and priorities. No named workflow, deployment scale, or AI-specific savings result was disclosed.", "Metric": "No AI-specific quantified result disclosed", "Source_IDs": "R052; S25", "Evidence": "Official transcript excerpt reviewed; third-party transcript is corroborating rendering", "Origin": "Q1 2026 earnings-call remarks", "Supplied_interpretation": "Management statement indicates strategic intent; it does not establish measured workforce savings or causation.", "Partners": "None disclosed", "Related_IDs": []}]}, "reason": "The reopened Q1 transcript does not support the $4B figure as an AI-linked outcome; preserve only the source-supported strategy statement.", "source_url": "https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q26transcript.pdf"}, {"id": "AI-104", "fields": {"Source_date": "2025-01-14", "Source_type": "partner announcement", "Time_scope": "Dated source: 2025-01-14; strategic investment announced, KMI operating rollout not established."}, "reason": "Corrects the initiative source date to the date of the cited Flyscan announcement; April 16 is the date of a separate KMI earnings release.", "source_url": "https://flyscan.com/2025/01/14/flyscan-systems-announces-that-kinder-morgan-joins-its-strategic-investor-group-alongside-marathon-petroleum-enbridge-and-hatch-newswire/"}], "context_audit": {"reviewer": "Lead-agent context audit", "record_audits": [{"id": "CTX-REF-001", "company": "Saudi Aramco", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Historical March 4, 2024 secondary account; model size and intended tasks are reported claims. The planned larger model is not proof of completion.", "correction": "", "source_url": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-014", "company": "Chevron", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference.", "correction": "", "source_url": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce"}], "residual_gaps": ["Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference."]}, {"id": "CTX-REF-018", "company": "Shell", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "NVIDIA dates this technical article September 9, 2024, before the reporting window. Computational speedup and accuracy describe model tests, not field-wide economic returns.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/"}], "residual_gaps": []}, {"id": "CTX-REF-019", "company": "BP", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here.", "correction": "", "source_url": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/"}], "residual_gaps": ["Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here."]}, {"id": "CTX-REF-020", "company": "BP", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI.", "correction": "", "source_url": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html"}], "residual_gaps": ["Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI."]}, {"id": "CTX-REF-023", "company": "TotalEnergies", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://totalenergies.com/news/news/ia-expediting-energy-transition"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The referenced TotalEnergies article is dated January 24, 2025, not February 2024. It discusses a prior Copilot adoption and continuing supervised AI use; it is context alongside existing TotalEnergies dossiers.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://totalenergies.com/news/news/ia-expediting-energy-transition"}], "residual_gaps": []}, {"id": "CTX-REF-030", "company": "Petrobras", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Microsoft publication is December 5, 2023. It describes ChatPetrobras using Azure OpenAI; accessibility to workers is not a measured active-user count.", "correction": "", "source_url": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-066", "company": "Marathon Petroleum", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The undated Imubit page attributes the 25% drum-cycle result to Marathon at the 2024 AFPM Summit. It does not support the separate Flyscan investment claim; removed that claim.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes"}], "residual_gaps": []}, {"id": "CTX-REF-088", "company": "Palantir", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Issuer release reproduced by StockTitan is September 25, 2024, one day before the window. Retained as historical context, not a current-period new initiative.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html"}], "residual_gaps": []}, {"id": "CTX-REF-089", "company": "AWS", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "AWS article is August 7, 2024. The 30% bpx production ambition is not an Xpert result; removed it from the metric field.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/"}], "residual_gaps": []}, {"id": "CTX-REF-091", "company": "Databricks", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Databricks issuer release is April 9, 2024. Its organization count is company-wide, not energy adoption; excluded from initiative performance metrics.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-097", "company": "Enverus", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Enverus dates the launch February 20, 2024. Later product extensions require their own disclosure evidence.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-108", "company": "McKinsey & Company", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Oil & Gas Journal reports McKinsey scenarios on August 26, 2026. They are modeled industry opportunities, not realized AI returns; underlying McKinsey model was not independently audited.", "correction": "", "source_url": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-109", "company": "Rystad Energy", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Rystad describes four workflow categories for the $500 billion estimate. The earlier five-area wording came from a separate digitalization article and was removed; publication month is not visible in the opened page.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas"}], "residual_gaps": []}, {"id": "CTX-REF-110", "company": "Rystad Energy", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "World Oil attributes the December 5, 2025 estimate to Rystad. It is a digitalization opportunity including AI, not an AI-only savings result.", "correction": "", "source_url": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-111", "company": "Boston Consulting Group", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "BCG PDF executive summary presents 30–70% potential EBIT improvement. It is a scenario, not achieved industry profit; other chart statistics require a separate detailed methodology review.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-112", "company": "Bain & Company", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Bain publication is March 19, 2026, not February. Its headline says fewer than 20% scaling with results while body says a quarter scaling applications/reimagining functions; preserve the source distinction and mixed-sector survey population.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/"}], "residual_gaps": []}, {"id": "CTX-REF-113", "company": "EY", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "EY groups oil and gas AND chemicals executives together for the 85%/29% findings. Earlier text incorrectly treated these as oil-and-gas-only statistics.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey"}], "residual_gaps": []}, {"id": "CTX-REF-114", "company": "EY", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "EY article is April 13, 2026; December 2025 is the survey wave. The statistics describe energy respondents and should not be presented as an oil-and-gas-only sample.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption"}], "residual_gaps": []}, {"id": "CTX-REF-115", "company": "EY", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "EY January 6, 2025 article supports the 96% data/access/silo challenge figures. Survey methodology was not fully reviewed; this is context, not a deployment.", "correction": "", "source_url": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-116", "company": "Deloitte", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Deloitte presents IT-spending forecasts and selected maintenance examples. Spending share is a projection; isolated vendor-derived maintenance results are not typical industry outcomes.", "correction": "", "source_url": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-117", "company": "Deloitte", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Deloitte describes 36% of task hours as human-in-the-loop: 26% mechanical/electronic automation plus 10% generative/agentic AI. It is incorrect to imply one-third of all tasks are agentic-AI-led.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html"}], "residual_gaps": []}, {"id": "CTX-REF-118", "company": "KPMG", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "KPMG landing-page figures are 40%, 80% and 82% for the energy, natural resources and chemicals CEO cohort, not exclusively oil and gas.", "correction": "", "source_url": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-119", "company": "DNV", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "DNV table percentages are NET agreement (agree minus disagree). Its prose also distinguishes 12% live use and 3% advanced use. Removed simple adoption-percentage interpretation of the net-agreement table.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/"}], "residual_gaps": []}, {"id": "CTX-REF-120", "company": "DNV", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Oilfield Technology reports DNV intentions for the year ahead, not observed completed deployment. The surveyed population spans energy companies, not solely oil and gas.", "correction": "", "source_url": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-121", "company": "DNV Cyber", "outcome": "supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "DNV January 22, 2025 release supports the retained 375-respondent energy-sector cyber figures; these are respondent views rather than measured incident rates.", "correction": "", "source_url": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/"}], "residual_gaps": []}, {"id": "CTX-REF-122", "company": "International Energy Agency", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.iea.org/reports/energy-and-ai"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "IEA report landing page dated April 10, 2025 was reopened. The detailed supercomputer statistics were not checked in the underlying chapter during this context audit.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.iea.org/reports/energy-and-ai"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-123", "company": "PwC", "outcome": "supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "PwC April 13, 2026 release supports 74%/20% and the 1,217-executive, 25-sector population. This is a cross-industry self-report study, not oil-and-gas operating evidence.", "correction": "", "source_url": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html"}], "residual_gaps": []}, {"id": "CTX-REF-124", "company": "Novi Labs", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Novi vendor survey supports headline adoption and well-forecasting figures among more than 150 US reservoir engineers. Self-selected/vendor survey is not a census of operators; publication month not confirmed.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-125", "company": "Shell", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Microsoft case is November 7, 2023, not June. DIY is low-code/citizen development; the disclosed developer count does not establish AI adoption.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk"}], "residual_gaps": []}, {"id": "CTX-REF-127", "company": "Society of Petroleum Engineers", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Event page establishes April 7–10, 2026 forum and topics. Agenda is not evidence of completed operator deployments or measured benefits.", "correction": "", "source_url": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-128", "company": "The Open Group OSDU Forum", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "The Open Group post describes work and forward priorities. Community implementation and certification should not be represented as completed deliverables based on forward-looking testimonials.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/"}], "residual_gaps": []}, {"id": "CTX-REF-129", "company": "Rice University / Chevron", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Rice interview supports Chevron Digital Scholars and a reservoir-decline capstone. Exact publication date and complete cohort metrics were not re-established from the reviewed excerpt.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-130", "company": "US Department of Energy", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "DOE 26-challenge announcement is February 12, 2026. Removed unsupported $5 billion funding and 17-lab numeric claims from this single-source record.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and"}], "residual_gaps": []}, {"id": "CTX-REF-131", "company": "North Sea Transition Authority", "outcome": "partially supported", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "NSTA source supports a £107,000 grant and planned chatbot/data-viewer work. A grant and planned assessment are not proof of production deployment.", "correction": "", "source_url": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/"}], "residual_gaps": ["No full independent verification of all original details; read the scope of the check above."]}, {"id": "CTX-REF-132", "company": "Havtil (Norwegian Ocean Industry Authority)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "Havtil article supports risk considerations and supervised use. Removed the separate alleged 2025 dropped-object pilot, which the linked 2024 article does not establish.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/"}], "residual_gaps": []}, {"id": "CTX-REF-134", "company": "Journal of Petroleum Technology (SPE)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "SPE commentary cites a cross-industry generative-AI study for 95% without measurable P&L impact. It must not be generalized to an observed 95% failure rate for all oil-and-gas AI.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas"}], "residual_gaps": []}, {"id": "CTX-REF-135", "company": "Enabled Emissions Campaign / Nature study", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference.", "correction": "", "source_url": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/"}], "residual_gaps": ["Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference."]}, {"id": "CTX-REF-136", "company": "Microsoft (shareholders and ex-employees)", "outcome": "corrected", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products"], "findings": [{"severity": "medium", "field": "Context evidence", "issue": "As You Sow December 4, 2025 statement establishes an investor proposal and advocacy claims. Removed unsupported cloud-market-share and production-uplift figures from this reference.", "correction": "Current context entry revised; earlier wording retained in baseline audit snapshot.", "source_url": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products"}], "residual_gaps": []}, {"id": "CTX-REF-137", "company": "ConocoPhillips", "outcome": "unverified", "checks": ["source access", "date eligibility", "population and metric scope", "AI attribution", "context classification"], "source_urls_reviewed": ["https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/"], "findings": [{"severity": "low", "field": "Context evidence", "issue": "Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained.", "correction": "", "source_url": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/"}], "residual_gaps": ["Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained."]}], "source_audits": [{"url": "https://www.middleeastainews.com/p/aramco-launches-largest-industrial-llm", "access": "excerpt reviewed", "title": "Aramco launches Metabrain 250B-parameter industrial LLM", "publication_date": "2024-03", "claims_checked": ["Historical March 4, 2024 secondary account; model size and intended tasks are reported claims. The planned larger model is not proof of completion."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://inspectioneering.com/news/2025-02-13/11441/chevron-to-lay-off-up-to-20-of-its-global-workforce", "access": "inaccessible", "title": "Up to 20% workforce cut, technology and hubs cited", "publication_date": "2025-02", "claims_checked": ["Supplied source returned 403. Workforce reductions cannot be attributed to AI from the retained reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://developer.nvidia.com/blog/spotlight-shell-accelerates-co2-storage-modeling-100000x-using-nvidia-physicsnemo/", "access": "excerpt reviewed", "title": "NVIDIA physics-ML speeds CO2 storage modeling 100,000x", "publication_date": "2024-09-09", "claims_checked": ["NVIDIA dates this technical article September 9, 2024, before the reporting window. Computational speedup and accuracy describe model tests, not field-wide economic returns."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://investors.palantir.com/news-details/2024/Palantir-and-bp-Agree-to-5-Year-Strategic-Relationship-With-New-AI-Capabilities/", "access": "inaccessible", "title": "bp renews Palantir for 5 years with AIP LLMs", "publication_date": "2024-09", "claims_checked": ["Palantir URL returned navigation without the announcement body. The five-year renewal and sensor claims were not freshly verified here."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://uk.finance.yahoo.com/news/bp-increases-staff-cuts-6-083702505.html", "access": "inaccessible", "title": "bp expands job cuts as AI drives efficiency", "publication_date": "2025-08", "claims_checked": ["Supplied Yahoo URL returned 429. No fresh support was established for the job-cut numbers or attribution to AI."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://totalenergies.com/news/news/ia-expediting-energy-transition", "access": "excerpt reviewed", "title": "Early adopter of Microsoft Copilot across the workforce", "publication_date": "2025-01-24", "claims_checked": ["The referenced TotalEnergies article is dated January 24, 2025, not February 2024. It discusses a prior Copilot adoption and continuing supervised AI use; it is context alongside existing TotalEnergies dossiers."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://news.microsoft.com/source/latam/company-news-pt-br/petrobras-cria-ferramenta-com-inteligencia-artificial-generativa-para-apoiar-mais-de-100-mil-trabalhadores/?lang=pt-br", "access": "excerpt reviewed", "title": "ChatPetrobras GenAI rolled out to 110,000 workers", "publication_date": "2023-12", "claims_checked": ["Microsoft publication is December 5, 2023. It describes ChatPetrobras using Azure OpenAI; accessibility to workers is not a measured active-user count."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://imubit.com/blog/opportunities-positive-roi-application-ai-refinery-processes", "access": "excerpt reviewed", "title": "Imubit closed-loop AI cuts suboptimal coke-drum cycles", "publication_date": "Undated; refers to 2024 AFPM Summit", "claims_checked": ["The undated Imubit page attributes the 25% drum-cycle result to Marathon at the 2024 AFPM Summit. It does not support the separate Flyscan investment claim; removed that claim."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.stocktitan.net/news/PLTR/apa-corporation-expands-partnership-with-palantir-to-leverage-ai-tqwva36dmbm1.html", "access": "excerpt reviewed", "title": "APA extends multi-year Palantir deal, adds AIP", "publication_date": "2024-09-25", "claims_checked": ["Issuer release reproduced by StockTitan is September 25, 2024, one day before the window. Retained as historical context, not a current-period new initiative."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://aws.amazon.com/blogs/industries/how-bpx-energy-uses-amazon-bedrock-to-transform-oil-and-gas-production-insights/", "access": "excerpt reviewed", "title": "bpx 'Xpert' genAI chat on Amazon Bedrock", "publication_date": "2024-08-07", "claims_checked": ["AWS article is August 7, 2024. The 30% bpx production ambition is not an Xpert result; removed it from the metric field."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.prnewswire.com/news-releases/databricks-launches-data-intelligence-platform-for-energy-bringing-generative-ai-capabilities-to-the-energy-sector-302111159.html", "access": "excerpt reviewed", "title": "Data Intelligence Platform for Energy with genAI", "publication_date": "2024-04-09", "claims_checked": ["Databricks issuer release is April 9, 2024. Its organization count is company-wide, not energy adoption; excluded from initiative performance metrics."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.enverus.com/newsroom/enverus-instant-analyst-technology-added-to-its-ai-powered-product-portfolio/", "access": "excerpt reviewed", "title": "Instant Analyst genAI with cited answers", "publication_date": "2024-02-20", "claims_checked": ["Enverus dates the launch February 20, 2024. Later product extensions require their own disclosure evidence."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.ogj.com/general-interest/news/55400981/analysis-ai-could-unlock-230-billion-in-annual-upstream-oil-and-gas-value", "access": "excerpt reviewed", "title": "AI could unlock $230B a year in upstream", "publication_date": "2026-08", "claims_checked": ["Oil & Gas Journal reports McKinsey scenarios on August 26, 2026. They are modeled industry opportunities, not realized AI returns; underlying McKinsey model was not independently audited."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas", "access": "excerpt reviewed", "title": "Digital and AI: a $500B upstream opportunity", "publication_date": "Undated source; 2026–2030 forecast period", "claims_checked": ["Rystad describes four workflow categories for the $500 billion estimate. The earlier five-area wording came from a separate digitalization article and was removed; publication month is not visible in the opened page."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.worldoil.com/news/2025/12/5/digitalization-may-unlock-320-billion-in-oil-and-gas-savings-by-2030-analysts-say/", "access": "excerpt reviewed", "title": "Digitalization could save $320B by 2030", "publication_date": "2025-12", "claims_checked": ["World Oil attributes the December 5, 2025 estimate to Rystad. It is a digitalization opportunity including AI, not an AI-only savings result."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf", "access": "excerpt reviewed", "title": "AI reshapes O&G winners: Deploy, Reshape, Invent", "publication_date": "2025-08", "claims_checked": ["BCG PDF executive summary presents 30–70% potential EBIT improvement. It is a scenario, not achieved industry profit; other chart statistics require a separate detailed methodology review."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.bain.com/insights/energy-agenda-2026-returns-restructuring-and-resilience/", "access": "excerpt reviewed", "title": "Energy execs: fewer than 20% scaling AI with results", "publication_date": "2026-03-19", "claims_checked": ["Bain publication is March 19, 2026, not February. Its headline says fewer than 20% scaling with results while body says a quarter scaling applications/reimagining functions; preserve the source distinction and mixed-sector survey population."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.ey.com/en_us/newsroom/2025/03/ey-launches-the-ey-future-of-energy-survey", "access": "excerpt reviewed", "title": "O&G execs: 85% say reskilling decides success; 29% do it", "publication_date": "2025-03", "claims_checked": ["EY groups oil and gas AND chemicals executives together for the 85%/29% findings. Earlier text incorrectly treated these as oil-and-gas-only statistics."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.ey.com/en_us/insights/energy-resources/energy-cautiously-enters-the-next-stage-of-ai-adoption", "access": "excerpt reviewed", "title": "Energy leaders struggle to link AI to productivity", "publication_date": "2026-04-13", "claims_checked": ["EY article is April 13, 2026; December 2025 is the survey wave. The statistics describe energy respondents and should not be presented as an oil-and-gas-only sample."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.ey.com/en_us/insights/oil-gas/building-out-the-ai-assisted-energy-company", "access": "excerpt reviewed", "title": "96% of O&G execs cite data and vendor challenges", "publication_date": "2025-01", "claims_checked": ["EY January 6, 2025 article supports the 96% data/access/silo challenge figures. Survey methodology was not fully reviewed; this is context, not a deployment."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.deloitte.com/us/en/insights/industry/oil-and-gas/oil-and-gas-industry-outlook.html", "access": "excerpt reviewed", "title": "AI heading past 50% of O&G IT spend by 2029", "publication_date": "2025-10", "claims_checked": ["Deloitte presents IT-spending forecasts and selected maintenance examples. Spending share is a projection; isolated vendor-derived maintenance results are not typical industry outcomes."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.prnewswire.com/news-releases/deloitte-us-energy--chemicals-industry-charts-growth-pathworkforce-to-grow-4-1-by-2033--industry-embraces-ai-1-2-million-workers-need-upskilling-302466392.html", "access": "excerpt reviewed", "title": "1.2M US energy and chemicals workers need upskilling", "publication_date": "2025-05", "claims_checked": ["Deloitte describes 36% of task hours as human-in-the-loop: 26% mechanical/electronic automation plus 10% generative/agentic AI. It is incorrect to imply one-third of all tasks are agentic-AI-led."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://kpmg.com/in/en/insights/2025/11/kpmg-2025-global-energy-natural-resources-and-chemicals-ceo-outlook.html", "access": "excerpt reviewed", "title": "Energy CEOs: only 40% actively reskilling for AI", "publication_date": "2025-11", "claims_checked": ["KPMG landing-page figures are 40%, 80% and 82% for the energy, natural resources and chemicals CEO cohort, not exclusively oil and gas."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "access": "excerpt reviewed", "title": "Only 12% of O&G firms use AI in live operations", "publication_date": "2024-10", "claims_checked": ["DNV table percentages are NET agreement (agree minus disagree). Its prose also distinguishes 12% live use and 3% advanced use. Removed simple adoption-percentage interpretation of the net-agreement table."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.oilfieldtechnology.com/digital-oilfield/01092025/offshore-energy-industry-urged-to-prioritise-safety-in-the-era-of-ai/", "access": "excerpt reviewed", "title": "60% plan AI deployment; human-AI interaction a key risk", "publication_date": "2025-09", "claims_checked": ["Oilfield Technology reports DNV intentions for the year ahead, not observed completed deployment. The surveyed population spans energy companies, not solely oil and gas."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.dnv.com/cyber/insights/news/energy-companies-boosting-investment-in-cybersecurity-arms-race-to-manage-the-greatest-risk-to-the-industry-today/", "access": "excerpt reviewed", "title": "65% of energy firms see cyber as top risk", "publication_date": "2025-01", "claims_checked": ["DNV January 22, 2025 release supports the retained 375-respondent energy-sector cyber figures; these are respondent views rather than measured incident rates."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.iea.org/reports/energy-and-ai", "access": "excerpt reviewed", "title": "IEA: oil and gas an early AI adopter", "publication_date": "2025-04", "claims_checked": ["IEA report landing page dated April 10, 2025 was reopened. The detailed supercomputer statistics were not checked in the underlying chapter during this context audit."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-performance-study.html", "access": "excerpt reviewed", "title": "20% of companies capture 74% of AI value", "publication_date": "2026-04", "claims_checked": ["PwC April 13, 2026 release supports 74%/20% and the 1,217-executive, 25-sector population. This is a cross-industry self-report study, not oil-and-gas operating evidence."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://novilabs.com/blog/the-state-of-ai-adoption-in-reservoir-engineering-2024-2025-comparative-analysis/", "access": "excerpt reviewed", "title": "Reservoir engineer AI use rises to 56%", "publication_date": "Undated article; compares 2024–2025 survey waves", "claims_checked": ["Novi vendor survey supports headline adoption and well-forecasting figures among more than 150 US reservoir engineers. Self-selected/vendor survey is not a census of operators; publication month not confirmed."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.microsoft.com/en/customers/story/1699045592642116808-shell-energy-power-platform-uk", "access": "excerpt reviewed", "title": "Shell's DIY citizen developer program passes 4,000 builders", "publication_date": "2023-11-07", "claims_checked": ["Microsoft case is November 7, 2023, not June. DIY is low-code/citizen development; the disclosed developer count does not establish AI adoption."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.spe-events.org/forum/Artificial-Intelligence-Upstream-2030-Beyond", "access": "excerpt reviewed", "title": "SPE Forum: AI in Upstream E&P 2030", "publication_date": "2026-04", "claims_checked": ["Event page establishes April 7–10, 2026 forum and topics. Agenda is not evidence of completed operator deployments or measured benefits."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://blog.opengroup.org/2025/12/04/osdu-forum-highlights-from-the-open-group-houston-summit-2025/", "access": "excerpt reviewed", "title": "OSDU pushes an AI-ready open data platform", "publication_date": "2025-12", "claims_checked": ["The Open Group post describes work and forward priorities. Community implementation and certification should not be represented as completed deliverables based on forward-looking testimonials."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://d2k.rice.edu/news/leveraging-machine-learning-accelerate-chevrons-digital-transformation", "access": "excerpt reviewed", "title": "Chevron Digital Scholars earn Rice data-science master's", "publication_date": "Historical context; exact article date not re-established", "claims_checked": ["Rice interview supports Chevron Digital Scholars and a reservoir-decline capstone. Exact publication date and complete cohort metrics were not re-established from the reviewed excerpt."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.energy.gov/undersecretaryforscience/articles/energy-department-announces-26-genesis-mission-science-and", "access": "excerpt reviewed", "title": "Genesis Mission targets AI for subsurface energy", "publication_date": "2026-02-12", "claims_checked": ["DOE 26-challenge announcement is February 12, 2026. Removed unsupported $5 billion funding and 17-lab numeric claims from this single-source record."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.nstauthority.co.uk/news-publications/innovative-project-to-improve-access-to-north-sea-data/", "access": "excerpt reviewed", "title": "NSTA funds AI chatbot for North Sea data", "publication_date": "2025-10", "claims_checked": ["NSTA source supports a £107,000 grant and planned chatbot/data-viewer work. A grant and planned assessment are not proof of production deployment."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2024/ai-and-risk-in-norways-petroleum-industry/", "access": "excerpt reviewed", "title": "Norway regulator flags AI black-box and data risks", "publication_date": "2024-03", "claims_checked": ["Havtil article supports risk considerations and supervised use. Removed the separate alleged 2025 dropped-object pilot, which the linked 2024 article does not establish."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://jpt.spe.org/twa/from-pilot-to-production-understanding-the-challenges-of-scaling-ai-in-oil-and-gas", "access": "excerpt reviewed", "title": "Why O&G AI pilots fail to scale", "publication_date": "2026-08", "claims_checked": ["SPE commentary cites a cross-industry generative-AI study for 95% without measurable P&L impact. It must not be generalized to an observed 95% failure rate for all oil-and-gas AI."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://insideclimatenews.org/news/13082026/ai-enables-oil-and-gas-production/", "access": "inaccessible", "title": "Study: AI in oil and gas adds up to 1.8 Gt CO2", "publication_date": "2026-08", "claims_checked": ["Supplied article was inaccessible. Underlying study, model assumptions and enabled-emissions estimates were not independently checked; retained only as an unresolved reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.asyousow.org/press-releases/2025/12/4/microsoft-agm-investors-raise-concern-about-company-fueling-climate-change-with-its-ai-technology-products", "access": "excerpt reviewed", "title": "Investors and ex-staff challenge Microsoft's O&G AI work", "publication_date": "2025-12", "claims_checked": ["As You Sow December 4, 2025 statement establishes an investor proposal and advocacy claims. Removed unsupported cloud-market-share and production-uplift figures from this reference."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}, {"url": "https://www.offshore-technology.com/news/conocophillips-to-reduce-workforce/", "access": "inaccessible", "title": "ConocoPhillips cuts up to 25% of staff in BCG-led overhaul", "publication_date": "2025-09", "claims_checked": ["Supplied source was inaccessible. Retained restructuring claim is not established as AI-caused; no new verification of numbers was obtained."], "notes": "Context-only review; excluded from initiative totals. Original extraction is preserved in the prior edition.", "partition": "context"}], "patches": [{"id": "CTX-REF-001", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Proprietary operational history is the moat; NOCs with deep archives are building domain LLMs rather than relying only on general-purpose models."}}, {"id": "CTX-REF-014", "fields": {"Evidence": "Source access unresolved", "summary": "Unverified inherited reference: Chevron said it would cut 15-20% of its global workforce (up to ~9,000 roles) by end-2026 to save $2-3B. Levers included leveraging technology to raise productivity and changing where work is done via global capability centers such as Manila, Buenos Aires and its Bengaluru ENGINE center.", "metrics": "Unverified inherited figures; source access unresolved.", "takeaway": "Analyst interpretation: Majors are pairing technology-enabled productivity with centralized service hubs; AI enablement pitches should target work consolidated into these centers."}}, {"id": "CTX-REF-018", "fields": {"date": "2024-09-09", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Physics-ML surrogates turn slow simulation into fast scenario screening, a strong pattern for subsurface and CCS workflows."}}, {"id": "CTX-REF-019", "fields": {"Evidence": "Source access unresolved", "summary": "Unverified inherited reference: bp renewed its decade-long Palantir relationship with a five-year strategic agreement centered on Palantir's Artificial Intelligence Platform (AIP). AIP brings LLMs into bp's operations on top of digital twins fed by more than 2 million sensors to speed human decision-making.", "metrics": "Unverified inherited figures; source access unresolved.", "takeaway": "Analyst interpretation: Operators with an existing ontology and digital-twin layer can add LLM copilots quickly; the data model is the durable asset."}}, {"id": "CTX-REF-020", "fields": {"Evidence": "Source access unresolved", "summary": "Unverified inherited reference: bp raised expected job losses to about 6,200 (roughly 15% of office-based staff) plus contractor cuts. CEO Murray Auchincloss said technology is helping improve capital productivity and drive cost reductions, with AI part of the overhaul toward a $5B cost-cut target by end-2027.", "metrics": "Unverified inherited figures; source access unresolved.", "takeaway": "Analyst interpretation: AI is now explicitly linked to headcount plans at majors; change management and redeployment belong in the business case alongside productivity."}}, {"id": "CTX-REF-023", "fields": {"date": "2025-01-24", "Category": "Context / survey / enabling ecosystem", "title": "TotalEnergies describes supervised AI and prior Copilot adoption", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Horizontal copilots plus a central digital factory for vertical use cases is the standard two-track model among majors."}}, {"id": "CTX-REF-030", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: A secure internal ChatGPT-style wrapper extended to contractors reaches the whole workforce quickly and cheaply."}}, {"id": "CTX-REF-066", "fields": {"date": "Undated; refers to 2024 AFPM Summit", "summary": "Imubit reports that Marathon Petroleum presented a 25% reduction in suboptimal coke-drum cycles at the 2024 AFPM Summit. The vendor describes AI-based closed-loop process optimization. The linked page is undated and does not establish a separate Flyscan investment.", "vendors": "Imubit", "metrics": "Vendor-reported: 25% fewer suboptimal drum cycles; facility, measurement period and realized savings not specified here.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Reinforcement-learning optimizers layered on existing APC can deliver measurable unit gains without replacing controls."}}, {"id": "CTX-REF-088", "fields": {"date": "2024-09-25", "Category": "Historical context outside target window", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Palantir expands inside E&Ps from an operational data layer into back-office genAI such as invoice and contract checks, so land-and-expand from one ontology is the model to beat."}}, {"id": "CTX-REF-089", "fields": {"date": "2024-08-07", "metrics": "No quantified Xpert-specific operating result established by this review.", "takeaway": "Analyst interpretation: conversational access to operating data is a useful workflow; company production growth cannot be attributed to the assistant.", "Evidence": "Context source excerpt reviewed"}}, {"id": "CTX-REF-091", "fields": {"date": "2024-04-09", "metrics": "No energy-specific customer count or operating benefit established by this launch release.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Majors' AI and ML work often runs on Databricks, so building natively on its lakehouse (Unity Catalog, model serving) lowers the barrier to buying."}}, {"id": "CTX-REF-097", "fields": {"date": "2024-02-20", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Answers with citations over proprietary data are what buyers now expect as a minimum for O&G genAI."}}, {"id": "CTX-REF-108", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Fewer than 20% of deployments track AI value, so building value measurement into the enablement program is part of the product."}}, {"id": "CTX-REF-109", "fields": {"date": "Undated source; 2026–2030 forecast period", "summary": "Rystad estimates nearly $500 billion of cumulative value from digitalization and AI in E&P during 2026–2030, with cost reductions and production increases contributing roughly equally. It groups workflows into asset development; operations and maintenance; exploration and reservoir development; and drilling, wells and production. This is a modeled opportunity, not realized industry returns.", "metrics": "Rystad forecast: about $500 billion cumulative value in 2026–2030; $80 billion additional annual value in 2030 versus 2025.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Frame use-case portfolios around the few value pools the analysts agree on (drilling, maintenance, reservoir) instead of spreading effort thin."}}, {"id": "CTX-REF-110", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Smaller operators face cost and cyber barriers, so modular, low-infrastructure enablement offerings are an underserved gap."}}, {"id": "CTX-REF-111", "fields": {"summary": "BCG presents a potential 30–70% improvement in oil-and-gas EBIT over five years and organizes AI strategy into deployment of tools, redesign of workflows and new business models. This is consulting analysis rather than measured sector-wide profitability.", "metrics": "BCG modeled potential: 30–70% EBIT improvement over five years; not an achieved return.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Seat licences (Deploy) are table stakes. Most of the value sits in redesigning a few core workflows, so enablement has to go past prompt training."}}, {"id": "CTX-REF-112", "fields": {"date": "2026-03-19", "summary": "Bain surveyed 859 executives across oil and gas, utilities, chemicals, mining and agribusiness in December 2025–January 2026. Its headline says fewer than 20% are scaling AI with results; the body separately says about a quarter are scaling applications and reimagining functions with measurable impact. These formulations are retained as distinct source statements, not interchangeable adoption rates.", "metrics": "Mixed-sector survey, n=859. Headline: fewer than 20% scaling with results. Body: about one-quarter scaling applications/reimagining functions; fewer than 10% companywide transformation.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: The top barrier is unclear business value, not technology, so tie each use case to an owner and a P&L metric before building it."}}, {"id": "CTX-REF-113", "fields": {"title": "Oil, gas and chemicals executives: 85% value reskilling; 29% retrain", "summary": "EY’s survey of 1,020 energy-industry respondents in November–December 2024 found a gap between reskilling intentions and activity. Within the combined oil-and-gas-and-chemicals executive cohort, 85% said reskilling would determine success, while 29% reported current retraining and 58% said culture impeded digital adoption.", "metrics": "Combined oil/gas/chemicals cohort: 85% reskilling importance; 29% retraining; 58% cultural impediment. Total energy survey n=1,020.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: The market gap is in doing the reskilling, not in convincing people it matters."}}, {"id": "CTX-REF-114", "fields": {"date": "2026-04-13", "metrics": "December 2025 survey wave: energy leaders 56% versus all-industry 35% struggle with AI-productivity attribution; 72% versus 50% see a training need. Not oil-and-gas-only.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Energy lags other sectors on measuring AI gains, so offering an attribution method alongside training is a differentiator."}}, {"id": "CTX-REF-115", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Plan the pilot-to-production handoff (data plumbing, vendor orchestration, ownership) as part of any use case, not as an afterthought."}}, {"id": "CTX-REF-116", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: As budgets swing toward AI, the scarce input becomes people who can use it, which is a direct opening for enablement."}}, {"id": "CTX-REF-117", "fields": {"summary": "Deloitte forecasts that 1.2 million US energy-and-chemicals workers need upskilling and 534,000 need reskilling. Of analyzed task hours, 64% remain human-driven; 36% involve humans in the loop, split between mechanical/electronic automation (26%) and generative/agentic AI (10%). These are workforce projections, not observed job displacement.", "metrics": "Projected: 1.2 million upskilling; 534,000 reskilling; 26% mechanical/electronic automation task hours and 10% generative/agentic AI task hours.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Upskilling demand is workforce-wide, not only for engineers, so programs need role-based tracks that include field and operations staff."}}, {"id": "CTX-REF-118", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: CEO awareness has outrun action on skills, so position enablement as how the AI strategy gets executed."}}, {"id": "CTX-REF-119", "fields": {"summary": "DNV describes low operational AI maturity among its oil-and-gas respondents: 12% reported live day-to-day use and 3% advanced integration, while 47% described planning or pilots. Its separate strategic-intention table reports net agreement, defined as agreement minus disagreement; those values are not simple adoption shares.", "metrics": "Reported maturity: 12% live use; 3% advanced use; 47% planning/piloting. Separate 62% digital-strategy and 47% forward-intention figures are net agreement.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Assurance (showing that AI is safe, secure and validated) is itself an adoption lever in a safety-critical industry."}}, {"id": "CTX-REF-120", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Teach how to supervise AI (when to override, how to verify), not just how to use it."}}, {"id": "CTX-REF-121", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Any agentic or OT-adjacent AI program has to be designed with OT security from the start; cyber is a gating objection."}}, {"id": "CTX-REF-122", "fields": {"summary": "IEA’s April 10, 2025 Energy and AI report examines both electricity needs for AI and AI’s potential to change energy operations. This review confirmed the report and scope but did not recheck the detailed oil-and-gas supercomputer statistics in its underlying chapter.", "metrics": "Detailed supercomputer counts are withheld pending chapter-level verification.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Even the IEA ranks skills alongside data and security as a core barrier, which supports treating enablement as infrastructure."}}, {"id": "CTX-REF-123", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: AI value is winner-take-most, so the case for enablement is competitive positioning, not just efficiency."}}, {"id": "CTX-REF-124", "fields": {"date": "Undated article; compares 2024–2025 survey waves", "metrics": "Vendor survey: adoption 43% to 56%; forecasting use 58% to 75%; 84% report public-data challenges. More than 150 US reservoir engineers; not an operator census.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Engineers now own data quality (74%), so enablement should teach data stewardship alongside AI tools."}}, {"id": "CTX-REF-125", "fields": {"date": "2023-11-07", "Category": "Historical context outside target window", "summary": "Microsoft’s November 2023 Shell case describes a Power Platform citizen-development program with more than 4,000 active DIY developers. It concerns low-code applications and automation; the developer count is not an AI-specific user count.", "metrics": "More than 4,000 active DIY developers, partner-reported in 2023; not an AI-adoption metric.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Shell's playbook (CoE, embedded coaches, persona-based paths, hackathons) transfers directly from low-code to GenAI and agent building."}}, {"id": "CTX-REF-127", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: The profession's own agenda is about trust, tacit knowledge and human-AI teaming, so enablement content should address these as well as tools."}}, {"id": "CTX-REF-128", "fields": {"summary": "The Open Group’s December 4, 2025 summit recap describes an OSDU community focus on analytics, AI-ready data access and seismic-interpretation workflows. Testimonials identify community implementation and certification as future priorities; the post alone does not establish those deliverables as complete.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Subsurface AI enablement depends on standardized data foundations such as OSDU; data literacy belongs in the curriculum."}}, {"id": "CTX-REF-129", "fields": {"date": "Historical context; exact article date not re-established", "metrics": "Cohort counts not reverified in this pass; retained as workforce-development context.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Deep, degree-level programs for a small cadre of domain experts complement broad AI literacy and create in-house translators."}}, {"id": "CTX-REF-130", "fields": {"date": "2026-02-12", "summary": "DOE announced 26 Genesis Mission science and technology challenges on February 12, 2026. One challenge concerns AI modeling of underground environments for responsible, cost-effective development of subsurface energy resources. This is a research agenda, not evidence of a deployed operator system or realized production benefit.", "metrics": "26 research challenges; no project-specific funding or operating result established by this source.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Federal subsurface AI programs are a source of open models, datasets and training partners for US operators."}}, {"id": "CTX-REF-131", "fields": {"Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Regulators are building AI access layers over public subsurface data; enablement programs should teach staff to use them."}}, {"id": "CTX-REF-132", "fields": {"Category": "Historical context outside target window", "summary": "Havtil’s 2024 article discusses AI in Norwegian petroleum operations, including data quality, opaque models and human supervision. It warns that an operator may become overreliant on automated recommendations. The linked article does not substantiate the separate 2025 incident-data pilot previously combined with this record.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Regulators expect companies to own AI risk, so enablement needs explainability, verification habits and escalation rules."}}, {"id": "CTX-REF-134", "fields": {"summary": "SPE’s The Way Ahead commentary discusses barriers to scaling AI, including data readiness, organizational alignment, workforce skills and cyber risks. Its cited 95% finding concerns a separate cross-industry generative-AI study and measurable P&L impact; it is not evidence that 95% of oil-and-gas AI projects fail.", "metrics": "No oil-and-gas-wide failure percentage established by this commentary.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Scaling is a governance and change-management problem, so budget for data plumbing and trust-building, not just models."}}, {"id": "CTX-REF-135", "fields": {"Evidence": "Source access unresolved", "summary": "Unverified inherited reference: A peer-reviewed Nature study reported by Inside Climate News estimates that AI tools that make extraction faster and cheaper add 0.47-1.8 Gt of CO2 a year. That is 1.2-4.8% of 2024 global energy-related emissions and up to 13 times data-center emissions. The authors call these enabled emissions and want them recognized, measured and governed.", "metrics": "Unverified inherited figures; source access unresolved.", "takeaway": "Analyst interpretation: Expect ESG scrutiny of O&G AI programs; framing that covers emissions, safety and methane use cases reduces reputational risk."}}, {"id": "CTX-REF-136", "fields": {"summary": "As You Sow’s December 4, 2025 statement promoted a shareholder proposal asking Microsoft to assess risks from AI products used in oil-and-gas development. This is advocacy and governance context; it does not independently establish operating outcomes or Microsoft’s market share.", "metrics": "No operational benefit, market-share figure or measured emissions result established by this source.", "Evidence": "Context source excerpt reviewed", "takeaway": "Analyst interpretation: Vendor reputational pressure can shift tool availability and terms, so avoid single-vendor lock-in in enablement stacks."}}, {"id": "CTX-REF-137", "fields": {"Evidence": "Source access unresolved", "summary": "Unverified inherited reference: ConocoPhillips said it will cut 20-25% of its roughly 13,000-person workforce under a restructuring called Competitive Edge, advised by BCG, citing rising per-barrel costs. The cuts were not attributed to AI, but they came alongside similar reductions at Chevron and bp. EY's surveys find that employee job-security concerns inhibit AI adoption and experimentation.", "metrics": "Unverified inherited figures; source access unresolved.", "takeaway": "Analyst interpretation: When layoffs coincide with AI rollouts, enablement programs need explicit messaging that AI is for augmentation, or adoption will stall."}}], "method": "Every one of the 40 supplied context URLs was opened. Readable passages were checked for date, sample, attribution and classification. Excerpts and inaccessible sources are identified; this is not independent verification of underlying surveys or performance."}, "supplemental_findings": [{"id": "AI-006", "source_id": "R040", "outcome": "partially supported", "finding": "Devon management links AI investment/platforms to production projects beginning to scale in Q1/Q2 2026, but the remarks do not explicitly say the discussed gas-lift trials were AI-driven. Keep gas-lift optimization and AI-platform work as related but not proven identical workflows; the $1B business optimization target is not AI-attributed.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/02/18/devon-dvn-q4-2025-earnings-call-transcript/"}, {"id": "AI-021", "source_id": "S11", "outcome": "partially supported", "finding": "C3 AI FY2025 Form 10-K was reopened. It verifies the April 2025 renewal, co-selling/co-investment language, and that Baker Hughes is no longer exclusive reseller; separate May 28 issuer release adds term through June 2028, named deployed customers and internal-use applications. Benefits and customer scope remain vendor-reported/aggregate.", "source_url": "https://www.sec.gov/Archives/edgar/data/1577526/000162828025032604/ai-20250430.htm"}, {"id": "AI-044", "source_id": "R050", "outcome": "supported", "finding": "The Q4 2024 call names current Gulf of America AI use in supply-chain management, asset integrity, and reservoir characterization and an enterprise AI center of excellence. Keep Algeria seismic dataset as future-use potential; no results or implementation detail were disclosed.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/02/19/occidental-petroleum-oxy-q4-2024-earnings-call-tra/"}, {"id": "AI-057", "source_id": "S36", "outcome": "supported with scope limitation", "finding": "Ownership source confirms Charlesbank as PakEnergy private-equity partner and documents Petrosight well-lifecycle software acquisition; it does not support AI in PakEnergy accounting workflows. This is corroborating context only, not a factual AI correction.", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}, {"id": "AI-058", "source_id": "S36", "outcome": "supported with scope limitation", "finding": "The reopened acquisition release is ownership/deal context only and does not substantiate the PakCAPTURE ticket/receipt extraction workflow; retain the existing AI workflow on its own cited evidence.", "source_url": "https://pakenergy.com/blog/pakenergy-expands-solutions-for-energy-customers-with-petrosight-acquisition"}, {"id": "AI-076", "source_id": "R016", "outcome": "partially supported", "finding": "C3 AI’s June 4, 2026 release confirms existing Shell predictive maintenance on 13,000+ pieces of equipment, deployed on Azure, and says the new multi-year agreement will introduce agent-based root cause analysis/remediation. Do not describe the new agentic component as already deployed; “hundreds of millions” is C3 AI executive-attributed, not independently validated.", "source_url": "https://www.businesswire.com/news/home/20260604461527/en/C3-AI-and-Shell-Expand-Collaboration-Scaling-Reliability-AI-Deployment-Across-Global-Asset-Operations"}, {"id": "AI-078", "source_id": "R024", "outcome": "partially supported", "finding": "TotalEnergies’ Sep 26, 2025 announcement plans three-year scale-out of Cognite data/AI across all operated upstream assets. It describes scope and intended value chain but is an expansion/scale-up announcement, not proof that worldwide rollout was already complete.", "source_url": "https://totalenergies.com/news/press-releases/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale"}, {"id": "AI-096", "source_id": "R058", "outcome": "partially supported", "finding": "Q1 2026 transcript gives more operational detail than the inherited “vendor-led cautious integration strategy” title: daily real-time data monitored by control-room/field teams, frac water/sand pressures/volumes and logistics, and AI-assisted preferred lateral targeting. Reported 40M data points/day is company-stated scale, not an outcome metric.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/07/matador-mtdr-q1-2026-earnings-call-transcript/"}, {"id": "AI-115", "source_id": "R084", "outcome": "partially supported", "finding": "The Apr 11, 2025 article describes intended integration of Nabors RigCLOUD edge/cloud/rig instrumentation with Corva AI analytics. It does not document deployed customer count, basin/rig scope, or realized gains; retain as partner/vendor offering/rollout stage.", "source_url": "https://worldoil.com/news/2025/4/11/nabors-corva-ai-join-to-advance-digital-transformation-in-drilling-industry/"}, {"id": "AI-125", "source_id": "R102", "outcome": "partially supported", "finding": "May 7, 2026 article adds a concrete Winn Resources deployment: monthly Texas RRC W-10/G-10 submissions, data-gap visibility, and human review; Collide/Winn-reported processing time reduction is >95%, with a quoted 50-well filing completed in 20 minutes. Other listed workflows are future targets. Evidence is secondary-only in this supplemental source.", "source_url": "https://energycapitalhtx.com/collide-riggs-ai-platform"}, {"id": "AI-021", "issue": "The C3 AI FY2026 Form 10-K says Baker Hughes is no longer C3 AI’s exclusive reseller after April 2025 amendment; the filing still describes a strategic relationship.", "correction": "Retain the multi-year relationship/renewal as reported, but update reseller status to non-exclusive from April 2025 onward. Do not describe the relationship as terminated.", "source_url": "https://ir.c3.ai/static-files/e49d6691-4430-4b62-aace-de6c7692f497"}, {"id": "AI-005", "issue": "Chevron’s cited Project Kilby page could not be reopened; the direct URL returned 403.", "correction": "Retain prior verified project information only as previously reviewed; mark this supplemental source not reverified, not disproved. Its category remains AI-demand infrastructure, not AI adoption.", "source_url": "https://www.chevron.com/newsroom/2026/q2/chevron-signs-20-year-power-agreement-with-microsoft-for-west-texas-data-center"}, {"id": "AI-010", "issue": "The JPT/SPE report confirms a signed strategic collaboration aimed at developing agentic AI and unified upstream workflows; it does not establish an operating deployment.", "correction": "Keep the profile at collaboration/development stage and attribute it to secondary editorial corroboration; avoid wording that suggests Shell assets already use the joint agentic system.", "source_url": "https://jpt.spe.org/slb-shell-agree-to-collaborate-on-agentic-ai-efforts"}, {"id": "AI-032", "issue": "Repsol’s page says Copilot for Web reached ~25,000 employees while M365 Copilot licenses are progressively scaling; a 550-person, four-month pilot reported 121 minutes/week per participant. The page describes >60 use cases in process, not 60+ in production.", "correction": "Keep product reach distinct from M365 licensing and pilot results. Do not attribute the 121 minutes to reduced labor cost or use this page to support a “60+ in production/22 scaled” count.", "source_url": "https://www.repsol.com/en/energy-move-forward/innovation/adoption-generative-ai-repsol/index.cshtml"}, {"id": "AI-006", "issue": "Devon’s Q1 2026 transcript adds management detail: Smart Gas Lift uses AI models/physics-based calculation in a closed loop; 2025 pilot reported 2–3% uplift, and full Delaware Basin implementation exceeded 850 wells. Current uplift above pilot was not quantified.", "correction": "No contradiction to “secondary-sourced”; ensure pilot uplift and company-wide $1B business-optimization savings remain separate, with only the pilot’s 2–3% attributed to Smart Gas Lift.", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/05/06/devon-energy-dvn-q1-2026-earnings-transcript/"}, {"id": "AI-044", "issue": "Oxy management describes AI already helping subsurface work in Permian/onshore, while the complex Gulf of America project is being developed with execution expected in the following one to two years.", "correction": "Separate existing onshore assistance from the forward-looking Gulf project; do not describe Gulf AI project as live/deployed. Retain source as secondary transcript.", "source_url": "https://www.fool.com/earnings/call-transcripts/2025/08/07/occidental-oxy-q2-2025-earnings-call-transcript/"}, {"id": "AI-099", "issue": "The International Mining article supports almost 120 current autonomous trucks at Base Plant and 2025 expansion, but neither AI/ML basis nor the 140-truck planned target is stated there.", "correction": "Keep autonomous haulage distinct from AI. If retaining 140 as a plan, source it separately; this secondary article supports near-120 current fleet, not that target.", "source_url": "https://im-mining.com/2025/07/04/suncor-on-its-record-autonomous-ultraclass-truck-fleet/"}, {"id": "AI-020", "issue": "Baker Hughes Cordant announcement URL returned 403 during direct reopen.", "correction": "Mark this supplemental URL not reverified; do not treat access failure as contradiction. Preserve any earlier review status separately from this pass.", "source_url": "https://www.bakerhughes.com/company/news/baker-hughes-expands-collaboration-microsoft-accelerate-energy-and-industrial"}, {"id": "AI-116", "issue": "Issuer earnings release confirms FlexRobotics deployment on a Permian rig for a supermajor, but does not establish AI in the deployed robot. H&P’s separate product page calls FlexRobotics automation and lists AI zone monitoring as an add-on camera feature.", "correction": "Keep AI only for the explicitly described camera-based monitoring capability and qualify it as add-on/product capability; do not imply it was part of the customer deployment absent evidence. The deployed base robotics workflow is automation.", "source_url": "https://www.businesswire.com/news/home/20260204290127/en/Helmerich-Payne-Inc.-Announces-Fiscal-First-Quarter-Results"}, {"severity": "medium", "field": "source attribution / workflow specificity", "issue": "R043's IADC archive listing verifies only the article title/byline; the body could not be reopened, and the title concerns post-merger AI/synergies rather than the imported description of well-design/completions-sequencing agents.", "correction": "Do not use R043 as fresh verification of the named well-design workflow. Keep earlier presentation evidence as not reverified this pass rather than disproved; reconcile the R043 title/summary mismatch if retaining that citation.", "source_id": "R043", "id": "AI-007"}, {"severity": "medium", "field": "disclosure date / result metric", "issue": "Chevron's own article is dated 2025-11-12, not the supplied December month. It states >1 million files, 4 TB, and seconds-fast retrieval; it does not say studies fell from weeks to minutes.", "correction": "Date REF-013 to 2025-11-12 and remove 'weeks to minutes' as an achieved metric unless another source directly supports it. Keep the corpus size and retrieval workflow.", "source_id": "R013", "id": "AI-074"}, {"severity": "high", "field": "reported_metric / evidence provenance", "issue": "R052 confirms general management statements about AI's growing role in workforce efficiency, but it does not substantiate a $4B target or connect that target to AI. The Motley Fool URL is a third-party transcript; the official Oxy-hosted transcript corroborates the remarks but provides no AI-specific economics.", "correction": "Remove the $4B figure from the AI profile's Reported_metric and from the R052 disclosure finding; state that no AI-specific quantified outcome was disclosed. Limit the record to management's stated strategic direction and its generic workforce/process-efficiency context.", "source_id": "R052", "id": "AI-091"}, {"severity": "high", "field": "AI basis / category / date", "issue": "The Suncor page describes autonomous trucks using GPS, wireless/LTE and perception technologies, with control-room oversight. It contains no AI/artificial-intelligence reference. The page is undated, so neither a 2025 nor 2026 publication date can be inferred; 140 trucks is a future target with no dated year on the page.", "correction": "Replace the AI framing with 'Automation / AI not established'; retain the operating haulage evidence, nearly-120 snapshot, human control-room role and future 140 target without assigning a publication/year to the page. Remove the claim that AI optimizes driving conditions and unverified workforce job-loss claims from the disclosure.", "source_id": "R062", "id": "AI-099"}, {"severity": "medium", "field": "source date", "issue": "The profile's cited Flyscan URL is dated 2025-01-14, while Source_date is 2025-04-16, the date of the separate KMI Q1 earnings release.", "correction": "Set the initiative Source_date to 2025-01-14 and keep the profile's caveat that investment/partnering does not establish KMI deployment.", "source_id": "R068", "id": "AI-104"}, {"severity": "low", "field": "source provenance", "issue": "The user-supplied GlobeNewswire link failed, but Weatherford's issuer-hosted copy of the October 1, 2025 release was reviewed and supports the portfolio-launch details.", "correction": "Retain launch facts with the Weatherford investor-relations copy as primary evidence; describe components as offered capabilities, not customer deployment or outcomes.", "source_id": "R082", "id": "AI-113"}, {"severity": "low", "field": "attribution", "issue": "World Oil is secondary coverage and repeats Aker BP/Cognite claims, including the >70% RCA-time reduction; it is not independent validation.", "correction": "Keep the result labeled partner/company-reported and unbenchmarked; the existing Cognite primary partner release remains the preferred supporting citation.", "source_id": "R092", "id": "AI-049"}], "new_additions_independent_review": {"reviewed_file": "work/research-swarm/round2/additions-root.json", "reviewer": "partition-2 independent review", "review_date": "2026-09-27", "sources_reopened": [{"url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "access": "full text reviewed", "claims_checked": ["Leucipa AI-powered production management, field optimization and forecasting; Lucy described as a planned pilot with GenAI chat interface and real-time production insights; deployment beginning at Operations Support Center.", "Snowflake data platform adoption; ML model forecasting; real-time drilling data/models hosted; SCADA signals used to detect issues and alert Operations Support Center."], "notes": "The issuer’s announcement supports both packets but distinguishes an operating platform/deployment from the planned Lucy pilot.", "title": "EXE Announces Strategic Partnerships to Strengthen Data-Driven Operational Efficiency", "publication_date": "2026-03-03"}, {"url": "https://www.slb.com/newsroom/press-release/2026/pr-2026-0609-slb-qualcomm", "access": "full text reviewed", "claims_checked": ["SLB/Qualcomm edge-AI collaboration and product/workflow development scope."], "notes": "Same exact initiative and announcement as assigned existing AI-046; a second addition would duplicate the existing record.", "title": "SLB Collaborates with Qualcomm on Edge AI Solutions for Energy Operations", "publication_date": "2026-06-09"}, {"url": "https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution", "access": "full text reviewed", "claims_checked": ["Evergreen Data Solution product launch and AI-assisted classification, curation, contextualization and validation of subsurface data."], "notes": "Commercial product capability, not customer deployment proof.", "title": "Viridien redefines subsurface data readiness at scale with Evergreen Data Solution", "publication_date": "2026-06-10"}, {"url": "https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract", "access": "full text reviewed", "claims_checked": ["TGS multi-year AI seismic foundation-model contract award to an unnamed international operator; scope includes structural and stratigraphic interpretation and TGS multi-client data."], "notes": "Customer remains unnamed; no contract value or field result disclosed.", "title": "TGS Awarded Multi-Year AI Seismic Foundation Model Contract", "publication_date": "2026-08-18"}, {"url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results", "access": "excerpt reviewed", "claims_checked": ["SLB Q1 2025 earnings release lines 285–286: Woodside award for integrated delivery of 18 Trion ultra-deepwater wells; AI-enabled drilling capabilities included; service start scheduled early 2026 and delivery via Performance Live."], "notes": "The cited release supports the packet’s timing as a plan and the contracted scope; it does not prove services commenced or outcomes were achieved.", "title": "SLB Announces First-Quarter 2025 Results", "publication_date": "2025-04-25"}, {"url": "https://investors.technipenergies.com/shareholder-information", "access": "excerpt reviewed", "claims_checked": ["2025 shareholder/management messages say DAP leverages advanced analytics and large language models; DAP launched in 2025; target €100m cost savings and incremental revenues by 2028; AI for All implemented."], "notes": "The €100m is a 2028 target, not realized savings. The source supplies no named AI workflow or AI-specific attribution for savings.", "title": "Technip Energies shareholder information — 2025 chair and CEO messages", "publication_date": "Undated; see individual source review"}], "initiative_reviews": [{"key": "expand-leucipa-lucy", "outcome": "supported with stage limitation", "findings": ["Leucipa is described as AI-powered and deployment begins through the Operations Support Center. Lucy is explicitly a pilot (“will also pilot”), so do not write that Lucy is already deployed."], "duplicates": []}, {"key": "expand-snowflake-drilling-ml", "outcome": "supported", "findings": ["Issuer source supports ML models for forecasting and real-time data/model hosting for drilling, plus SCADA alerts to the Operations Support Center. No result metric or specific model detail is disclosed."], "duplicates": ["Same March 2026 issuer announcement as Leucipa/Lucy, but a distinct data/ML workflow; not an exact duplicate."]}, {"key": "slb-qualcomm-edge-ai", "outcome": "duplicate candidate", "findings": ["Exact SLB/Qualcomm edge AI collaboration is already assigned as AI-046 and sourced to the same official SLB release. Do not count as a new initiative; merge any incremental details into AI-046."], "duplicates": ["AI-046"]}, {"key": "viridien-evergreen-data", "outcome": "supported as commercial product launch", "findings": ["Launch and AI-assisted subsurface data preparation capability supported. Label as a vendor product, not an operator deployment; no customer/results disclosed."], "duplicates": []}, {"key": "tgs-seismic-foundation-contract", "outcome": "supported as vendor-reported contract award", "findings": ["Multi-year contract for an unnamed international operator and seismic interpretation scope supported. No value, customer identity or outcome is disclosed."], "duplicates": []}, {"key": "woodside-trion-ai-drilling", "outcome": "supported as planned contract scope", "findings": ["SLB states award covers 18 ultra-deepwater wells and includes AI-enabled drilling; services were scheduled to start in early 2026. Keep source attribution as SLB, and do not infer work actually commenced or that the contract produced measured results."], "duplicates": []}, {"key": "technip-dap-ai-for-all", "outcome": "supported with economics/date caveat", "findings": ["Technip’s 2025 letter supports DAP launched in 2025, advanced analytics/LLM use, AI for All implementation and a €100m cost-savings-plus-incremental-revenue target by 2028. The target is not achieved savings; no workflow-level result or AI-only attribution is provided."], "duplicates": []}], "summary": "Six of seven additions are distinct and source-supported within the stated stage limits. SLB/Qualcomm duplicates existing AI-046 and should not be counted as a new record. Key restrictions: Lucy is planned pilot; Woodside Trion is a contracted/scheduled scope with no start or outcome confirmation; Technip’s €100m is a combined 2028 target; Viridien and TGS claims are vendor-reported products/contracts rather than operator deployments."}, "new_profile_audits": [{"id": "AI-194", "company": "Expand Energy", "key": "expand-leucipa-lucy", "outcome": "Independent review completed", "reviewer_outcome": "supported with stage limitation", "findings": [{"field": "Independent review", "issue": "Leucipa is described as AI-powered and deployment begins through the Operations Support Center. Lucy is explicitly a pilot (“will also pilot”), so do not write that Lucy is already deployed.", "correction": "", "source_url": ""}], "residual_gaps": ["A deployment start and assistant pilot are different stages.", "No quantified production uplift or cost saving is established.", "Do not equate the broader vendor platform footprint with Expand’s actual deployment."], "source_urls_reviewed": ["https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-195", "company": "Expand Energy", "key": "expand-snowflake-drilling-ml", "outcome": "Independent review completed", "reviewer_outcome": "supported", "findings": [{"field": "Independent review", "issue": "Issuer source supports ML models for forecasting and real-time data/model hosting for drilling, plus SCADA alerts to the Operations Support Center. No result metric or specific model detail is disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Data integration benefits cannot be attributed entirely to AI.", "No independently measured drilling or failure-rate improvement is provided."], "source_urls_reviewed": ["https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-196", "company": "Viridien", "key": "viridien-evergreen-data", "outcome": "Independent review completed", "reviewer_outcome": "supported as commercial product launch", "findings": [{"field": "Independent review", "issue": "Launch and AI-assisted subsurface data preparation capability supported. Label as a vendor product, not an operator deployment; no customer/results disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Data-readiness claims are vendor statements.", "No time-saving benchmark, accuracy test or contract economics provided."], "source_urls_reviewed": ["https://www.viridiengroup.com/company/newsroom/news/viridien-redefines-subsurface-data-readiness-scale-evergreen-data-solution"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-197", "company": "TGS", "key": "tgs-seismic-foundation-contract", "outcome": "Independent review completed", "reviewer_outcome": "supported as vendor-reported contract award", "findings": [{"field": "Independent review", "issue": "Multi-year contract for an unnamed international operator and seismic interpretation scope supported. No value, customer identity or outcome is disclosed.", "correction": "", "source_url": ""}], "residual_gaps": ["Contract award establishes commercial adoption, not measured interpretation quality.", "No model size, benchmark, reserve impact or financial terms disclosed."], "source_urls_reviewed": ["https://www.tgs.com/press-releases/tgs-awarded-multi-year-ai-seismic-foundation-model-contract"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-198", "company": "Woodside", "key": "woodside-trion-ai-drilling", "outcome": "Independent review completed", "reviewer_outcome": "supported as planned contract scope", "findings": [{"field": "Independent review", "issue": "SLB states award covers 18 ultra-deepwater wells and includes AI-enabled drilling; services were scheduled to start in early 2026. Keep source attribution as SLB, and do not infer work actually commenced or that the contract produced measured results.", "correction": "", "source_url": ""}], "residual_gaps": ["The cited April2025 release is an award and forward schedule, not a current progress certificate.", "No AI-attributable efficiency or well-quality result is disclosed."], "source_urls_reviewed": ["https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-first-quarter-2025-results"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-199", "company": "Technip Energies", "key": "technip-dap-ai-for-all", "outcome": "Independent review completed", "reviewer_outcome": "supported with economics/date caveat", "findings": [{"field": "Independent review", "issue": "Technip’s 2025 letter supports DAP launched in 2025, advanced analytics/LLM use, AI for All implementation and a €100m cost-savings-plus-incremental-revenue target by 2028. The target is not achieved savings; no workflow-level result or AI-only attribution is provided.", "correction": "", "source_url": ""}], "residual_gaps": ["The web page does not expose an exact publication date.", "Financial target combines digital cost savings and incremental revenue; it is not AI-only realized savings.", "Do not treat the workforce program as evidence all employees are active AI users."], "source_urls_reviewed": ["https://investors.technipenergies.com/shareholder-information"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-189", "company": "KazMunayGas", "key": "kmg-refinery-ai-predictive-maintenance", "outcome": "Independent review completed", "reviewer_outcome": "supported with stage caveats", "findings": [{"field": "stage/timeline", "issue": "The May 22, 2026 KMG release distinguishes 2025 pilots at Atyrau and Pavlodar from 2026 rollout waves. It says Shymkent began compressor monitoring in April 2026, Atyrau/Pavlodar each planned >100 critical units over three stages, and 10 units were connected at the polypropylene plant. The packet preserves these distinct scopes and does not claim full rollout complete.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "asset-level workflow / date", "issue": "The Atyrau refinery’s Dec. 25, 2025 release says its Honeywell pilot launched in October (in context, October 2025), reports compressor anomaly detection before scheduled maintenance, and cites a separate pump-inlet pressure anomaly. The packet describes the compressor case without asserting quantified avoided downtime.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["KMG and its subsidiary are the reporting parties; no independent performance validation located.", "ANPZ article omits the year of the October pilot start; December 2025 publication does not justify a stronger date claim.", "ANPZ’s 28-unit initial circuit and KMG’s >100-unit phased plans are different scopes and should not be summed or conflated.", "No model accuracy, avoided downtime, savings or spend disclosed."], "source_urls_reviewed": ["https://www.kmg.kz/ru/press-center/press-releases/npz-prediktivnaya-analitika-vnedrenie/", "https://www.anpz.kz/en/press_center/news/76904/ii-kontrol-oborudovaniya-na-atyrauskom-npz-prediktivnaya-analitika-v-deystvii/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-190", "company": "QatarEnergy LNG", "key": "qatarenergy-lng-machinery-fleet-ai", "outcome": "Independent review completed", "reviewer_outcome": "supported; review-status update needed", "findings": [{"field": "source review status", "issue": "The cited QatarEnergy LNG Issue 167 PDF opened as a complete 13-page primary-source document in this review. Page 5 explicitly says “expand the successful AI pilot across all assets” and “use of legacy data for predictive purposes in rotating equipment.” Packet limitations are accurate, but its statement “Only an excerpt of the magazine PDF was reviewed” is stale and should be updated to full-text reviewed.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "stage", "issue": "The primary source supports a completed/successful pilot and an expansion objective; it does not support completed asset-wide deployment, nor identify the pilot asset, dates, or results. Packet stage is correctly bounded.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The document indicates a pilot and future expansion, not confirmed asset-wide deployment.", "Partner, location, AI technique, economics and operating metrics are not disclosed.", "The PDF is available and searchable in full; the relevant strategy passage is brief and does not disclose technical or outcome detail."], "source_urls_reviewed": ["https://www.qatarenergylng.qa/Portals/0/DNNGalleryPro/uploads/2025/7/16/ThePioneer167EN_DS.pdf"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-191", "company": "Ambyint", "key": "ambyint-infinitypl-methane-upgrade", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; correct two overstatements", "findings": [{"field": "funding metric", "issue": "The two linked sources state $4.1 million, not $4.08 million. The more precise $4.08m value in disclosed_results is unsupported by this source pair; either use $4.1m or supply an official source that establishes $4.08m.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "workflow / human role", "issue": "Both linked sources support an existing InfinityPL platform upgrade to predict inefficiencies, detect complex anomalies and reduce continuous manual intervention. Neither linked source states that the new upgrade will autonomously manage well setpoints or describes operator approval/override controls. Remove or mark setpoint control as undisclosed unless a further source is cited.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "deployment stage / results", "issue": "The sources consistently describe a funded upgrade and Tourmaline as a planned deployment partner. Up-to-90% methane-venting reduction is prospective, not an achieved outcome; the packet handles this distinction correctly. The existing 17-producer footprint refers to the existing platform, not the upgraded methane feature.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The up-to-90% figure is an expected reduction, not a verified result.", "No Tourmaline deployment date, number of wells, baseline, measurement protocol, or post-deployment result is disclosed.", "Technical and safety-control details are not disclosed.", "Sources do not establish autonomous setpoint management or describe operator approval/override controls for the upgrade."], "source_urls_reviewed": ["https://www.ambyint.com/insights/advance-ai-methane-emission-reduction-technology/", "https://www.eralberta.ca/media-releases/alberta-invests-7-million-to-tackle-methane-emissions-in-the-oil-and-gas-sector/"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-192", "company": "Kpler", "key": "kpler-ai-chartering-email-extraction", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; remove unsupported current-page metric", "findings": [{"field": "disclosed_results", "issue": "The packet’s “99% email parsing accuracy” is not supported by the linked current Kpler AI page. Directly searched that page for “accuracy” and “99”; no accuracy metric appeared. Remove the 99% result unless an observed source supports it.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "stage / dates", "issue": "Kpler’s Oct. 16, 2025 event page supports email/PDF/spreadsheet parsing and describes the AI chartering agent as “launching Q4.” The packet correctly states that a dated launch or commercial availability was not independently confirmed; preserve Q4 as planned, not achieved.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "source scope", "issue": "The current solutions page is undated and supports general conversational Kpler AI/data foundations, not email parsing accuracy or the dated launch status of the agent. The packet correctly treats its date as unknown, but its support array should not use it to support the 99% result.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The Q4 2025 agent was announced as planned; launch/commercial availability is not confirmed by dated primary evidence reviewed here.", "No evidence that recommendations execute trades or replace charterer approval.", "No email-parsing accuracy figure appears in the reviewed current product page; no independent accuracy or productivity metric was identified."], "source_urls_reviewed": ["https://www.kpler.com/event/kpler-maritime-briefing-2025-navigating-risk-harnessing-ai", "https://www.kpler.com/solutions/ai"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-193", "company": "Kodiak Gas Services", "key": "ifs-kodiak-material-replenisher", "outcome": "Independent review completed", "reviewer_outcome": "partially supported; deployment date and results attribution need correction", "findings": [{"field": "deployment date", "issue": "The Nov. 27, 2025 IFS Cloud release gives a conditional scenario (“If half of our workforce engages…”) but does not state that Kodiak had rolled out Material Replenisher by that date. The Apr. 23, 2026 IFS Loops release says Kodiak “rolled out” the agent. Change the deployment timeline to “by 2026-04-23 / date not specified”; keep Nov. 2025 only for the conditional estimate.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "disclosed_results", "issue": "The 2025 release attributes $3m/year and 90,000 hours to a conditional assumption of half the workforce using the agent daily. The 2026 release changes attribution language and calls $3m annual ROI and 90,000 hours returned “the result” after saying Kodiak rolled out the agent. The packet currently says the 2026 release merely repeats potential estimates and does not treat them as realized; that understates the later vendor claim. Separate the 2025 conditional estimate from the 2026 IFS-reported realized result, while making clear the latter remains vendor-reported and lacks methodology or independent validation.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}, {"field": "deployment scope / workflow", "issue": "The 2026 source confirms 4.5 million horsepower, 800 technicians, prior time spent searching for parts, and conversational find-and-order functionality. The linked two-source set does not substantiate the packet’s “150+ warehouses,” nearby-stock search, or specific IFS Cloud inventory/warehouse-location data integration. These may be true but require a cited source; otherwise remove or mark undisclosed.", "correction": "Current dossier incorporates the accepted correction or stated limitation.", "source_url": ""}], "residual_gaps": ["The April 2026 result is vendor-reported; no measurement method, baseline, timeframe or independent customer validation is provided.", "The 2025 release gives the same numbers as a conditional scenario and does not date the rollout.", "Warehouse count, nearby-stock search, data-system integration details, adoption rates and control boundaries are not disclosed in the linked sources."], "source_urls_reviewed": ["https://www.ifs.com/en/insights/news/ifs-cloud-25r2", "https://www.ifs.com/en/insights/news/ifs-loops-launches-agent-studio"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}, {"id": "AI-188", "company": "Permian Resources", "key": "permian-resources-ai-subsurface-placement", "outcome": "Independent review completed", "reviewer_outcome": "supported with disclosure limits", "findings": [{"field": "Independent review", "issue": "The November 5, 2025 presentation, printed slide 5, explicitly describes AI-driven subsurface insight and precision wellbore placement. Adjacent Haley outperformance is not attributable to AI. Corrected department to reservoir engineering with exploration and drilling secondary links.", "correction": "", "source_url": ""}], "residual_gaps": ["Single company presentation disclosure, with no model, data, validation, or deployment-count detail.", "Do not attribute adjacent Haley outperformance to AI."], "source_urls_reviewed": ["https://permianres.com/wp-content/uploads/2025/11/PR-Q325-Earnings-Presentation_FINAL.pdf"], "checks": ["Dates", "AI basis", "stage", "metrics", "duplicates"]}]}, "followup": {"date": "2026-09-28", "added": ["AI-200", "AI-201", "AI-202"], "enriched": ["AI-006"], "method": "Three Luna research assignments; separate candidate and workflow-mapping audits; root synthesis and integration. Useful undated and pre-window cases remain supplemental evidence, excluded from the new-profile count."}, "missing_followup": {"date": "2026-09-29", "baseline_profiles": 202, "added": ["AI-203", "AI-204", "AI-205", "AI-206", "AI-207", "AI-208", "AI-209", "AI-210", "AI-211", "AI-212"], "enriched": ["AI-165", "AI-125", "AI-143"], "classification_repairs": 63, "eqt_call_gap_closed": true, "search_log": [{"company": "Marathon Petroleum / Marathon Pipe Line", "query": "Targeted company / AI public-source search", "result": "Hit: AI-assisted pipeline ROW threat monitoring; added above.", "url": "https://www.marathonpetroleum.com/Newsroom/Company-News/Our-People-How-Paige-Chenevey-drives-success-at-Marathon-Pipe-Line/"}, {"company": "ORLEN Group", "query": "Targeted company / AI public-source search", "result": "Hit: enterprise AI rollout; refinery data-refining/prediction pilot; AI-based tanker inspection challenge is only a challenge in the July 2025 source and is not counted as deployed AI.", "url": "https://www.orlen.pl/en/about-the-company/media/press-releases/current/2025/july-2025/orlens-transformation-to-gain-momentum-with-cees-largest-ai-deployment"}, {"company": "Apache / APA", "query": "Targeted company / AI public-source search", "result": "Hit: executive interview makes broad AI claims for emissions reporting, logistics and reservoir decisions; implementation specifics absent.", "url": "https://egyptoil-gas.com/wp-content/uploads/2025/10/Egypt-Leading-Oil-Gas-Excellence.pdf"}, {"company": "Murphy Oil", "query": "Targeted company / AI public-source search", "result": "Hit: AI/ML in cybersecurity monitoring; not hydrocarbon process AI.", "url": "https://www.murphyoilcorp.com/wp-content/uploads/2025/08/2025_Murphy-Sustainability-Report.pdf"}, {"company": "HF Sinclair", "query": "Targeted company / AI public-source search", "result": "No primary operator disclosure located in this pass. Search surfaced recruiting and employee LinkedIn references to AI/ML process optimization and predictive maintenance, but no operator-authored public implementation report in scope; not counted.", "url": "https://www.hfsinclair.com/"}, {"company": "Harbour Energy", "query": "Targeted company / AI public-source search", "result": "No qualifying operator AI disclosure located in targeted official-site search; searches returned no-hit.", "url": "https://www.harbourenergy.com/"}, {"company": "DNO", "query": "Targeted company / AI public-source search", "result": "No qualifying operator AI disclosure located in targeted official-site search; searches returned no-hit.", "url": "https://www.dno.no/"}, {"company": "MOL Group", "query": "Targeted company / AI public-source search", "result": "No operating AI deployment located; official governance material mentions AI mainly as cyber risk, not a deployed workflow.", "url": "https://molgroup.info/en/"}, {"company": "SOCAR", "query": "Targeted company / AI public-source search", "result": "No qualifying company deployment located. SOCAR-hosted journals surfaced academic AI studies, which were not company implementations.", "url": "https://www.socar.az/"}, {"company": "Sonatrach", "query": "Targeted company / AI public-source search", "result": "No in-window deployed initiative located. Official 2025 newsletter references a July 2024 AI-industry symposium, outside window and not evidence of deployment.", "url": "https://sonatrach.com/wp-content/uploads/2025/06/SH-NEWS-VF-N%C2%B045.pdf"}, {"company": "Galp", "query": "Targeted company / AI public-source search", "result": "Official Galp/IBM AI seismic-advisor disclosure found, but original press release is from 2017 (outside window); current 2025 AI solar analytics relates to power/renewables and was excluded from O&G operator initiatives.", "url": "https://www.galp.com/corp/en/media/press-releases/press-release/id/834/galp-and-ibm-build-cutting-edge-artificial-intelligence-advisor-to-speed-up-appraisal-of-oil-and-gas-prospects-and-discoveries"}, {"company": "OMV Petrom", "query": "Targeted company / AI public-source search", "result": "2025 annual report search mostly returned AI governance/cyber oversight; broader OMV 2025 AI record already exists in corpus (AI-084, AI-132–135), so no duplicate subsidiary record added.", "url": "https://www.omvpetrom.com/downloads/2026/03/bb67c931-07a4-7243-67e5-8b861da2bd08/10%20OMV%20Petrom%20Group%20Annual%20Report%202025%20en%20full.pdf"}, {"company": "Marathon Petroleum legacy refining AI/ML", "query": "Targeted company / AI public-source search", "result": "Earlier 2024 article on refinery AI/ML predates window; not counted. 2025 pipeline use case above qualifies.", "url": "https://www.marathonpetroleum.com/Newsroom/Company-News/Mechanics-and-Maintenance-Les-Davis/"}, {"company": "Apache / APA Palantir partnership", "query": "Targeted company / AI public-source search", "result": "Excluded from date window: AI partnership extension announced 2024-09-25, four days before scope start; later vendor repost does not change disclosure date.", "url": "https://palantir2020ipo.q4web.com/news-details/2024/APA-Corporation-Expands-Partnership-with-Palantir-To-Leverage-AI-Technology-Across-Their-Oil-and-Gas-Operations/default.aspx"}, {"company": "Workflow / company search", "query": "oil gas AI accounts payable invoice processing operator deployment 2025 named; oil gas AI joint interest billing JIB automation company 2025", "result": "No newly verified named AI deployment from AP/JIB found in this pass. Results surfaced product marketing and older dated Red Wolf/Enverus case (2020-origin story), so excluded from new/enriched set as not a window disclosure.", "url": ""}, {"company": "Workflow / company search", "query": "oil gas AI land title legal contract workflow operator deployment 2025; oil gas operator AI revenue accounting deployment 2025", "result": "No qualifying newly disclosed named title/legal or live revenue-accounting deployment verified. Existing AI-057 lease provision processing is a dated product capability, not customer deployment. Winn/Collide describes revenue statement extraction only as planned.", "url": ""}, {"company": "Workflow / company search", "query": "oil gas AI nomination scheduling trading logistics deployment operator 2025; AI chartering vessel email parsing energy", "result": "Kpler event source verifies described chartering capabilities but no customer deployment; agent launch was planned for Q4 2025 and later status is unverified. Existing AI-103 has a Williams one-month marketing test; AI-170 pipeline decision support and AI-192 Kpler product record already exist.", "url": ""}, {"company": "Workflow / company search", "query": "oil gas AI procurement customer operator deployment 2025; AI sourcing optimization named energy company", "result": "C3 AI customer story provides detailed Baker Hughes sourcing workflow; Microsoft customer story provides TotalEnergies BuyerCompanion specifics and MARGE metrics. Both corresponding records already existed and are enriched.", "url": ""}, {"company": "Workflow / company search", "query": "private equity backed oil gas portfolio AI operational deployment named company 2025 2026", "result": "Searches surfaced sponsor/portfolio ownership and general PE AI strategy reports but no primary source naming an O&G PE-backed portfolio company with an operational AI workflow and attributable operating outcomes. Winn Resources is private but sponsor status is not stated in the source and therefore is not counted as PE-backed.", "url": ""}, {"company": "Workflow / company search", "query": "AI mineral title accuracy oil gas attorney validated; AI title land legal oil gas operator deployment", "result": "Lead: Toptal case study at https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy reports an unnamed oil-and-gas technology firm, 200,000 records, 2,000 attorney-validated answers, and 93% fewer false positives. Source is anonymous and undated, so disclosure window, customer identity, production status, baseline definition, and relevance as an operator deployment are unverified; excluded from counted records. AskLandon is an undated product site with no customer outcome; excluded.", "url": ""}, {"company": "Workflow / company search", "query": "Diversified Energy OneDEC AI producer statements revenue distribution contract repository", "result": "Found operator's May 2026 sustainability report with explicit workflow details and governance architecture; no workflow metric or rollout milestone. Added in-window record.", "url": ""}, {"company": "Workflow / company search", "query": "Trafigura ZeroNorth AI fleet rollout July 2025 official company source", "result": "Found named 9 July 2025 operator press release: ZeroNorth AI voyage optimization planned across >350 controlled vessels, plus emissions and reporting tools. Rollout planned; no realized metric.", "url": ""}, {"company": "Workflow / company search", "query": "SM Energy Midland Basin gas-lift ML AI step-rate test injection rates Rule 425 transcript 2025", "result": "Found 20 Nov 2025 COO transcript filed by SM Energy: compression-linked ML learns from step-rate tests and recommends injection rates; deployed in production in Midland Basin; Civitas extension only planned.", "url": ""}, {"company": "Workflow / company search", "query": "Vital Energy AI ESP SoftServe AWS deep neural network production uplift case", "result": "Found named vendor case with ML workflow and vendor-reported BPD numbers, but no publication date; segregated as undated supplemental, outside in-window count.", "url": ""}, {"company": "Workflow / company search", "query": "Mercuria CurlHG ML OCR shipping documentation ETRM AI", "result": "Found Mercuria first-party CSR technology disclosure describing CurlHG OCR/ML inside ETRM. Undated; segregated as supplemental; no metric.", "url": ""}, {"company": "Workflow / company search", "query": "AI deployment primary sources Amplify Energy, Diversified Energy, BKV, Mach Natural Resources, Vital Energy, Hilcorp, Flywheel Energy, Crescent Energy, Civitas, historical PDC Energy", "result": "Diversified, Vital, and Civitas/SM leads surfaced as above. Amplify and Mach: searched company names with AI/ML/deployment and inspected available operation/investor material; no in-window named operational AI case located. BKV surfaced governance / legacy SCADA consolidation and 2024 sustainability report discussing tests/MVPs in predictive maintenance, emissions and GenAI, but not a sufficiently specified new AI workflow with in-window operating outcomes. Hilcorp, Flywheel: no qualifying workflow found; Flywheel search returned lubricant-engine-oil case unrelated to AI. Crescent: no named AI workflow in company/SEC results. Historical PDC is outside its independent existence as Chevron acquired it in 2023; no in-window separate PDC deployment located.", "url": ""}, {"company": "Workflow / company search", "query": "W Energy AI field data customer case; Quorum AI named operator case accounting; PakEnergy AI customer operator; Enverus AI customer case private oil gas company", "result": "W Energy search found named field data gathering/customer anecdotes (including Range safety compliance) but no source establishing these cases as AI, so excluded. Quorum Venator story describes ERP/data-foundation modernization and future AI readiness, not AI deployment. PakEnergy Ironwood case describes AP automation/document capture but not AI explicitly, so excluded from AI-specific tally. Enverus case hits were older or business-process outsourcing rather than AI; excluded.", "url": ""}, {"company": "Workflow / company search", "query": "Trafigura, Vitol, Gunvor, Mercuria AI deployment energy trading shipping OCR ML official sources", "result": "Trafigura and Mercuria records above. Vitol: targeted official-site searches did not surface named deployed AI workflow. Gunvor: hits were 2025 management/financial disclosures, a generic privacy notice allowing AI tools, and undated product-like AI glossary; no named operating workflow. No-hits are limited to indexed public search and do not prove absence.", "url": ""}, {"company": "Snam", "query": "Oil and gas AI technical plant-document extraction", "result": "Named operator and full technical preprint found; candidate retained as development/evaluation, not fleet deployment.", "url": "https://arxiv.org/html/2511.08609v1"}, {"company": "WEF external case catalog", "query": "Reconcile AI Energy Foresight Tool with oil and gas register", "result": "Publisher links a 130+ case energy-wide Tableau catalog. Static extraction unavailable and PDF timed out. Catalog remains unreconciled; no new cases counted.", "url": "https://www.weforum.org/publications/from-paradox-to-progress-a-net-positive-ai-energy-framework/"}], "supplemental": [{"company": "Vital Energy", "title": "Human-approved recommendations and a nine-well autonomous pump trial", "url": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted", "status": "Dated evidence update · existing supplemental workflow · not a new profile", "finding": "JPT’s 1 October 2024 account adds detail to the existing vendor case: recommendations reached more than 200 wells, while autonomous setpoint writes were trialed on nine. Six wells with the highest autonomous utilization exceeded forecast oil production by more than 3%.", "limit": "Operator account reported by JPT/SPE, not an independently validated trial. Forecast comparison is not a controlled counterfactual. Reliability improvements were expressly not fully attributed to AI. No implementation cost or payback disclosed; trial dates and control safeguards remain unspecified.", "prior_evidence": {"company": "Vital Energy", "title": "Deep-neural-network electrical-submersible-pump optimization recommendations (phase-one case)", "url": "https://www.softserveinc.com/en-us/resources/boost-oil-production-by-optimizing-esp-with-ai", "status": "Undated · outside new-profile count", "finding": "Deep-neural-network and optimization models forecast oil/gas/water production, simulate controls, and recommend daily ESP operating parameters within limits such as motor temperature, voltage, intake pressure, gas and water volumes. Users see simulations and can provide feedback/constraints.", "limit": "Undated vendor case; results vendor-reported and cannot be placed in 2024-09-29–2026-09-29 window. This is recommendation/testing evidence, not proof of autonomous control or independently measured uplift."}, "source_id": "SW3-006"}, {"company": "Mercuria", "title": "CurlHG ML/OCR extraction of shipping documents into ETRM", "url": "https://csr.mercuria.com/home/prosperity/technology-innovation/", "status": "Undated · outside new-profile count", "finding": "OCR and machine learning read and digitize shipping documents inside Mercuria's ETRM systems, removing manual re-entry of commodity characteristics and intended to improve consistency, accuracy, compliance, and traceability in global operations.", "limit": "Undated operator primary page; cannot establish in-window timing. Machine learning is explicit, but there are no measured results or implementation controls described."}, {"company": "Oman LNG", "title": "RT-OP process optimization", "url": "https://www.yokogawa.com/library/resources/references/suc-omanlng-rt-op/", "status": "AI not established · excluded", "finding": "First-principles process optimization and APC integration is described.", "limit": "The reviewed case does not establish AI/ML. It is undated; completion in 2024 does not establish an in-window disclosure."}, {"company": "Unnamed midstream operator", "title": "Saltwater-disposal pump optimization", "url": "https://jpt.spe.org/case-study-ai-driven-optimization-of-swd-pumping-efficiency-using-kpi-based-time-series-analytics", "status": "Already in report library · GAPF002", "finding": "Detailed provider case adds operating constraints, including the tradeoff between slower pumping and throughput.", "limit": "Unnamed customer; already held as technical literature, so not counted as a newly identified company."}, {"company": "Unnamed LNG terminal", "title": "Reinforcement-learning scheduling study", "url": "https://www.sciencedirect.com/science/article/pii/S009813542500523X", "status": "Abstract-level lead · not counted", "finding": "Research reports a shadow scheduling comparison.", "limit": "Full methods and site identity have not been verified; no inference of sustained commercial deployment."}, {"company": "Unnamed title-technology firm", "title": "Mineral-title model evaluation", "url": "https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy", "status": "Anonymous and undated · not counted", "finding": "Provider describes attorney-validated evaluation data for mineral-title extraction.", "limit": "Identity, disclosure date and operator deployment remain unresolved."}], "rediscovered": ["AI-167", "AI-192"], "findings": [{"title": "Financial and commercial workflows were underrepresented.", "text": "Diversified describes producer-statement and contract-data work; Trafigura adds a planned shipping deployment. The records distinguish platform capability from verified use and realized savings.", "ids": ["AI-210", "AI-211", "AI-125"]}, {"title": "Some company omissions were search and labeling gaps.", "text": "Marathon Pipe Line, ORLEN, APA, Murphy, Snam and SM Energy were absent. Winn’s customer case existed under its vendor name and is now findable under Winn Resources.", "ids": ["AI-203", "AI-204", "AI-212", "AI-125"]}, {"title": "Operating evidence needs a denominator.", "text": "TotalEnergies reports five predicted events in an initial pilot. PETRONAS reports monitored equipment count. Neither source supplies implementation cost, a false-alarm rate or a cash-payback calculation.", "ids": ["AI-207", "AI-209"]}], "unresolved": ["A complete company-by-quarter earnings-call and presentation census remains open beyond the EQT call ledger.", "Named failed deployments and measured project costs/payback remain scarce in the reviewed public evidence.", "Private-equity ownership still needs to be traced to named portfolio deployments and comparable operating outcomes.", "Land/title, joint-interest billing, revenue accounting and trade settlement still lack enough named, dated deployment evidence.", "The WEF energy-wide external case catalog remains unreconciled; its 130+ cases are not counted as oil-and-gas initiatives."], "method": "Three Luna research agents, targeted second passes, independent source checks, root deduplication and integration. No-hit searches establish only the result of this pass; reported company results are not independently audited."}, "swarm2": {"date": "2026-09-29", "baseline_profiles": 212, "summary": "This pass adds named trading, field-service, gas-processing and commercial-planning workflows. The stronger disclosures reveal operating scale, but still rarely disclose the cost of achieving it.", "method": "Three Luna researchers, follow-up source audits, original-PDF review, visual checking of reported figures and deduplication against all 212 baseline profiles and their detailed research packets. This remains a targeted review, not a global census.", "findings": [{"title": "Useful AI is reaching commercial operations.", "text": "Koch describes chat-to-ETRM gas-trade capture. Petrobras discloses machine-learning revenue forecasting. These address specific work and decisions; faster entry and forecast accuracy do not establish cash returns.", "ids": ["AI-213", "AI-221"]}, {"title": "A sponsor relationship can be traced to an actual workflow.", "text": "Sentinel Peak and Quantum co-developed steam-injection optimization. The sponsor report supplies a named portfolio case, but its operating results are from 2023 and explicitly unverified.", "ids": ["AI-214"]}, {"title": "The financial claims need different labels.", "text": "ADNOC Gas reports value delivered to date. Petrobras reports captured value and avoided production losses. Occidental reports estimated annual benefits across a mix of deployed and unfinished projects. These figures cannot be added into an industry savings total.", "ids": ["AI-216", "AI-219", "AI-153"]}], "excluded": [{"title": "Shell reliability: already captured in the detailed register", "reason": "AI-076 already includes the June 2026 C3 AI source, equipment coverage and vendor economic-value claim. An undated Shell transcript is useful context but does not establish a new dated result.", "url": "https://www.shell.com/what-we-do/digitalisation.html"}, {"title": "Petrobras: neighboring digital results are not automatically AI", "reason": "Slide 158 also reports procurement automation, reservoir-risk solutions and refinery digital twins. It does not explicitly establish AI for those individual workflows; their dollar figures are excluded from these AI records.", "url": "https://www.sec.gov/Archives/edgar/data/1119639/000129281425004085/pbr20251127_6k1.htm"}, {"title": "Anonymous invoice, land/title and marine cases", "reason": "Practical vendor descriptions were found, but named customer, deployment date or explicit AI evidence was missing. These do not close the operator-attribution gap.", "url": "https://smartbridge.com/invoice-processing-automation-case-study/"}, {"title": "Weatherford: projected value and unnamed customer", "reason": "A vendor case describes a West Texas pilot and expansion, but the customer is unnamed and the headline annual value is prospective. It is not added as a named deployment with realized savings.", "url": "https://www.weatherford.com/real-results/production/foresite-production-optimization-platform-drives-$17-7m-in-annual-incremental-value-through-improve/"}, {"title": "WEF case catalogue: source link now returns 404", "reason": "The official publication’s Tableau link resolved to a 404 page during this review. Its 130+ energy-wide cases remain unreconciled and are not counted as oil-and-gas initiatives.", "url": "https://www.weforum.org/publications/from-paradox-to-progress-a-net-positive-ai-energy-framework/"}], "gaps": ["Project-level implementation cost, ongoing support cost and realized net cash benefit remain absent from the newly reviewed cases.", "No adequately documented named failed deployment was established in this pass; selective public disclosure remains a limitation.", "Named, dated land/title, JIB and AFE AI deployments remain thinly evidenced.", "Sentinel Peak improves PE portfolio attribution, but a complete sponsor-by-portfolio-company census is still missing.", "The global company-quarter earnings-call and investor-presentation census remains incomplete."], "search_log": [{"target": "Existing register and prior gap-file deduplication", "result": "Confirmed existing AI-076 Shell reliability, AI-146 Petrobras Smart Tocha, AI-040 ADNOC ENERGYai; exact economics duplicates AI-141 Aramco Fadhili, AI-201 Chevron Kaybob, and AI-202 ConocoPhillips Montney excluded. No matching ADNOC Gas RTO record found.", "url": "", "queries": []}, {"target": "Oil and gas AI savings/payback/value search and primary-source opens", "result": "Retained Shell (June 2026 C3 follow-up plus Shell-hosted executive transcript), ADNOC Gas FY2025 integrated-report highlight, and Petrobras November 2025 6-K/business-plan slide. No disclosed cost/payback figure among these three.", "url": "", "queries": []}, {"target": "Aramco Fadhili; Chevron Kaybob; ConocoPhillips Montney economics", "result": "Good concrete economics but already represented as AI-141 / AI-201 / AI-202 in the existing 212-profile dataset; excluded as duplicates.", "url": "", "queries": []}, {"target": "Weatherford ForeSite $17.7m value case", "result": "Primary vendor source says an unnamed Fortune 500 West Texas operator piloted 80+ wells, expanded to nearly 1,000, and expects at least $17.7m annualized value only after full rollout. Rejected as a new named-company disclosure and because the headline dollars are prospective, not realized cash.", "url": "", "queries": []}, {"target": "Petrobras digital value plan and exact phrases", "result": "Retained as a measured value disclosure enriching AI-146. Direct official PDF page fetch was too large for the browser reader; relevant slide 158 was also retrieved in text/indexed copies, while SEC filing copy was opened.", "url": "", "queries": []}, {"target": "Eni 2025 energy/just-transition reporting, per parallel note", "result": "Not retained: primary PDF access and exact page/result attribution not independently verified during this pass; avoid duplicating AI-030 until source text is directly verified.", "url": "", "queries": []}, {"target": "Quantum Capital Group / Quantum Energy Partners; Sentinel Peak Resources", "result": "One qualifying and additional named operator workflow found.", "url": "", "queries": ["Sentinel Peak Resources AI steam injection Quantum 2025", "Sentinel Peak Resources Quantum portfolio company AI", "Sentinel Peak Resources artificial intelligence steam injection emissions AI"]}, {"target": "EnCap Investments", "result": "No qualifying candidate found in this limited search; absence is not evidence that none exists.", "url": "", "queries": ["site:encapinvestments.com AI artificial intelligence portfolio oil gas workflow"]}, {"target": "EIG Partners / Breakwater / Repsol upstream", "result": "No qualifying in-window workflow verified.", "url": "", "queries": ["site:eigpartners.com AI oil gas portfolio artificial intelligence operations"]}, {"target": "Tailwater Capital portfolio", "result": "No qualifying in-window operational workflow found in the surfaced materials.", "url": "", "queries": ["site:tailwatercapital.com artificial intelligence portfolio oil gas AI"]}, {"target": "Carlyle energy portfolio", "result": "No qualifying O&G/energy portfolio workflow verified in the surfaced sources.", "url": "", "queries": ["site:carlyle.com oil gas AI portfolio company deployed AI workflow oilfield 2025 2026"]}, {"target": "Warburg Pincus, NGP, Kimmeridge, Blackstone, Apollo", "result": "No additional qualifying candidate verified. Detailed per-firm source review for Warburg, Kimmeridge, Blackstone, Apollo and NGP was not completed in this swarm; gap explicitly retained.", "url": "", "queries": ["site:warburgpincus.com oil gas portfolio artificial intelligence AI operations", "site:ngpenergycapital.com AI portfolio company oil gas artificial intelligence", "Kimmeridge AI portfolio operator deployment oil gas 2025", "Blackstone Enverus AI oil and gas 2025 portfolio company", "Apollo oil gas portfolio AI operations 2025 2026"]}, {"target": "Workflows", "result": "Identified Koch/ipushpull physical gas trade booking, land-title vendors without named end customers, and pre-window JIB digitization case.", "url": "", "queries": ["site:*.com oil gas AI land title division order accounting AI operator 2025", "AI JIB invoice accounting oil gas customer case study 2025 operator", "AI refinery operations LNG marine logistics oil gas operator 2025 AI customer", "AI landman title abstract oil gas company AI title workflow 2025"]}, {"target": "Workflows", "result": "Identified anonymous Smartbridge invoice automation, anonymized/unnamed marine pilots, ADNOC AiPSO, and broad C3 AI deployment claims without use-case attribution.", "url": "", "queries": ["2025 oil gas AI LNG cargo optimization operator named customer AI company", "AI maritime shipping oil tanker voyage optimization customer 2025 AI energy", "2025 refinery artificial intelligence named refinery customer AI", "oil gas invoice AI automation operator case study named 2025"]}, {"target": "Workflows", "result": "Full source text was opened for each; exact locators are included above for candidate and strongest exclusions.", "url": "https://ipushpull.com/blog/ipushpull-launches-llm-enabled-agent-to-automate-physical-gas-trade-booking-via-chat", "queries": []}, {"company": "WEF / Accenture", "query": "Follow the official 130+ AI-energy case repository from the WEF publication", "result": "Publication links to the Tableau AIEnergy2025/AIEnergy catalogue. Browser navigation resolved to Tableau 404 Not Found on review. This pass cannot reconcile its entries or claim an oil-and-gas case count.", "url": "https://www.weforum.org/publications/from-paradox-to-progress-a-net-positive-ai-energy-framework/", "catalogue_url": "https://public.tableau.com/app/profile/world.economic.forum.data.analytics/viz/AIEnergy2025/AIEnergy", "target": "WEF / Accenture"}, {"target": "Occidental proxy and earnings call", "result": "Official issuer documents establish the 2025 portfolio stage split and the separate routeless field workflow; estimated benefits are kept separate from realized results.", "url": "https://www.oxy.com/siteassets/documents/investors/2026-proxy-statement.pdf"}], "added": ["AI-213", "AI-214", "AI-215", "AI-216", "AI-217", "AI-218", "AI-219", "AI-220", "AI-221"], "enriched": ["AI-153", "AI-146"]}, "swarm3": {"date": "2026-09-30", "baseline_profiles": 221, "summary": "Six additional profiles cover pipeline power, aerial surveillance, water injection, drilling oversight, legal AI selection and workforce adoption. A dated Vital Energy account strengthens an existing supplemental case. The evidence distinguishes what is operating, what was only announced, and what still lacks measured economics.", "method": "Three Luna researchers and lead source review; cross-checks against 221 profiles, detailed research packets, supplemental cases and the report library. Research covers disclosures in the last 24 months through September 2026; earlier implementation dates are retained. This targeted pass does not complete a global census.", "findings": [{"title": "Operating scale needs a denominator.", "text": "ADNOC reports its drilling platform across more than 120 rigs. Williams reports employee AI use above 90%. The first measures assets covered and the second measures adoption; neither establishes autonomous operation or net cash benefit.", "ids": ["AI-226", "AI-224"]}, {"title": "Several useful applications sit inside established engineering work.", "text": "Enbridge’s power decisions and ONGC’s injection recommendations support existing operating choices. The disclosures identify practical uses, but provide no quantified savings or production uplift for these workflows.", "ids": ["AI-222", "AI-225"]}, {"title": "Legal AI now has a named upstream buyer; deployment evidence remains thin.", "text": "Harvey names Devon’s legal, land, transactional and compliance functions. Its dated selection announcement does not show go-live, workflow outcomes or cost savings.", "ids": ["AI-227"]}], "supplemental": [{"company": "Vital Energy", "title": "Human-approved recommendations and a nine-well autonomous pump trial", "url": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted", "status": "Dated evidence update · existing supplemental workflow · not a new profile", "finding": "JPT’s 1 October 2024 account adds detail to the existing vendor case: recommendations reached more than 200 wells, while autonomous setpoint writes were trialed on nine. Six wells with the highest autonomous utilization exceeded forecast oil production by more than 3%.", "limit": "Operator account reported by JPT/SPE, not an independently validated trial. Forecast comparison is not a controlled counterfactual. Reliability improvements were expressly not fully attributed to AI. No implementation cost or payback disclosed; trial dates and control safeguards remain unspecified.", "source_id": "SW3-006"}], "excluded": [{"title": "Dagang Oilfield (PetroChina/CNPC) machine-learning production prediction", "reason": "Superseded 1 October 2026: original-page decoding resolved; accepted as AI-236. Previous review: Promising: CNPC News Center article dated 2026-02-04 reports trial across multiple blocks and forecast-accuracy improvement of 15%. Full primary page fetch failed with Unicode decoding error, so no candidate accepted on search-returned article text alone.", "url": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml"}, {"title": "Oil India DRIVE 2.0 digital program", "reason": "FY2024-25 annual report says 11 prior digital projects were implemented but frames broader AI adoption and the Command-and-Control Centre under DRIVE 2.0 as planned; no named qualifying deployed AI workflow found in the report.", "url": "https://www.oil-india.com/files/financial_results_documents/OIL_India_Annual_Report_2024_25.pdf"}, {"title": "PETRONAS PROTEAN predictive rotating-equipment analytics", "reason": "Exact substantive duplicate of existing AI-209; annual-report update already captured the 350+ equipment scope, Malaysia/international deployment, predictive-failure workflow, and lack of outcome metric.", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf"}, {"title": "Petrobras AI workflows in the 2026-30 Business Plan", "reason": "The specific well-configuration ML, control-room predictive monitoring, domestic revenue forecast, and flare-control disclosures already match existing Petrobras dataset records; no additional nonduplicate operator workflow identified.", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0"}, {"title": "KPMG cognitive contract management for a U.S. oil and gas company", "reason": "Substantive named-workflow evidence and control via subject-matter-expert dashboard review, but the customer is explicitly anonymous and KPMG page provides no dated publication/update. Cannot meet named end-customer requirement or verify window.", "url": "https://kpmg.com/us/en/how-we-work/client-stories/energy-cognitive-contract-management.html"}, {"title": "Permian operator / BIS Grooper land-data extraction", "reason": "Customer is anonymized (described only as a large-cap independent producer headquartered in Midland). The accessible case page has no dated publication and its contents do not establish a new disclosure or update in the requested window.", "url": "https://www.casestudies.com/company/bis-inc/case-study/permian-operator-accelerates-acquisition-data-onboarding-with-grooper-ai-powered-land-data-extraction"}, {"title": "Chesapeake Energy / Enverus OpenContract, OpenTicket, OpenInvoice", "reason": "The page says OpenContract/OpenTicket implementation occurred in 2022; the case page itself has no publication/update date. Its named account (now Expand Energy) overlaps an existing customer identity in the research corpus. Evidence describes business-process automation and does not substantiate AI. Reported process adoption and discount capture are not a quantified realized net economic return.", "url": "https://www.enverus.com/customer-story/reduce-manual-work-and-price-leakage-oil-and-gas/"}, {"title": "Oil India Limited / Avaali invoice-to-pay supplier portal", "reason": "Named end customer and go-live described, but the page has no date, preventing in-window validation. It describes process automation and invoice extraction without establishing AI; no quantified operating or economic result is reported.", "url": "https://www.avaali.com/case-studies/oil-powers-digital-supplier-ecosystem"}, {"title": "Henry Resources / Rand Group JIB and revenue imports", "reason": "The named case establishes oil-and-gas financial workflow improvement, but neither the customer quote nor case attributes AI to this implementation, gives a date, or quantifies outcomes. Page's AI/MCP section describes possible current product connectivity, not Henry's deployed usage. Keep as a lead for a potential dated, non-AI customer update only.", "url": "https://www.randgroup.com/apps/oracle-netsuite/erp/netsuite-joint-interest-billing/"}, {"title": "Certas Energy / SS&C Blue Prism invoice processing", "reason": "Named energy distributor, but no publication date or in-window substantive update is shown. It is explicitly intelligent automation/RPA plus OCR, not evidence of an AI model. Vendor process-time claims do not specify independent validation or labor-cost realization.", "url": "https://www.blueprism.com/resources/case-studies/certas-rpa-oil-gas/"}, {"title": "AI-driven SWD pump efficiency case (Neuralix / unnamed operator)", "reason": "Duplicate; already documented in report library as GAPF002 and in work/missing-sept29/operating.json.", "url": "https://jpt.spe.org/case-study-ai-driven-optimization-of-swd-pumping-efficiency-using-kpi-based-time-series-analytics"}, {"title": "ExxonMobil Permian closed-loop gas-lift optimization (SPE 219553)", "reason": "Superseded 1 October 2026: exact-workflow audit supports new profile AI-235, related to AI-022. Previous review: Technical article contains a named gas-lift control case, but no implementation cost. Its exact overlap with the existing ExxonMobil records has not been resolved; retained for reconciliation, not counted as a confirmed duplicate or new profile.", "url": "https://jpt.spe.org/unsung-hero-artificial-lift-boosts-production-without-breaking-the-bank-restricted"}, {"title": "Transocean 2025 10-K disclosure of no AI use in offshore drilling", "reason": "Already represented by AI-161. Retain its explicit operational adoption boundary there; no duplicate.", "url": "https://www.sec.gov/Archives/edgar/data/1451505/000145150526000018/rig-20251231x10k.htm"}, {"title": "Equinor 2025 annual report AI and maintenance portfolio disclosures", "reason": "Portfolio and overlapping initiative evidence; no new specifically attributable AI adoption-cost figure found.", "url": "https://www.sec.gov/Archives/edgar/data/1140625/000114062526000013/exhibit154equinor2025ann.htm"}, {"title": "TotalEnergies 2025 Universal Registration Document digital/industrial AI overview", "reason": "Company-level scale/context, not project adoption economics. Port Arthur control-room pilot already AI-207.", "url": "https://www.sec.gov/Archives/edgar/data/879764/000110465926035876/tot-20251231xex15d1.htm"}, {"title": "Conceptual offshore asset-integrity AI cost-benefit framework", "reason": "Not a named operator deployment or measured pilot; paper says the framework is conceptual.", "url": "https://link.springer.com/article/10.1007/s41872-025-00353-2"}, {"title": "Repsol / Harvey: useful operating detail, disclosure date unresolved", "reason": "The named legal workflow differs from existing Repsol profiles. The source describes a pre-window 2024 rollout and later results, but no exact publication date is shown. Retained for date verification; excluded from this dated-profile increment.", "url": "https://www.harvey.ai/customers/repsol"}], "gaps": ["No explicit named-operator implementation cost or payback found in this focused pass.", "Named finance, JIB, AFE, title and procurement deployments with task-level outcomes remain thinly evidenced. Devon’s legal announcement does not close those gaps.", "No documented named failed AI deployment was established; public disclosures remain selective.", "Repsol’s legal AI case and the Dagang production-forecasting source need date/access verification.", "A complete company-quarter earnings-call, investor-presentation and sponsor-portfolio census remains unfinished."], "search_log": [{"target": "ONGC water flood optimization Heera South Heera AI ML data physics Tachyus annual report 2024-25", "result": "Located the operator's FY2024-25 annual report filed on NSE. Opened the full 709-page primary PDF and reviewed the field-specific passage: pilot in Heera, recommendations in South Heera, subsequent North/Mid Heera use, and routine performance monitoring. Promoted as one distinct operator deployment.", "url": ""}, {"target": "Oil India Limited annual report 2024-25 artificial intelligence ML operations workflow", "result": "Opened the full primary FY2024-25 report. It reports 11 implemented DRIVE digital projects, but describes DRIVE 2.0 and more inclusive AI adoption as future work; the report does not identify a sufficiently specific, deployed AI operating workflow. Excluded.", "url": ""}, {"target": "Petrobras 2025 machine learning well configurations predictive monitoring integrated operations control rooms", "result": "Located Petrobras 2026-30 business-plan evidence, but dataset already contains distinct matching Petrobras records for machine-learning well configurations (swarm2-ai-219), AI predictive monitoring / integrated operations control rooms (swarm2-ai-220), revenue forecasting (swarm2-ai-221), and Smart Tocha (AI-146). No additional record promoted.", "url": ""}, {"target": "Dagang Oilfield machine-learning production prediction model 2025 forecast accuracy 15%", "result": "CNPC News Center result describes the model in multiple blocks on trial and reports a 15% increase in forecast accuracy, plus 11 AI applications. The primary page could not be opened due a Unicode decoding error; excluded under the full-source-read rule despite a promising indexed article.", "url": ""}, {"target": "PETRONAS PROTEAN 2025 integrated report deployment 350 equipment items", "result": "The PETRONAS annual-report disclosure of machine-learning failure prediction across Malaysia/international assets and more than 350 rotating-equipment items is already the substance of AI-209. Treated as a duplicate, not a new addition.", "url": ""}, {"target": "oil gas operator AI land administration title abstracts customer deployment 2024 2025", "result": "Returned generic land AI vendors and a BIS/Grooper Permian acquisition-data case using an anonymized operator; no qualifying named, in-window customer disclosure found.", "url": ""}, {"target": "oil gas operator AI invoice processing procurement customer case study 2025", "result": "Returned Chesapeake/Enverus AP-field-ticket automation, anonymized Calgary exploration-company AP case, utilities/energy AP examples, and named legacy automation. Chesapeake implementation is stated as 2022 and source is undated; other examples were anonymous, outside category, or did not substantiate AI deployment.", "url": ""}, {"target": "AI revenue accounting JIB automation operator customer 2024 2025", "result": "Returned JIB software product pages and Henry Resources/Rand Group accounting integration. Named Henry case demonstrates JIB/revenue imports and ERP process improvements but does not evidence deployed AI; page's current AI assistant discussion is a product capability, not customer use.", "url": ""}, {"target": "AI land administration oil gas operator deployed customer lease 2025", "result": "Returned KPMG named-workflow details but anonymized North American oil and gas client and no publication date; excluded for customer identity and period verification.", "url": ""}, {"target": "AI oil gas accounts payable named 2025 invoice automation; JIB customer implementation; title customer implementation", "result": "Returned supplier invoice automation (OIL India), Certas intelligent automation, and older/non-AI process automation. OIL India source is undated and describes automation/OCR, not clearly AI; Certas source does not establish disclosure date and describes OCR plus RPA. Both withheld.", "url": ""}, {"target": "site:annualreports.com oil gas 2025 artificial intelligence implementation cost operational project", "result": "Reviewed Equinor 2025 annual report: portfolio-level ML and business-value language, plus named maintenance/efficiency examples, but no project adoption cost or new case suitable for this focused request. Existing profile coverage overlaps.", "url": ""}, {"target": "site:sec.gov/Archives oil gas 2025 artificial intelligence pilot cost operations project", "result": "Reviewed Transocean 2025 10-K result: it explicitly says AI is not currently used in offshore drilling operations and describes administrative AI plus governance. Already represented as AI-161; no new profile. No project-level operating AI adoption cost found.", "url": ""}, {"target": "named operator annual report AI operational control boundaries human in the loop 2025", "result": "Reviewed TotalEnergies 2025 Universal Registration Document: broad industrial AI scale and Digital Factory/infrastructure description; no project-level adoption cost or new control-boundary case surfaced. Specific Port Arthur pilot already AI-207.", "url": ""}, {"target": "SPE 219528 autonomous ESP Vital Energy SPE paper human oversight deployment cost", "result": "JPT/SPE article dated 2024-10-01 and SWPSC case abstract reviewed. Found named Vital Energy human-approval then nine-well autonomous FSP transition; LOE qualitatively unaffected but no project implementation cost or payback disclosed. Relevant as an in-window evidence enrichment to pre-existing undated supplemental Vital workflow.", "url": ""}, {"target": "SPE 219553 gas lift ExxonMobil machine learning adoption cost human approval", "result": "JPT article describes 1,300+ Permian wells, three-month human approval stage (1.87% average uplift) and subsequent three-month stage with most changes automatically implemented (about 2.2%). No project-level adoption cost. ExxonMobil has existing AI initiatives; no new profile or cost record added.", "url": ""}, {"target": "SPE papers oil gas AI explicit deployment costs adoption costs 2024 2025", "result": "Focused searches surfaced field-performance papers but no defensible named-operator AI case with explicit license, integration, implementation cost, or payback. Generic deployment ROI and non-AI lift comparisons excluded.", "url": ""}, {"target": "Enbridge midstream power and right-of-way workflows", "result": "Original 7 October 2024 company release read and independently audited against the full register. Two specific workflows differ from the existing in-line deformation-feature record.", "url": "https://www.enbridge.com/media-center/news/details?id=123831&lang=en"}, {"target": "Williams companywide AI adoption", "result": "Operator page reports >90% employee use; exact publication date comes from URL path. Kept distinct from Sequent gas marketing and gas-to-power infrastructure; no productivity return inferred.", "url": "https://www.williams.com/2026/06/01/williams-tops-90-ai-adoption-as-approved-tools-become-part-of-everyday-work/"}, {"target": "ADNOC real-time drilling operations", "result": "Original 4 August 2026 release read by lead and regional researcher; ADNOC deployment is distinct from ONGC DrillOps, ENERGYai and AiPSO.", "url": "https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance/"}, {"target": "Harvey / Icertis / Sirion / SAP legal and procurement search", "result": "Devon’s August 2025 selection announcement qualifies at announcement stage only. Repsol remains undated; SAP SATORP procurement material did not establish AI.", "url": "https://www.harvey.ai/blog/devon-energy-selects-harveys-legal-ai-platform"}], "added": ["AI-222", "AI-223", "AI-224", "AI-225", "AI-226", "AI-227"], "enriched": []}, "swarm4": {"date": "2026-10-01", "summary": "The missing detail was concentrated in specific operating workflows, back-office processes and the evidence behind value claims. This pass adds refinery prediction and blending, well surveillance, gas-lift control, document-based tax recovery and production forecasting. The remaining priorities are a complete document census, named accounting and land deployments, and measured returns after costs.", "findings": [{"title": "The company register still misses consequential operators.", "text": "Moeve and Presidio add operating context that was absent. Moeve names refinery applications and has documented Carlyle-controlled ownership; Presidio distinguishes live surveillance from broader automation targets.", "ids": ["AI-228", "AI-229", "AI-231"]}, {"title": "Back-office AI has concrete workflows and different kinds of value.", "text": "SLB connects document extraction to airline tax processing and export-duty recovery, and uses AI to match internal projects. The $10 million duty-recovery claim has no stated period or implementation cost; the invoice bot’s one-FTE equivalent is time capacity, not a disclosed staff reduction.", "ids": ["AI-232", "AI-233", "AI-234"]}, {"title": "Detailed operating studies improve the report more than another AI partnership.", "text": "ExxonMobil’s specific gas-lift case was incorrectly excluded by a broad company-level match in the prior pass. Dagang’s original Chinese disclosure is now readable and supports a separate forecasting trial. The register preserves historical results and unspecified measurement bases.", "ids": ["AI-235", "AI-236"]}], "priorities": [{"gap": "1. Coverage by company and reporting period", "status": "Moeve and Presidio were absent from the company register. This pass adds specific workflows; it does not complete a global census.", "next": "A company-by-quarter checklist of earnings calls, presentations and annual reports, recording searched documents and absent disclosures. Harbour, Cheniere, INPEX and Galp remain priority gaps."}, {"gap": "2. Accounting, land and legal operations", "status": "SLB supplies three concrete internal workflows. First E&P is a named operator lead with an inaccessible original source. Only two baseline profiles were classified as land/legal.", "next": "Named deployments for JIB, AFE controls, revenue accounting, ownership/title and lease obligations; original case text, dates, human review and customer confirmation."}, {"gap": "3. Net financial returns", "status": "SLB reports export-duty recovery; lift cases report production changes. Neither establishes fully costed, incremental net ROI.", "next": "Implementation and recurring cost, benefit period, counterfactual, failed cases and reconciliation to cash or operating expense. Separate capacity released from headcount removed."}, {"gap": "4. Private-equity causation", "status": "Moeve’s report documents Carlyle-controlled ownership. Presidio has historical sponsor backing. Those facts do not show sponsor-led AI execution.", "next": "Sponsor or portfolio-company evidence naming the sponsor’s operating role, spend, rollout mandate and attributable portfolio-level results."}, {"gap": "5. Deployment identity and control boundaries", "status": "The new technical cases expose differences between advisory tools, automated setpoints and full-field claims. Some platform-to-field crosswalks remain unresolved.", "next": "Which wells are active, which decisions are automatic, operator overrides, fallback rules and explicit links between named products and field studies."}], "supplemental": [{"company": "First E&P", "title": "Supplier-invoice AI portal", "finding": "Wragby’s indexed case text names a Nigerian operator, AI Builder extraction, approvals and ERP integration. It reports a 65% turnaround reduction and 500+ processed invoices.", "limit": "The legacy original site does not resolve. Only indexed text was accessible; vendor figures and July 2025 date remain provisional. Not added to the verified initiative register.", "url": "https://wragbyrevamped.azurewebsites.net/?p=14689"}, {"company": "Repsol", "title": "Harvey legal workflow", "finding": "Harvey describes a January 2024 pilot, April 2024 rollout and later legal-team adoption. This is a useful lead for contract and legal work.", "limit": "Updated in fifth pass: dated 30 March 2026 interview supports first register profile AI-237. Stronger metrics on this undated page remain undated. Previous limit: The page is undated; a phrase such as “two years later” is insufficient to invent a publication date. Keep separate from dated deployments pending chronology verification.", "url": "https://www.harvey.ai/customers/repsol"}, {"company": "ADNOC Onshore / AIQ", "title": "Autonomous gas-lift control — SPE 230130", "finding": "Indexed abstract text describes centralized process control and cloud analytics in a field with more than 200 gas-lift wells.", "limit": "Direct OnePetro access returned 403 and the JPT highlight was unavailable. Indexed details are not equivalent to a directly reviewed source; held outside the initiative register. Possible overlap with existing ADNOC platforms remains unresolved.", "url": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/D021S043R003/793441"}], "excluded": [{"title": "Harbour Energy: governance is not an operating deployment", "reason": "FY2025 annual report p.76 discusses AI risks and controls. The separately surfaced 2026 downhole telemetry paper could not be read; no Harbour profile added.", "url": "https://www.harbourenergy.com/media/a11hxbdn/harbour-energy-annual-report-accounts-2025_web.pdf"}, {"title": "Moeve: adjacent savings and digital projects", "reason": "The report’s 27,000-tonne CO2 reduction belongs to energy-recovery projects. Finance4All, CLM and terminal management are not individually identified as AI; none is counted as an AI result.", "url": "https://www.moeveglobal.com/stfls/corporativo/FICHEROS/consolidated-management-report2025.pdf"}, {"title": "Moeve PredictIA: undated development page", "reason": "Concrete research on blending and hydrocracking, but no exact disclosure date or production deployment shown; not separately counted against potentially overlapping refinery work.", "url": "https://www.moeveglobal.com/en/innovation/innovation-spaces/innovation-center/predictia-project"}, {"title": "Unresolved lead", "reason": "Official article search result previously reported multi-block model trials and 15% higher forecast accuracy, but direct opening still fails with Unicode decoding error. Search result text alone does not satisfy the user's source-review rule.", "url": ""}, {"title": "Unresolved lead", "reason": "Already in baseline as AI-201 and AI-202 with the same April 1, 2026 JPT primary article, named workflows, and quantified evidence; not a new addition or update.", "url": ""}, {"title": "Unresolved lead", "reason": "Already in baseline as AI-168 with the same Aug. 5, 2025 case study and current reported outcome fields; not a new in-window disclosure.", "url": ""}, {"title": "Unresolved lead", "reason": "The vendor case study anonymizes the operator, preventing named-company attribution and meaningful deduplication.", "url": ""}, {"title": "Unresolved lead", "reason": "Vendor result snippet describes a data-platform case and faster analyst-built analytics, but the full primary customer story returned a cache miss and no sufficiently concrete production-operation decision workflow was verified.", "url": ""}, {"title": "Unresolved lead", "reason": "Strong production and revenue metrics, but the operator is anonymized and the case describes lift-service / digital-platform actions without establishing the stated results as an AI deployment. Excluded from this AI-operator cohort.", "url": ""}, {"title": "Cheniere Energy / Arkestro procurement", "reason": "The Jan. 24, 2025 Emerj interview is explicitly Arkestro-sponsored. Full interview read: it pairs Cheniere's Supply Chain Center of Excellence lead Faraz Shahid and Arkestro CEO Neil Lustig for general discussion of digital skills, change management, and multi-technology procurement. It gives no Cheniere workflow deployment, Arkestro selection, customer quote about using Arkestro, or performance result. Arkestro primary press/news search returned no Cheniere customer announcement. Do not imply a supplier relationship from co-interview alone.", "url": "https://emerj.com/driving-procurement-efficiencies-for-oil-gas-faraz-shahid-cheniere-energy-neil-lustig-arkestro/"}, {"title": "Reliance Oilfield Services / Stampli", "reason": "A secondary case-study aggregator says AI/ML extraction helped AP scale to 400+ invoices/month without added staff, but accessible page links to the underlying case and no dated in-window disclosure is established. Not promoted without direct full primary case and date verification.", "url": ""}, {"title": "Anonymous midstream/utility/Calgary AP examples", "reason": "USEReady, UE Labs and Intelbyte sources describe useful invoice workflows/metrics but customer identity is withheld, or customer is not an oil/gas end user. They do not meet named-customer requirement.", "url": ""}, {"title": "KPMG lease AI, BIS Grooper land extraction, Kolena land/JIB examples", "reason": "Relevant lease/JIB text automation is described, but the customer is anonymous or identified only by broad role/basin; publication/update date or named operator is absent. Prior swarm3 business file documents route details; not repeated as candidates.", "url": ""}, {"title": "Generic Coupa/SAP/Icertis/Arkestro procurement and JIB features", "reason": "Product or industry material is not evidence of a named oil/gas customer's production AI workflow. Named SATORP, INPEX, and Henry Resources materials are conventional procurement/accounting digitization or AI product availability, not a qualifying dated AI deployment in the target workflow.", "url": ""}, {"title": "Devon Energy AI-006 closed-loop gas lift", "reason": "Similar technical category, different operator, asset population and source. It cannot be used as a dedup reason for ExxonMobil SPE 219553.", "url": ""}, {"title": "ONGC AI-152 AI/ML gas-lift tender", "reason": "Different operator and procurement stage; an RFP is not ExxonMobil's operating deployment.", "url": ""}, {"title": "SM Energy AI-212 machine-learning gas-lift optimization", "reason": "Different operator. Do not collapse distinct company deployments based on the same lift-control workflow family.", "url": ""}, {"title": "ADNOC RoboWell offshore NASR case in 2024 Sustainability Report", "reason": "Retained as related supplemental context only, not a third new profile. The report identifies offshore NASR and the SPE case identifies a mature onshore field; no source links them to the same system or deployment.", "url": ""}, {"title": "Vital Energy SPE 219528 autonomous ESP", "reason": "Out of scope for this pass because parent explicitly excluded the Vital supplemental case; it is a different operator and artificial-lift workflow.", "url": ""}, {"title": "Neuralix GAPF002 SWD pumping efficiency", "reason": "Explicitly excluded per instruction; known duplicate and unrelated to the candidates retained here.", "url": ""}, {"title": "AI-201, AI-202, AI-141 cases", "reason": "Explicitly excluded per instruction; no duplicate case reused.", "url": ""}], "gaps": ["A company-by-quarter checklist of earnings calls, presentations and annual reports, recording searched documents and absent disclosures. Harbour, Cheniere, INPEX and Galp remain priority gaps.", "Named deployments for JIB, AFE controls, revenue accounting, ownership/title and lease obligations; original case text, dates, human review and customer confirmation.", "Implementation and recurring cost, benefit period, counterfactual, failed cases and reconciliation to cash or operating expense. Separate capacity released from headcount removed.", "Sponsor or portfolio-company evidence naming the sponsor’s operating role, spend, rollout mandate and attributable portfolio-level results.", "Which wells are active, which decisions are automatic, operator overrides, fallback rules and explicit links between named products and field studies."], "method": "Three Luna researchers worked in parallel on business functions, regional operators and technical/economic evidence. Findings were checked against all 227 baseline profiles and attached research packets, then reviewed before integration. Dates are disclosure or update dates unless stated otherwise. Source review is not independent validation of company outcomes. This targeted pass does not establish exhaustive company or quarter coverage.", "search_log": [{"target": "Absent operators: Cheniere, Venture Global, Harbour", "result": "Cheniere procurement interview referred to business researcher. Harbour annual-report AI mentions center on risk/governance; a 2026 technical paper remains inaccessible. Venture Global search returned AI-demand framing and unverified third-party operational claims, not a verified deployment.", "queries": []}, {"target": "Moeve/Cepsa, Galp and INPEX company-source search", "result": "Moeve absent across profiles and packets. Original release and FY2025 report establish three practical initiatives/announcements. Galp/INPEX prior coverage already records timing gaps; no fresh confirmed additions in this root search.", "queries": []}, {"target": "https://news.cnpc.com.cn/system/2026/02/04/030185200.shtml", "result": "Retried opening the public CNPC News Center primary page directly. It still fails with a Unicode decoding error. No alternate access route or access-control workaround attempted. Excluded under the full-source-read requirement.", "queries": []}, {"target": "Canadian private oil producer AI machine learning operations optimization case study operator 2025 Alberta", "result": "Returned vendor/platform material and known public-operator cases, but no new named private Canadian operator deployment verified. Whitecap/Shakudo surfaced as a new public-operator infrastructure story; full-page fetch returned cache miss and the source does not establish a sufficiently specific operational workflow. Excluded.", "queries": []}, {"target": "private equity backed oil gas operator artificial intelligence production optimization 2025 operator deployment", "result": "Returned private-equity-backed but anonymized Permian case-study materials, already disclosed operator examples, and named Presidio production-intelligence material. Presidio SEC source opened and reviewed in full; promoted with scope and target caveats.", "queries": []}, {"target": "private operator machine learning production optimization 2025 named operator upstream; JPT Montney production optimization", "result": "The SPE/JPT April 1, 2026 OPX Ai article gives detailed Kaybob and Montney workflows and results, but its exact source URL and those deployments are already recorded as AI-201 and AI-202 in the 227-profile baseline. Not re-added.", "queries": []}, {"target": "PE-backed operator oil and gas AI case study; private equity backed Permian operator ROAM-AI", "result": "A vendor case study for a PE-backed Permian operator does not name the customer. Excluded because the operator cannot be independently identified or deduplicated and the full operator attribution needed for a profile is absent.", "queries": []}, {"target": "Presidio SEC 2026 investor presentation AI deployment FTW Technologies well surveillance", "result": "Opened and read the full SEC-hosted investor presentation, dated June 2026, including its legal disclaimer and slides 15–18. The live surveillance workflow and live/target ambiguity are recorded in the candidate and limits. Deep-deduped against baseline, prior swarm, and report-library files; no Presidio/FTW record found.", "queries": []}, {"target": "Named operator AI invoice/AP and finance workflows", "result": "Found Wragby/First E&P named AI invoice portal; result page indexed the full case text and date. Other returns were anonymous or undated KPMG/USEReady/Intelbyte/AP examples.", "queries": ["oil and gas AI accounts payable named customer case study invoice automation 2025", "oil and gas operator AI finance accounting JIB customer deployment case study named 2024 2025 2026", "oil and gas land title lease AI named operator customer story deployment", "oil energy legal AI contract review customer deployment procurement case study named operator"]}, {"target": "Repsol/Harvey exact disclosure chronology", "result": "No dated secondary announcement surfaced. Original Harvey page remains undated and states January 2024 pilot, April 2024 rollout, followed by 'Two years later'; retain as undated supplemental.", "queries": ["Harvey Repsol 'Two years later' date 96% adoption published 2026", "site:harvey.ai/customers/repsol 2026 Harvey Repsol adoption date"]}, {"target": "Procurement/CLM legal AI vendors Icertis/Sirion/Coupa", "result": "Icertis surfaced Enbridge, with undated contract-management case and only generic early-adopter AI statement; no workflow-specific AI use. Sirion results were non-energy/anonymous; Coupa SATORP/INPEX material did not establish deployed AI.", "queries": ["Icertis oil gas AI contract management named customer oil company case study", "oil gas procurement AI customer case study Coupa named operator AI contract review"]}, {"target": "JIB, revenue accounting and invoice automation", "result": "Found SLB/Microsoft Learn airline invoice/tax processing case (explicit last updated Aug 8, 2025) and First E&P/Wragby. Others were anonymous, outside scope, conventional automation, older, or product capability.", "queries": ["oil gas AI revenue accounting customer deployment 2025 operator", "joint interest billing AI customer oil and gas implemented AI", "named E&P company AI invoice processing accounts payable case study 2024 2025", "oil gas energy customer AI procurement case study invoices contract 2025"]}, {"target": "First E&P source provenance and confirmation", "result": "Only Wragby's case surfaced with named First E&P and exact case date; no separate First E&P source found. Wragby page full indexed text accessible, legacy page URL direct open returns cache miss.", "queries": ["First E&P invoice AI portal Wragby Microsoft Power Platform case study", "site:first-ep.com.ng AI invoice processing finance First E&P", "First E&P AI Builder invoices Wragby July 2025"]}, {"target": "Cheniere/Arkestro primary-source verification", "result": "No Arkestro primary customer announcement, customer case, or named Cheniere deployment source found. Emerj Jan 24 2025 sponsored interview is about sector practices/change/talent; it does not say Cheniere deployed Arkestro. Arkestro press release names energy companies in aggregate, not customers.", "queries": ["site:arkestro.com Cheniere Energy procurement", "Arkestro Cheniere Energy procurement AI announcement Faraz Shahid", "Cheniere Arkestro procurement AI", "Faraz Shahid Arkestro Cheniere procurement"]}, {"target": "Additional energy invoice workflows", "result": "Microsoft Learn source confirms specific SLB AI Builder invoice/tax workflow. First E&P found again, same Wragby story. Cheniere/Arkestro not promoted.", "queries": ["AI invoice automation oil gas operator Microsoft Learn customer story", "First E&P Power Platform AI Builder invoice Wragby"]}, {"target": "ExxonMobil SPE 219553 gas lift exact title author primary paper", "result": "Located complete accessible abstract for SPE 219553-MS and its DOI/authors. Abstract says >1,300 ExxonMobil Permian wells, remote injection setpoint tests, QC, model-tracked uplift and ~2.0% average oil uplift; the page says full text is not available.", "queries": []}, {"target": "site:onepetro.org SPE-219553-MS OR SPE 219553", "result": "Confirmed the work is SPE 219553-MS and the 2025 JPT highlight traces directly to it; OnePetro paper body not accessible in this pass.", "queries": []}, {"target": "SPE 219553 ExxonMobil 1300 wells human approval 1.87 2.2 JPT", "result": "Read the full accessible relevant section of JPT's 2024-10-01 feature. It gives both three-month stages, field-person approval in the first, mostly automatic changes in the second, and reports 1.87% then ~2.2% average uplift.", "queries": []}, {"target": "Gas Lift Optimization in the Permian Uses Machine Learning 2025 JPT", "result": "Read the entire publicly accessible 2025-10-01 JPT highlight: >1,300 operators' wells, remote gas-lift setpoints, automated data acquisition, no major expenditure framing, and approx. 2.0% average uplift. It is an opening highlight rather than the complete SPE paper.", "queries": []}, {"target": "ADNOC AIQ SPE-230130-MS autonomous gas lift 200 wells challenges", "result": "Located and read the complete accessible OnePetro abstract. It names ADNOC Onshore and AIQ authors, >200 gas-lift wells, centralized APC/cloud analytics, failure/fallback, DCS/OT-IT integration, and sparse-data model limits; no numeric outcome or economics disclosed.", "queries": []}, {"target": "SPE 230130 JPT autonomous gas lift ADNOC October 2026", "result": "Opened and read the publicly accessible JPT 2026-10-01 article through its field overview. It confirms the mature onshore Abu Dhabi context (>200 gas-lift producers) and explicitly says the system operates within operational and safety constraints; further text is member-exclusive.", "queries": []}, {"target": "ADNOC 2024 Robowell AI gas lift offshore NASR 30% official report", "result": "Read the full official report page. It reports 2024 offshore NASR RoboWell with up to 30% gas-lift-consumption optimization and planned 500+ well expansion. This could be related to AIQ's gas-lift work, but SPE 230130 is an unnamed onshore field and does not identify RoboWell; no common-system assertion made.", "queries": []}, {"target": "ExxonMobil AI records gas lift 1300 Permian SPE 219553 broad AI-022", "result": "Searched all 227 baseline initiative names and detailed fields: only AI-022 is broad ExxonMobil Permian/operating-site AI; no named closed-loop gas-lift workflow or SPE 219553 was found. Other lift records were Devon AI-006, ONGC tender AI-152, and SM Energy AI-212, different operator/stage.", "queries": []}, {"target": "ADNOC onshore AIQ gas lift AI autonomous control existing baseline AiPSO RoboWell", "result": "Checked all ADNOC baseline initiatives and research packets. AI-011 is ADNOC/SLB AiPSO, AI-040 subsurface ENERGYai; neither establishes identity with AIQ-authored SPE 230130. The RoboWell record is separate offshore NASR context.", "queries": []}, {"target": "SPE papers named operator AI deployment integration failures human control oil gas 2024 2025", "result": "The focused primary-source searches yielded ADNOC Onshore/AIQ SPE 230130 as an additional specific technical-limit case. No separate named primary operator paper with new quantified AI adoption cost or payback was defensible in this bounded pass.", "queries": []}], "corrections": [{"subject": "ExxonMobil SPE 219553", "decision": "Prior broad-family exclusion superseded; accepted as AI-235, related to AI-022."}, {"subject": "Dagang original article", "decision": "Prior access limitation resolved by decoding the original CNPC page; accepted as AI-236."}], "added": ["AI-228", "AI-229", "AI-230", "AI-231", "AI-232", "AI-233", "AI-234", "AI-235", "AI-236"], "enriched": []}, "swarm5": {"date": "2026-10-01", "summary": "The strongest additions sit outside the usual production-optimization story: revenue accounting, legal adoption and worksite safety. Offshore operators and an INPEX CCS model also fill company gaps. This pass adds operating controls, dated adoption evidence and a document-coverage ledger; it still does not establish a complete global census or project-level net returns.", "findings": [{"title": "Revenue accounting now has a named, period-specific case.", "text": "Spearpoint reports processing 82 statements from May through August 2026, with human exception review and export to WolfePak. Its controller’s 20–30 hours a week estimate is staff capacity, not a disclosed payroll saving.", "ids": ["AI-238"]}, {"title": "Adoption and operating safeguards deserve their own evidence.", "text": "Repsol’s March 2026 interview resolves the dated-adoption gap. Saipem’s notices spell out human review and model-improvement conditions. Those details explain how work changes more clearly than broad AI commitments.", "ids": ["AI-237", "AI-241"]}, {"title": "An operating platform is not proof that its new AI is deployed.", "text": "MODEC’s Shape Reef expansion is a 36-month development project. SBM Offshore is scaling a data foundation while broader AI integration remains prospective. INPEX separately documents a CCS model and a deployed Ichthys information-retrieval tool; neither has a disclosed net financial return.", "ids": ["AI-243", "AI-244", "AI-240"]}], "priorities": [{"gap": "Revenue accounting: stronger evidence, unresolved controls", "status": "Spearpoint adds a named, dated revenue-statement workflow with source-page traceability and human exception review.", "next": "Error rates, rejected records, reviewer workload, posting controls and customer-specific fees; named JIB and AFE deployments remain missing."}, {"gap": "Safety: operating controls are now visible", "status": "Saipem notices describe local processing, human verification and retraining triggers. MODEC’s new safety models remain a development project.", "next": "Alert precision, missed events, response time, deployment count and attributable safety impact."}, {"gap": "Undercovered operators: document coverage, not absence claims", "status": "INPEX, Harbour, Cheniere and Galp now have a document-level review ledger. A dated INPEX CCS-model disclosure qualifies.", "next": "Remaining quarterly calls and presentations; original dated evidence for Galp’s specific AI examples and named Harbour/Cheniere deployments."}, {"gap": "Sponsor-led work: evidence does not yet establish returns", "status": "Sentinel Peak’s own page describes the existing steam-management programme, but its current emissions estimate is undated.", "next": "AI project cost, incremental cash benefit, matched operating periods and a causal attribution method."}, {"gap": "Opportunity-map coverage", "status": "Recent profiles omitted from the older opportunity index are now classified. Cases beyond McKinsey’s upstream categories remain visible without an assigned value.", "next": "Maintain one disposition per register profile. Workflow fit must remain separate from deployment stage, evidence quality and realized value."}], "document_coverage": [{"company": "Harbour Energy", "status": "Reviewed; no operational case promoted in this pass.", "documents_checked": [{"title": "Annual Report & Accounts 2025", "date": "2026-03-26", "url": "https://www.harbourenergy.com/media/a11hxbdn/harbour-energy-annual-report-accounts-2025_web.pdf", "review": "Full issuer PDF opened. AI references found in cybersecurity risk/governance discussion on printed p. 76; search of the report for “machine learning” found no match."}, {"title": "Results, reports and presentations index", "date": "Accessed 2026-10-01", "url": "https://www.harbourenergy.com/investors/results-reports-and-presentations/", "review": "Reviewed issuer index to identify 2025 annual report, 2025 results and 2026 operating updates; not every linked presentation/video was separately read."}], "coverage_note": "The reviewed report documents governance awareness of AI risk, not a named operational AI workflow. This is a document-specific finding and does not establish that Harbour has no internal operational AI."}, {"company": "INPEX", "status": "One candidate; prior undated strategy-page gap partly resolved with a dated report disclosure.", "documents_checked": [{"title": "Integrated Report 2024", "date": "2025-07-04", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf", "review": "Full original issuer PDF opened. Printed p. 41 has a named developed ML surrogate for CCS reservoir simulation; printed p. 41 also says a separate integrated monitoring pilot began in spring 2022, which is pre-window. The CCS model’s build/use dates and project deployment are unspecified."}, {"title": "Integrated report index", "date": "Accessed 2026-10-01", "url": "https://www.inpex.com/english/ir/library/integrated_report/", "review": "Issuer index gives 4 July 2025 publication date for 2024 report and 7 July 2026 for 2025 report."}, {"title": "Digital Strategy", "date": "Undated; accessed 2026-10-01", "url": "https://www.inpex.com/english/business/technology/dx/", "review": "Current company strategy page is not dateable to a deployment window; it remains unsuitable on its own for claims about when its examples began."}, {"title": "Integrated Report 2025 — Growth Strategy section", "date": "2026-07-07", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202512_en_03.pdf", "review": "Official index confirms report publication 7 July 2026. Full 21-page original Growth Strategy PDF opened; printed pp. 40–42 reviewed. Identifies developed/deployed Ichthys O&M AI tool, natural-language retrieval from daily reports and maintenance history, approximately 110,000 records, >90% classification accuracy, and retrieval time reduced from hours to seconds/minutes. It does not mention/update the CCS surrogate model."}], "coverage_note": "Dated disclosures now support two distinct INPEX cases: 2025 report release of an already-developed CCS simulation surrogate (no project deployment quantified), and 2026 report release of a deployed Ichthys O&M report/maintenance-history AI tool with reported process metrics. The latter’s deployment date, output review path beyond engineer approval of search quality, validated accuracy methodology and economic benefit remain undisclosed."}, {"company": "Galp", "status": "Dated report and live operator pages reviewed; concrete AI workflows on current innovation pages are not dateable to the requested disclosure window.", "documents_checked": [{"title": "Annual Integrated Report 2025", "date": "2026-03-23", "url": "https://www.galp.com/corp/Portals/0/Recursos/Investidores/IMR2025/AnnualIntegratedReport2025.pdf", "review": "Issuer report/publication index and original report search excerpts reviewed. It describes 2025 renewable-portfolio partnerships for advanced monitoring, predictive maintenance, AI and drones, but does not identify a separate AI workflow/model, deployment date or attributable result. It reports a 5 MW/20 MWh Alcoutim storage pilot operating from early 2025 but does not assign AI to it."}, {"title": "Key Innovation Projects", "date": "Undated; accessed 2026-10-01", "url": "https://www.galp.com/corp/en/innovation/key-innovation-projects", "review": "Full issuer page read. It says AI-based analytics for second-life batteries are implemented at the 20 MWh Alcoutim BESS and enable early fault detection/stable operation, and describes AI soiling analytics tests and transformer predictive-maintenance diagnostics. The page has no source date/update date; 2025 commissioning evidence for the BESS does not date its AI functions."}, {"title": "Galp 103 AI-enabled drone inspection pilot", "date": "Undated; accessed 2026-10-01", "url": "https://www.galp.com/pt/upcomingenergies/en/pilots/upstream-galp-103-ai-enabled-autonomous-drone-inspections-offshore-platforms", "review": "Full issuer innovation page read. It describes AI/RGB/thermal anomaly detection and onshore field testing, with offshore campaigns planned; publication date and in-window start are not established."}, {"title": "Galp and Powin install a large-scale Alcoutim battery system", "date": "2025-04-11", "url": "https://www.galp.com/corp/pt/media/comunicados-de-imprensa/comunicado/id/1602/galp-e-powin-instalam-sistema-de-baterias-de-grande-escala-em-parque-solar-de-alcoutim", "review": "Full operator release read. Confirms 5 MW/20 MWh storage system began injecting power and was commissioned; it describes storage, grid dispatch and stability, not AI."}], "coverage_note": "The undated current pages provide stronger technical detail than the dateable report, but dateable claims are deliberately not combined into a dated AI deployment candidate."}, {"company": "Cheniere Energy", "status": "Reviewed key primary filings; no operational AI case promoted in this pass.", "documents_checked": [{"title": "Cheniere Energy 2025 Form 10-K", "date": "2026-02-26", "url": "https://www.sec.gov/Archives/edgar/data/3570/000000357026000005/lng-20251231.htm", "review": "Full SEC HTML filing opened. Searches for “artificial intelligence,” “machine learning,” “predictive maintenance,” “digital twin,” and “analytics” returned no matches."}, {"title": "Investor presentations index", "date": "Accessed 2026-10-01", "url": "https://lngir.cheniere.com/presentations", "review": "Reviewed index of 2024–2026 earnings and corporate presentations to target primary disclosure documents. Presentations were not individually promoted as evidence absent a read source with a named AI workflow."}, {"title": "November 2024 Form 8-K / Code of Business Conduct amendment", "date": "2024-11-13", "url": "https://www.sec.gov/Archives/edgar/data/3570/000000357024000127/lng-20241113.htm", "review": "SEC filing read; mentions AI in conduct and protection-of-information policies, not as an operating workflow."}], "coverage_note": "These documents do not disclose a named operational AI use case. This is a gap in the reviewed material, not a finding that Cheniere has no AI use."}], "supplemental": [{"company": "First E&P", "title": "Supplier invoice portal — source remains inaccessible", "finding": "The previously indexed Wragby case remains a named finance lead.", "limit": "A fresh direct-source check failed. Existing indexed metrics remain outside the verified register.", "url": "https://wragbyrevamped.azurewebsites.net/?p=14689"}, {"company": "Galp", "title": "Battery analytics and offshore drone inspection", "finding": "Current issuer innovation pages describe specific AI applications, including battery analytics and an inspection pilot.", "limit": "The pages are undated. A dated battery commissioning release does not establish when the AI capability was used.", "url": "https://www.galp.com/corp/en/innovation/key-innovation-projects"}], "excluded": [{"title": "Unresolved lead", "reason": "Strong named Nigeria operator lead with claims of 65% turnaround reduction, 500+ invoices and >70% fewer vendor inquiries, but original page at legacy Wragby Azure domain is inaccessible through direct open. Search-result text is not substituted for full source read. Existing baseline and current-site research both flag this lead as inaccessible.", "url": ""}, {"title": "Unresolved lead", "reason": "Named workflow and 75% evaluation-time reduction, but page is approximately 2023 (outside 24-month window) and baseline contains no exact entry. Not a qualifying recent disclosure.", "url": ""}, {"title": "Unresolved lead", "reason": "Named customer deployment of lease, title-opinion and division-order search; the customer story describes standard land software and supplies no evidence AI is used in the workflow. Excluded from AI case set.", "url": ""}, {"title": "Unresolved lead", "reason": "Partnership is named but end customer is anonymous; no named operator deployment or measured operational results independently attributable to a customer.", "url": ""}, {"title": "Unresolved lead", "reason": "Dated vendor product announcement in scope but no named customer, no customer-side adoption evidence. Product lead only.", "url": ""}, {"title": "Unresolved lead", "reason": "2025 launch announcement contains GulfMark quote about its prior land-owner documentation burden but does not establish that GulfMark deployed the new AI product.", "url": ""}, {"title": "Galp AI analysis for second-life battery BESS at Alcoutim", "reason": "Full operator page states the AI platforms are implemented at the Alcoutim 20 MWh BESS and support early fault detection/stable operation. However, the page has no publication/update date; the dated 11 April 2025 commissioning release does not mention AI. In-window AI disclosure timing is therefore unverified.", "url": "https://www.galp.com/corp/en/innovation/key-innovation-projects"}, {"title": "Galp 103 AI-enabled autonomous offshore drone inspections", "reason": "Full operator page describes PixForce partnership, RGB/thermal inspection, AI anomaly detection, tested onshore at Porto de Açu and planned offshore campaigns. Page is undated and does not establish a launch/disclosure date within the requested window; the offshore campaigns are planned, not established as routine operations.", "url": "https://www.galp.com/pt/upcomingenergies/en/pilots/upstream-galp-103-ai-enabled-autonomous-drone-inspections-offshore-platforms"}, {"title": "Galp 2025 renewables AI partnerships", "reason": "Dated 2025 report describes partnerships for advanced monitoring, predictive maintenance, AI and drones in renewable generation and loss reduction but does not name the systems/workflows, state scope or isolate results; current website details are undated.", "url": "https://www.galp.com/corp/Portals/0/Recursos/Investidores/IMR2025/AnnualIntegratedReport2025.pdf"}, {"title": "INPEX integrated monitoring pilot", "reason": "Report says integrated operation-centre/smart-facility pilot launched in spring 2022 (printed p. 41); implementation predates the review window. The July 2025 report date does not make the underlying pilot new.", "url": "https://www.inpex.com/assets/documents/english/ir/library/integrated_report/inpex_integratedreport202412_en.pdf"}, {"title": "Oil India DRIVE 2.0 Digital Oil Field and Command & Control Centre", "reason": "Full current page describes a proposed program covering 380 producing wells and proposed AI/ML well performance, integrity and maintenance applications. The page presents initiatives as plans under DRIVE 2.0 and does not establish a named deployed workflow or dated in-window go-live.", "url": "https://www.oil-india.com/index.php/digitalfootprint"}, {"title": "Harbour Energy AI mentions", "reason": "The full 2025 report mentions AI as a cyber and information-security risk/control issue, not as an operational workflow; no candidate counted.", "url": "https://www.harbourenergy.com/media/a11hxbdn/harbour-energy-annual-report-accounts-2025_web.pdf"}, {"title": "Cheniere employee-policy AI disclosure", "reason": "The 13 November 2024 Form 8-K discusses artificial intelligence only in the Code of Business Conduct and Ethics changes on information protection; it is not an operational AI use case.", "url": "https://www.sec.gov/Archives/edgar/data/3570/000000357024000127/lng-20241113.htm"}, {"title": "Previously captured Petrobras planning cases", "reason": "The well-configuration ML, control-room predictive monitoring, flare control and revenue forecasting cases surfaced in the November 2025 investor plan are already represented in the baseline; no new workflow added.", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0"}, {"title": "Carlyle 2024 Annual Report: anonymous GenAI portfolio savings", "reason": "Reports >$10m annual savings at an unnamed North American portfolio company and a separate AI-augmented customer-service result at an unnamed industrials company. Neither is identified as oil/gas nor tied to a named operating workflow. The $50m summit figure is potential portfolio-wide bottom-line impact, not realized case-level savings.", "url": ""}, {"title": "Carlyle / KKR sponsor AI strategy, partner programs, and AI funds/investment totals", "reason": "Sponsor capability and allocation do not establish a named end-user operating deployment or AI project spend/return.", "url": ""}, {"title": "Quantum Capital Group AI at Work / Quantum IQ / production-surveillance platform", "reason": "AI-052–AI-055 already cover these sponsor tools at portfolio level. Public descriptions name no additional operator/outcome; Sentinel Peak's workflow is already separately represented as AI-214.", "url": ""}, {"title": "Presidio / FTW Technologies investor presentation", "reason": "Exact profile AI-231 already captures June 2026 primary presentation, including live AI surveillance, ~1% YTD toward 3–5% target, illustrative $2.5MM per 1% annual value, no cost/payback, and historical Morgan Stanley Energy Partners backing. It is not a new sponsor-directed workflow from this pass.", "url": ""}, {"title": "Apollo acquisition of Bold Production Services", "reason": "Official acquisition announcement names gas-treatment equipment/services and efficiency goals but does not disclose AI deployment or AI economics.", "url": ""}, {"title": "AI-enabled data-center power supply, generation and infrastructure investments", "reason": "These finance power/compute for AI demand, rather than applying AI to oil/gas end-user workflows.", "url": ""}, {"title": "Williams Maximo MAS9 upgrade", "reason": "A software upgrade carrying AI features does not prove that Williams activated a particular AI maintenance workflow.", "url": "https://www.williams.com/wp-content/uploads/sites/14/2025/07/Williams_2024SustainabilityReport_Updated.pdf"}, {"title": "Saipem REFLEX and D2NCR", "reason": "Relevant indexed annual-report text, but original full annual PDF was unavailable to this pass; retain as leads, not directly reviewed profiles.", "url": "https://www.saipem.com/sites/default/files/2026-04/Annual_Report_2025.pdf"}, {"title": "Saipem LiSa harness as a separate AI profile", "reason": "Sensors and algorithms are described, but explicit model-based AI for the harness was not established independently of the broader safety-technology wording.", "url": "https://www.saipem.com/sites/default/files/2026-03/Saipem_Sustainability_Overview_2025.pdf"}], "gaps": ["Error rates, rejected records, reviewer workload, posting controls and customer-specific fees; named JIB and AFE deployments remain missing.", "Alert precision, missed events, response time, deployment count and attributable safety impact.", "Remaining quarterly calls and presentations; original dated evidence for Galp’s specific AI examples and named Harbour/Cheniere deployments.", "AI project cost, incremental cash benefit, matched operating periods and a causal attribution method.", "Maintain one disposition per register profile. Workflow fit must remain separate from deployment stage, evidence quality and realized value."], "method": "Three Luna researchers covered business functions, undercovered operators, and private-equity economics. Primary documents and named vendor cases were checked against 236 profiles and their attached packets. A separate source audit checked new offshore and safety cases. Original-source access, date precision, practical use and performance attribution were reviewed separately. Existing records are preserved; undated corroboration is labeled as such.", "search_log": [{"target": "site:wragby.com First E&P AI land accounting deployment; First E&P Wragby AI source legacy site; Harvey Repsol legal AI date", "result": "Confirmed Repsol’s official 2024 announcement is dated 2024-04-04, outside the window, and found the undated Harvey customer page already retained as a supplemental in current-site research. Full-read a separate Harvey interview dated 2026-03-30 and used it as a dated status update to that existing supplemental, not a new workflow. First E&P search returned Wragby legacy-domain page as a lead, but direct source open failed; it remains an existing unresolved research lead.", "queries": []}, {"target": "oil gas JIB AI customer case study operator; oil and gas revenue accounting AI operator case study", "result": "Found Joltly/Spearpoint dated customer case, Enverus/Raisa older case, and generic or illustrative material. Opened Joltly full page; baseline-deduped Spearpoint. Raisa result is old (published ~2023) and not baseline candidate.", "queries": []}, {"target": "Repsol Harvey AI legal pilot 2024; Continental Resources UiPath owner relations agent; named operator owner relations AI", "result": "Opened Harvey Repsol page and full dated Harvey power-user interview; confirmed dated 2026-03-30 rollout update and reconciled it with original dated 2024 issuer announcement. UiPath Continental landing page opened/read but lacks date/outcome/visible operator speaker details.", "queries": []}, {"target": "named oil gas operator AI land administration division orders 2025; title workflows", "result": "PakEnergy Impact Energy land-management story describes classical records/owner-relations tooling but no AI in the customer workflow. EAG nuEra and other title products are vendor launches without named operating customer; excluded as product-only.", "queries": []}, {"target": "First E&P AI-powered Invoice Processing Portal Wragby exact page and current domain", "result": "Search indexed full case text and date July 2, 2025, but its only identified source is wragbyrevamped.azurewebsites.net/?p=14689; direct open failed twice (including canonical guess). Since the actual page could not be opened/read, do not promote under full-source rule; preserve as priority missing evidence.", "queries": []}, {"target": "PakEnergy AI land operator; Firm App oil gas named end customer; oil and gas title AI customer", "result": "PakEnergy lease-provision AI remains a dated product announcement without customer deployment. Firm App/Prosper evidence is a consultancy/platform partnership and anonymous operator mention, not a named operator deployment. Enverus courthouse product mentions GulfMark quote about older workflow and no named AI implementation; do not promote.", "queries": []}, {"target": "['site:harbourenergy.com AI machine learning operational production 2025 Harbour Energy annual report', 'site:inpex.com integrated report AI machine learning production operations 2025 INPEX', 'site:galp.com annual integrated report 2025 artificial intelligence operations refinery upstream', 'site:cheniere.com AI machine learning operational LNG 2025 Cheniere']", "result": "Identified Harbour 2025 report index/PDF, INPEX report and Galp 2025 report. Cheniere results initially surfaced broad/non-operator material; continued to primary sources.", "queries": ["site:harbourenergy.com AI machine learning operational production 2025 Harbour Energy annual report", "site:inpex.com integrated report AI machine learning production operations 2025 INPEX", "site:galp.com annual integrated report 2025 artificial intelligence operations refinery upstream", "site:cheniere.com AI machine learning operational LNG 2025 Cheniere"]}, {"target": "['INPEX Integrated Report 2025 machine learning', 'Galp Annual Integrated Report 2025 AI workflows']", "result": "Reviewed INPEX Integrated Report 2024 report case (published July 2025) and report index. Galp AI results included general AI claims and current innovation pages; date/workflow substantiation remained under review.", "queries": ["INPEX Integrated Report 2025 machine learning", "Galp Annual Integrated Report 2025 AI workflows"]}, {"target": "['Galp Alcoutim 20 MWh BESS AI second-life batteries official']", "result": "Found official Galp page with explicit AI second-life-battery analysis at operating Alcoutim BESS, but that page is undated. A dated April 2025 Galp release confirms battery commissioning only, not AI use.", "queries": ["Galp Alcoutim 20 MWh BESS AI second-life batteries official"]}, {"target": "['Cheniere LNG AI operations', 'Africa Latin America Asia independent oil and gas AI operational workflow']", "result": "Search was dominated by vendor-market material and previously captured ADNOC/PETRONAS/Petrobras workflows. No new qualifying independent-operator source surfaced.", "queries": ["Cheniere LNG AI operations", "Africa Latin America Asia independent oil and gas AI operational workflow"]}, {"target": "['site:galp.com AI-based analysis second-life batteries Alcoutim', 'Galp early fault detection BESS AI', 'Galp AI BESS second-life batteries']", "result": "Confirmed Galp’s current page language about AI applied to second-life batteries, early-fault detection, solar AI soiling analytics, and transformer diagnostics; dateability remains insufficient for a dated AI candidate.", "queries": ["site:galp.com AI-based analysis second-life batteries Alcoutim", "Galp early fault detection BESS AI", "Galp AI BESS second-life batteries"]}, {"target": "['Ecopetrol AI production operations 2025 operator reports', 'Petrobras 2025 machine-learning operating cases', 'Sonangol AI oilfield 2025', 'Eni AI operational workflow 2025']", "result": "Ecopetrol 2025 results and Petrobras 2026–30 plan did not yield a new workflow distinct from the baseline; Eni results were broad company-level material. The specific Petrobras planning cases found are already in existing profile coverage.", "queries": ["Ecopetrol AI production operations 2025 operator reports", "Petrobras 2025 machine-learning operating cases", "Sonangol AI oilfield 2025", "Eni AI operational workflow 2025"]}, {"target": "['Cheniere 2025 Form 10-K AI machine learning operations', 'Cheniere investor presentation 2025 AI LNG']", "result": "Read the 2025 SEC Form 10-K full HTML and searched five operational-AI terms. Found no matches. Cheniere’s Nov 2024 AI mention is employee conduct policy, not operational deployment. Investor presentation index was checked; no presentation promoted without a primary full-text workflow.", "queries": ["Cheniere 2025 Form 10-K AI machine learning operations", "Cheniere investor presentation 2025 AI LNG"]}, {"target": "['Galp Annual Integrated Report 2025 AI predictive maintenance', 'Galp Alcoutim fault detection AI official report']", "result": "Report search supports broad 2025 partnerships around advanced monitoring, predictive maintenance, AI and drones, but not a separable workflow. Report independently confirms Alcoutim storage commissioning, not AI integration.", "queries": ["Galp Annual Integrated Report 2025 AI predictive maintenance", "Galp Alcoutim fault detection AI official report"]}, {"target": "['Seplat Energy AI machine learning field operations 2025 annual report', 'Petrobras machine learning 2025 operational case', 'Oil India AI 2025 field operations', 'PETRONAS 2025 AI named deployments']", "result": "Seplat Integrated Report 2025 was reviewed via issuer publication result excerpts with no named AI workflow surfaced. Oil India’s current DRIVE 2.0 page details proposed digital-oilfield scope, but it is a forward program, not evidence of a deployed workflow. Petrobras and PETRONAS results substantially overlap existing dataset coverage.", "queries": ["Seplat Energy AI machine learning field operations 2025 annual report", "Petrobras machine learning 2025 operational case", "Oil India AI 2025 field operations", "PETRONAS 2025 AI named deployments"]}, {"target": "['INPEX Integrated Report 2025 digital strategy', 'INPEX Integrated Report 2025 machine learning CCS', 'INPEX Integrated Report 2025 AI Ichthys maintenance']", "result": "Opened the official 2025 Integrated Report Growth Strategy chapter and read its digital-strategy pages (printed pp. 40–42). Found a distinct deployed Ichthys O&M AI tool; the report does not discuss/update the CCS surrogate. Added that as a second candidate and corrected the 2024 report physical PDF page locator.", "queries": ["INPEX Integrated Report 2025 digital strategy", "INPEX Integrated Report 2025 machine learning CCS", "INPEX Integrated Report 2025 AI Ichthys maintenance"]}, {"target": "site:carlyle.com portfolio company oil gas artificial intelligence deployment operations AI upstream; site:kkr.com portfolio oil gas AI operations; Apollo portfolio energy AI; Quantum AI production optimization", "result": "Carlyle 2024 Annual Report surfaced an unnamed North American portfolio company's >$10m annual GenAI savings and an unnamed industrial portfolio company's $10m AI-augmented customer-service savings; industry-specific oil/gas identity is not disclosed. Carlyle's $50m potential summit impact is prospective and portfolio-aggregate. KKR's Capstone AI strategy is sponsor capability, not an end-user deployment. Quantum portfolio search points to already-profiled AI-214 and AI-052–AI-055.", "queries": []}, {"target": "Carlyle over $10 million annual savings portfolio GenAI identify company; Carlyle energy portfolio AI operations", "result": "Read Carlyle's indexed annual-report result: the annual-saving company is anonymous and the report elsewhere labels the customer-service case 'industrials'; neither can be defensibly assigned to an oil/gas operator. Excluded as anonymous/non-oilfield.", "queries": []}, {"target": "Sentinel Peak Resources Quantum intelligent oilfield AI steam injection optimization cost payback pilot", "result": "Found official Sentinel Peak Environment page and Quantum 2024 Stakeholder Report's Sentinel Peak case. Report identifies Quantum digital-team co-development, steam as ~30% of operating costs at end-2023, >3MM bbl lower steam injection and >120k MT Scope 1 GHG reduction in 2023, while warning all portfolio-company data were not independently verified. The operator page gives a current estimated 200k MT CO2e/year enabled by intelligent-oilfield technologies.", "queries": []}, {"target": "Crescent Energy AI artificial intelligence KKR field operations production optimization; EnCap portfolio AI production optimization", "result": "Crescent investor and SEC searches returned acquisition/operating playbook and general financial/operational results but no named AI workflow or AI-attributed savings. EnCap's portfolio source describes companies and technology strategies, not a specific AI end-user deployment.", "queries": []}, {"target": "Apollo, Warburg Pincus, NGP sponsor portfolio energy company operational AI deployment oil and gas", "result": "Official results disclosed broad Apollo portfolio AI governance and operating-partner remit, not a named oil/gas operational case. NGP materials include energy-company portfolio but no operator AI return/cost case; unrelated NGP Capital industrial AI vendor investment excluded.", "queries": []}, {"target": "Carlyle, KKR, Quantum, EnCap sponsor portfolio AI oilfield operating cost case", "result": "Only specific sponsor-linked, named upstream workflow found was Sentinel Peak/Quantum (already AI-214). Quantum report full case content was available in search rendering, but the current direct update relies on the fully opened Sentinel Peak HTML page; no new candidate inferred solely from indexed PDF rendering.", "queries": []}, {"target": "Midstream operator and contractor operating AI", "result": "Williams Maximo upgrade exposes product AI capability but not a specific enabled workflow; existing Enbridge/Williams/Kinder cases deduplicated.", "queries": []}, {"target": "Saipem safety and quality source depth", "result": "Original notices and reports establish video safety controls and a separate early drone survey. Main annual PDF exceeded web limit and direct download returned 403; REFLEX/D2NCR indexed-only details not promoted.", "queries": []}, {"target": "FPSO operators MODEC and SBM Offshore", "result": "Both names absent from full baseline profiles and packets. Original announcements distinguish development, data-platform scaling and prospective integration.", "queries": []}], "promotions": {"Repsol": "AI-237"}, "added": ["AI-237", "AI-238", "AI-239", "AI-240", "AI-241", "AI-242", "AI-243", "AI-244"], "enriched": ["AI-214"]}, "swarm6": {"summary": "The important gaps are becoming more specific: named finance deployments, comparable economics, and issuer-document coverage. This pass adds internal audit, talent matching and customer personalization; it also verifies Ecopetrol’s portfolio economics and the limits of its dispatch rollout. Several detailed vendor cases remain outside the dated register.", "findings": [{"title": "Cost and benefit figures need matching periods and scope.", "text": "Ecopetrol reports 2025 AI investment and benefits, but its 7% ROI is a five-year indicator and the reported investment excludes other AI spending. This is useful portfolio evidence, not a template for project payback.", "ids": ["AI-147"]}, {"title": "Practical work extends beyond the oilfield.", "text": "PTTEP identifies an internal-audit project; Ecopetrol describes employee-to-role matching; MOL reports tested personalization. Each addresses a defined workflow, but none discloses a measured AI-specific return.", "ids": ["AI-245", "AI-246", "AI-247"]}, {"title": "An operating center can contain a model still in development.", "text": "Ecopetrol’s CADE center is operating, while the Rubiales dispatch model was only entering MVP construction in the second half of 2025. The report does not establish fully autonomous production dispatch.", "ids": ["AI-148"]}], "priorities": [{"gap": "JIB, AFE and land/title", "status": "Named, dated deployed examples remain thin. Continental owner-relations and anonymous AFE cases are promising leads.", "next": "Obtain customer-side confirmation, dated material and task-level controls; do not promote product capability pages into deployments."}, {"gap": "Comparable economics", "status": "Ecopetrol adds aggregate investment and benefit data. Undated SLB and Halliburton cases provide scoped vendor claims.", "next": "Project-level cost perimeter, recurring spend, benefit timing, baseline and cash conversion."}, {"gap": "Issuer-source coverage", "status": "Five-company IR backlog distinguishes documents identified from documents read. Range’s Q2 call and July deck were reviewed.", "next": "Finish the linked EOG, Diamondback, Permian Resources and Antero materials. Company-profile counts are not a quarterly-source census."}, {"gap": "Failure and adoption evidence", "status": "No qualifying pilot-termination case found in this bounded search.", "next": "Named discontinued projects, reasons, sunk cost, override rates and sustained user adoption."}], "document_coverage": [{"company": "Ecopetrol", "status": "Three relevant updates: EVO workflow, AI portfolio economics and CADE phase clarification.", "documents_checked": [{"title": "2025 Integrated Management Report", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 report on 2025", "review": "Downloaded original 482-page PDF; read printed pp. 89, 229, 318 and 447; visually verified table 7.115 and its note."}], "coverage_note": "Relevant passages read; not an all-document or all-quarter company census."}, {"company": "PTTEP", "status": "Internal-audit project initiated; deployment not established.", "documents_checked": [{"title": "PTTEP 56-1 One Report 2025 — internal audit", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "2026 report on 2025", "review": "Downloaded original PDF; full printed p.223 / physical p.212 read. 27 February 2026 invitation corroborates report period."}], "coverage_note": "Relevant passages read; not an all-document or all-quarter company census."}, {"company": "MOL Group", "status": "Tested AI personalization is being scaled; platform adoption is not AI impact.", "documents_checked": [{"title": "MOL 2025 AGM documents — Consumer Services", "url": "https://molgroup.info/storage/documents/altalanos_dokumentumok/annual_meeting_documents/mol-agm-documents-2025-eng.pdf", "date": "2025 AGM material", "review": "Original PDF opened; Consumer Services paragraph on printed p.27 / physical p.29 read."}], "coverage_note": "Relevant passages read; not an all-document or all-quarter company census."}, {"company": "Range Resources", "status": "No qualifying operating AI workflow found in these two official investor materials. Do not create an initiative from the AI data-center demand forecast. This is a source-bounded negative result for the reviewed Q2 2026 transcript and July 2026 company deck, not a claim that Range has no AI activity elsewhere.", "documents_checked": [{"title": "Q2 2026 Range Resources Earnings Call — Corrected Transcript", "url": "https://ir.rangeresources.com/static-files/01ba1533-34c6-44d3-8145-5c484d81f6d9", "date": "2026-07-22", "review": "Opened and read full 20-page transcript; searched within full text for artificial intelligence, machine learning, automation and AI. No AI/ML/automation deployment, business workflow, technology implementation, or related operating metric disclosed in the transcript."}, {"title": "Company Presentation – July 21, 2026", "url": "https://ir.rangeresources.com/static-files/cd9b4c2c-f4fe-4118-9659-a4c3528b95b1", "date": "2026-07-21", "review": "Opened and read full 42-page presentation; searched within full text for artificial intelligence, machine learning, automation and AI. One substantive AI reference category: AI data centers as a driver of expected natural-gas/power demand. The presentation describes no internal AI/ML system, automation workflow, user, deployment, operating metric or realized outcome. Other technology/capacity discussion does not establish AI use."}], "coverage_note": "Only these two issuer materials were reviewed; not a claim that Range has no AI activity."}, {"company": "OMV Petrom", "status": "No discrete operating AI case from the reviewed AI-specific passages. Do not infer that OMV Petrom has no AI initiatives; its report does identify AI as an emerging risk and Board topic.", "documents_checked": [{"title": "OMV Petrom Group Annual Report 2025, Directors’ Report / Risk Management and Corporate Governance Statement", "url": "https://www.omvpetrom.com/downloads/2026/03/bb67c931-07a4-7243-67e5-8b861da2bd08/10%20OMV%20Petrom%20Group%20Annual%20Report%202025%20en%20full.pdf", "date": "2026 report on 2025; exact publication date not established", "review": "Directly opened primary-source PDF (298 pages). Read risk management pages 48–49 / PDF pp. 47–48 lines 2654–2733 and corporate governance statement pp. 80–81 lines 5342–5360 for AI-specific passages. These passages cover AI risks and governance, not operational AI workflows."}], "coverage_note": "Scoped negative result only: the AI passages reviewed were governance and cybersecurity risk disclosures, with no specific operating workflow. Not a company-wide absence claim."}], "supplemental": [{"company": "Continental Resources", "title": "UiPath owner-relations/royalty workflow agent with human approvals", "finding": "Strong future follow-up for owner relations; not a dated, reviewable in-window case on currently available page. Do not assert the webinar says production outcomes beyond its landing-page claims.", "limit": "Full page opened and read. Named customer and production deployment are claimed, but no displayed publication date, operator speaker, workflow specifics, metrics, or dated deployment evidence. Cannot establish that either publication or deployment falls inside window. Webinar recording was not accessible/reviewed.", "url": "https://start.uipath.com/How-Built-a-UiPath-Agent-to-Automate-Owner-Relations.html"}, {"company": "Corva customer (unnamed Middle Eastern national oil company)", "title": "Automated OpenWells daily drilling report ingestion and offset-well analysis used to set AFE benchmarks", "finding": "AFE-relevant lead, excluded from candidates pending customer identity and dated source/deployment confirmation.", "limit": "Full page opened and read. Vendor says first regional on-premise deployment, automated DDR ingestion, offset analysis and 9% drilling-efficiency gain. Customer is not named and page has no publication or deployment date; fails named-operator and dated-window requirements.", "url": "https://www.corva.ai/case-studies/corva-empowers-middle-eastern-noc-to-set-new-afe-benchmark"}, {"company": "Operators unnamed (TechVora/eXovy case)", "title": "RAG-grounded AI agents monitor active drilling/completion cost actuals against historical AFEs and flag discrepancies", "finding": "Interesting specific AFE workflow and economics, but customer anonymity, inaccessible primary page and absent date disqualify it.", "limit": "Search result exposed the full page text; direct open failed (cache miss), so not accepted as a full-page-reviewed source. The operator(s) are anonymous and source date absent. Claim is vendor-reported; excludes as candidate.", "url": "https://techvora.com/case-studies/oil-gas-cost-overrun-monitoring"}, {"company": "SLB / unnamed Permian operator", "title": "Chemical injection optimization: US$4.4m vendor-reported net chemical-spend saving", "finding": "Sensors, ML and weekly expert review across roughly 1,400 wells; SLB reports savings in the first nine months of an undated operating period.", "limit": "Anonymous customer and undated source. No disclosed implementation cost or AI-only increment; not attached to SLB’s separate Intelligent Production Studio profile.", "url": "https://www.slb.com/resource-library/case-study-with-navigation/production-chemicals/permian-basin-field-automations-cs"}], "excluded": [{"title": "Spearpoint Resources / Joltly", "reason": "Already represented as baseline AI-238; avoid duplicate profile padding."}, {"title": "Impact Energy / PakEnergy land-management case", "reason": "Named operator case but describes conventional land management and accounting software; no AI workflow established."}, {"title": "PakEnergy AI lease-provision feature", "reason": "Product launch/capability only; no named operator deployment."}, {"title": "Continental Resources / UiPath", "reason": "Potentially strong named owner-relations deployment, but reviewed landing page is undated, has no speaker/task-level details or visible outcome evidence, and recording not reviewed; date window cannot be verified."}, {"title": "Corva Middle Eastern NOC AFE", "reason": "Customer anonymous and source publication/deployment dates absent."}, {"title": "TechVora/eXovy AFE cost monitoring", "reason": "Customer anonymous, page undated and direct source open failed; search-result content alone is insufficient."}, {"title": "BIS Grooper Permian land-data case", "reason": "Operator unnamed; date/deployment timing unverified; secondary aggregator copy surfaced, original case not reviewed."}, {"title": "KPMG cognitive contract management", "reason": "Customer anonymous and no visible publication or deployment date."}, {"title": "Landman.ai / Landon", "reason": "Vendor product capability, without qualifying named operator customer and dated production evidence."}, {"title": "SLB chemical-injection case", "reason": "Retained as date-gated supplemental evidence only. Original vendor page has strong reported cash savings but no publication/deployment date, anonymous customer, no independent verification, no AI-only attribution, and no project or recurring cost. Cannot claim it is within the 2024-10-01 to 2026-10-01 window."}, {"title": "Halliburton AI brownfield drilling case", "reason": "Vendor states 14 hours and approximately USD 122,500 expenditure avoided, but case page is undated and customer anonymous. Time avoided is not itself cash saving; no project price or net return is disclosed."}, {"title": "ADNOC 2023 AI value announcement", "reason": "Official announcement claims USD 500m AI-generated value from more than 30 tools in 2023, outside the requested window; aggregate attribution and methodology do not establish a project-level economics case."}, {"title": "BayPine / OpenAI portfolio rollout", "reason": "Dated PE program offers ChatGPT Enterprise and deployment support to portfolio firms but reports no named energy portfolio company, deployment result, implementation cost, or realized savings. Sponsor program is not end-user oil-and-gas economics."}, {"title": "C3 AI / Baker Hughes renewal, sponsor platform announcements, AI funds, and general digital budgets", "reason": "These describe vendor partnerships, products, funding, or broad spending rather than a named operator's incremental AI implementation cost or realized return."}, {"title": "AI pilot discontinuations or failures", "reason": "Searches surfaced no dated primary-source disclosure in the target window that named a discontinued upstream/midstream AI pilot and gave a defensible reason or associated cost. Absence in this search is not evidence that failures did not occur."}, {"title": "Petrobras refinery digital twins — AI attribution not established", "reason": "Original slide 158 was downloaded and visually reviewed. It reports US$200m captured in one year from digital twins, under a broad AI-and-technologies heading, but the individual digital-twin case does not specify AI. Kept out of the AI register until an AI-specific source is found.", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0"}], "gaps": ["Named and dated operating evidence for JIB, AFE and land/title.", "AI-specific deployment cost, recurring spend and independently supported net returns.", "A company-by-quarter census of earnings calls and investor presentations.", "Public evidence on discontinued pilots and unsuccessful deployments."], "method": "Three Luna researchers searched primary sources. The lead reviewer downloaded large issuer PDFs, read relevant pages, visually checked financial tables, deduplicated workflows and separately audited the Ecopetrol claims. Search-only leads are labeled. The pass is bounded, not an exhaustive global or quarterly census.", "search_log": [{"target": "JIB / revenue accounting", "result": "No new named, dated AI deployment qualified. Spearpoint Resources/Joltly revenue-statement extraction is already baseline initiative AI-238 and was not duplicated. Rand Group's Henry Resources JIB case describes an ERP import/standardization implementation but not an AI deployment; its page's AI/MCP content is a product capability description, not evidence Henry used AI. ClarityArc and MindBridge pages are vendor capabilities rather than named customer deployments."}, {"target": "AFE / spend control", "result": "Corva NOC case is anonymous and undated. TechVora/eXovy AFE case is anonymous, undated and direct-open inaccessible. Neither qualifies despite workflow specificity and vendor-claimed results."}, {"target": "Land/title", "result": "Landman.ai and Landon describe purpose-built title products without named operator deployments or dated customer evidence. BIS/Grooper search result describes substantial AI-assisted lease-data onboarding at an unnamed Permian operator; source is a secondary case-study aggregator and date/deployment timing was not established. KPMG's anonymous 10,000-lease cognitive-contract deployment has no visible date and no named customer. No qualifying candidate."}, {"target": "Procurement", "result": "No named, dated operator AI procurement workflow surfaced in this bounded pass. A Kolena product page enumerates supply-chain/contract use cases, but its cited operator stories concern division orders/JIB/acquisition diligence and have no named organizations or dates."}, {"target": "site:corporate.exxonmobil.com AI investment cost artificial intelligence operations savings 2025; site:chevron.com AI pilot stopped; site:slb.com AI customer case study savings payback; site:halliburton.com AI customer case study savings", "result": "Found Halliburton's undated AI brownfield case and SLB's undated chemical-injection case. Other hits were non-AI savings cases, digital partnership statements, or AI-adjacent vendor information."}, {"target": "site:halliburton.com/en/resources AI USD saved operator case study; site:slb.com AI saved million case study 2025", "result": "Halliburton brownfield page reports 14 hours and roughly USD 122,500 avoided expenditure, without disclosure date or implementation price. SLB chemical-injection case reports USD 4.4m net chemical-spend savings in nine months and supporting treatment metrics, but is undated and anonymous."}, {"target": "oil gas AI pilot discontinued abandoned project 2024 2025 primary source; private equity oil gas portfolio AI implementation savings", "result": "No qualifying primary-source report of a discontinued upstream AI pilot or named PE-backed operator deployment with AI project cost/payback. Results largely covered broad sponsor programs, AI infrastructure investment, vendor claims, or non-energy portfolios."}, {"target": "site:halliburton.com Oman 2025 fully automated surface subsurface execution; SLB chemical injection 4.4 million date", "result": "Verified Halliburton/Nabors's dated 2025-04-15 primary release: qualitative ahead-of-plan/higher-ROP/lower-NPT statements only, no dollars. Opened SLB's original HTML and searched rendered page for datePublished, 2025, and 2026; no date evidence found."}, {"target": "Large-report source verification", "result": "Petrobras, Ecopetrol and PTTEP PDFs were downloaded after web-reader size limits; relevant source pages read. Petrobras digital-twin value excluded because AI attribution is not explicit."}, {"target": "Issuer backlog and Range follow-up", "result": "Five issuer IR indexes reviewed; Range call and presentation opened. Other linked calls/decks remain a backlog, not reviewed evidence."}], "backlog": [{"priority": 1, "company": "EOG Resources", "latest_event_title": "EOG Resources Second Quarter 2026 Webcast and Results", "event_date": "2026-08-05", "issuer_event_index_url": "https://investors.eogresources.com/events-and-presentations", "issuer_event_detail_url": "https://investors.eogresources.com/events-and-presentations?item=48", "issuer_results_release_url": "https://investors.eogresources.com/2026-08-04-EOG-Resources-Reports-Second-Quarter-2026-Results", "issuer_ir_home_url": "https://investors.eogresources.com/", "baseline_absent": ["https://investors.eogresources.com/events-and-presentations", "https://investors.eogresources.com/events-and-presentations?item=48", "https://investors.eogresources.com/2026-08-04-EOG-Resources-Reports-Second-Quarter-2026-Results"], "review_status": "Issuer events index and Q2 2026 release opened. Index directly lists call date, webcast, press release, presentation, guidance, statistics and filing links. Event transcript/audio not reviewed. Separate EOG profiles currently use third-party Motley Fool transcripts, so primary issuer event/source gap remains.", "next_action": "Open the event presentation and webcast transcript/audio, inspect the AI-related passage(s), and add primary-source locators only where material. Avoid redoing all quarters absent an identified gap."}, {"priority": 2, "company": "Diamondback Energy", "latest_event_title": "Diamondback Energy Second Quarter 2026 Earnings; August 2026 Investor Presentation", "event_date": "2026-08-04 (call); 2026-08-03 (presentation)", "issuer_event_index_url": "https://ir.diamondbackenergy.com/node/10756", "issuer_investor_home_url": "https://ir.diamondbackenergy.com/investors", "issuer_results_release_url": "https://ir.diamondbackenergy.com/node/17316/pdf", "baseline_absent": ["https://ir.diamondbackenergy.com/node/10756", "https://ir.diamondbackenergy.com/investors", "https://ir.diamondbackenergy.com/node/17316/pdf"], "review_status": "Issuer events/presentations index opened and read; it lists the Q2 2026 earnings call (Aug 4) and August 2026 investor presentation (Aug 3). Earnings release PDF URL reviewed from issuer IR search result. Presentation deck and webcast not opened. Existing profile uses third-party transcript, so primary source coverage is a gap.", "next_action": "Open the August 2026 presentation and Q2 call replay/transcript, scan AI/data/automation mentions, and cite relevant slide or timestamp if any."}, {"priority": 3, "company": "Permian Resources", "latest_event_title": "Q2’26 Earnings Call; Q2’26 Earnings Presentation", "event_date": "2026-08-06 (call); presentation listed with Q2’26 earnings materials", "issuer_event_index_url": "https://permianres.com/investor-relations/news-events/", "issuer_event_release_url": "https://permianres.com/permian-resources-announces-second-quarter-2026-earnings-conference-call/", "issuer_presentation_url": null, "baseline_absent": ["https://permianres.com/investor-relations/news-events/", "https://permianres.com/permian-resources-announces-second-quarter-2026-earnings-conference-call/"], "review_status": "Issuer IR news/events page and Q2 call notice opened. The page identifies the Q2’26 earnings presentation and call, with call on Aug 6. The exact presentation attachment URL was not exposed in extracted page content and remains unresolved; do not infer a file URL. Call recording not reviewed. Existing baseline has an issuer Q3 2025 PDF but not this latest call/event source.", "next_action": "Resolve the Q2 presentation attachment URL from the issuer page or site assets and review its contents; use the Q2 2026 call webcast/transcript for any relevant AI discussion."}, {"priority": 4, "company": "Range Resources", "latest_event_title": "Q2 2026 Range Resources Earnings Conference Call; Company Presentation – July 21, 2026", "event_date": "2026-07-22 (call/transcript); 2026-07-21 (presentation)", "issuer_event_index_url": "https://ir.rangeresources.com/news-and-events/webcasts", "issuer_event_detail_url": "https://ir.rangeresources.com/events/event-details/q2-2026-range-resources-earnings-conference-call/", "issuer_presentation_index_url": "https://ir.rangeresources.com/news-and-events/company-presentations", "baseline_absent": ["https://ir.rangeresources.com/news-and-events/webcasts", "https://ir.rangeresources.com/events/event-details/q2-2026-range-resources-earnings-conference-call/", "https://ir.rangeresources.com/news-and-events/company-presentations"], "review_status": "Issuer webcast/transcript index and presentation index opened; both linked documents subsequently opened and fully read. The Q2 2026 corrected transcript (July 22) and Company Presentation (July 21) were searched for artificial intelligence, machine learning, automation and AI. No internal AI workflow/deployment found; the deck's AI references concern forecast demand from AI data centers. See work/swarm6/range-review.json for findings and locators.", "next_action": "No immediate follow-up unless an independent lead points to another dated Range AI workflow; reviewed primary materials did not disclose one."}, {"priority": 5, "company": "Antero Resources", "latest_event_title": "Second Quarter 2026 Earnings Call; 2Q 2026 Earnings Presentation", "event_date": "2026-07-30 (call and presentation)", "issuer_event_index_url": "https://www.anteroresources.com/news-events/ir-calendar", "issuer_event_detail_url": null, "issuer_presentations_index_url": "https://www.anteroresources.com/news-events/presentations", "issuer_results_page_url": "https://www.anteroresources.com/news-events/press-releases/detail/259/antero-resources-announces-second-quarter-2026-financial", "baseline_absent": ["https://www.anteroresources.com/news-events/ir-calendar", "https://www.anteroresources.com/news-events/presentations", "https://www.anteroresources.com/news-events/press-releases/detail/259/antero-resources-announces-second-quarter-2026-financial"], "review_status": "Issuer calendar, presentation index, and results page opened. Calendar confirms Q2 call on July 30; presentation index lists 2Q 2026 Earnings Presentation also July 30; results page has attached presentation/webcast links. The Q2 calendar detail URL was not independently confirmed, so it is left null. Underlying materials not reviewed.", "next_action": "Resolve the official Q2 calendar detail URL if needed; open Q2 presentation and call recording/transcript and scan for named AI/analytics/business-workflow disclosures."}], "added": ["AI-245", "AI-246", "AI-247"], "enriched": ["AI-147", "AI-148"]}, "swarm7": {"summary": "This pass deepens six operator dossiers and follows the data platforms into specific upstream workflows. The strongest additions identify what the system reads, what it changes, who acts on its output, and which results are actually attributed to AI.", "findings": [{"title": "The platform relationship matters less than the workflow it enables.", "text": "Continental combines Databricks and UiPath for drilling-data validation. Expand uses Snowflake for forecasting and drilling analytics. Devon’s Snowflake–Whitson pressure calculations are physics-based: useful infrastructure, but not an established AI deployment.", "ids": ["AI-248", "AI-195"]}, {"title": "Coverage, operational effect and cash return are different measures.", "text": "Chord discloses broad rod-lift coverage and early run-time improvement. Devon discloses autonomous lift scale. Neither lets us assign the company’s broader savings program to AI. Conoco’s Montney percentages are supplier-reported early effects against an undisclosed baseline.", "ids": ["AI-097", "AI-006", "AI-202"]}, {"title": "Human decisions still define many practical applications.", "text": "Devon’s camera model alerts or dispatches an operator. Diamondback uses predicted contractor incident risk in eligibility decisions. These workflows warrant different controls and success measures from closed-loop drilling.", "ids": ["AI-249", "AI-250", "AI-251"]}], "priorities": [{"gap": "Comparable economics", "status": "Time savings, coverage and supplier-reported operating percentages; no new complete AI project cost/return case.", "next": "Implementation cost, recurring cloud/model spend, internal staffing, durable cash conversion and the correct baseline."}, {"gap": "Deployment dates and boundaries", "status": "Several informative platform customer stories are undated, or cover an integrated operator rather than upstream alone.", "next": "Operator confirmation of go-live, assets/users, evaluation period and planned versus completed expansion."}, {"gap": "Control and model performance", "status": "Sources describe alerts, approvals and some autonomous execution; most omit validation and exceptions.", "next": "False-alert rate, override rate, data/model drift, fallbacks, authorization and named process owner."}], "platforms": [{"company": "Continental Resources", "platform": "Databricks + UiPath", "kind": "Operating agent workflow; current detail undated", "workflow": "Genie/UiPath agents interpret and validate drilling-data emails, with robots routing exceptions and clean data.", "proof": "UiPath reports 200+ emails/day and approximately five minutes saved per formerly manual email.", "limit": "2025 award establishes recognition, not the date of current metrics. No independent performance audit or cash ROI.", "url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "ids": ["AI-248"]}, {"company": "Expand Energy", "platform": "Snowflake / Snowpark / Cortex", "kind": "Named ML use; fuller dashboard in development", "workflow": "Dynamic forecasts, drilling analytics, natural-language questions and SCADA-based site alerts.", "proof": "March 2026 operator release anchors the relationship; undated vendor case estimates 1,000+ labor hours annually in one unit.", "limit": "Keep the estimate separate from M&A platform benefits and from achieved production uplift.", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "ids": ["AI-195"]}, {"company": "Encino Energy", "platform": "Snowflake + Microsoft", "kind": "Data foundation; invoice AI live; agents rolling out", "workflow": "Field data in Snowflake feeds Microsoft apps. AI Builder handles invoice extraction; production-question agents were being built.", "proof": "Dated April 2025 Microsoft story; invoice app reportedly saves hundreds of hours annually for one employee.", "limit": "App-level time saving, not isolated AI value. Snowflake is not the inference engine in this account.", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "ids": ["AI-166"]}, {"company": "Devon Energy", "platform": "Snowflake + Whitson", "kind": "Operating physics workflow; AI not established", "workflow": "Two-way sharing of wellbore/time-series data supports daily bottomhole-pressure calculations for 5,000+ wells.", "proof": "Calculations return to Whitson+ and Devon SCADA; named engineering users.", "limit": "Undated story. $50,000 sensor comparison is not $250m of realized savings. Geoscience AI/ML appears as work in progress.", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/devon-energy/", "ids": ["AI-006", "AI-007"]}, {"company": "bp", "platform": "Databricks Genie / Unity Catalog", "kind": "Customer-presenter account; date unresolved", "workflow": "Natural-language maintenance questions over curated equipment/work-order data from 60 systems, including 47 SAP instances.", "proof": "Speaker reports questions answered in ~30 seconds versus weeks; MVP reached 400 users, 9,000-user rollout was an intention.", "limit": "Mixed upstream/downstream. Demo used mock data. No exact publication date, maintenance-outcome measure or project cost verified.", "url": "https://answers.databricks.com/accelerating-maintenance-conversational-ai-bp-uses-genie-databricks-WsDSWkNLmZs", "ids": []}], "document_coverage": [{"company": "Continental Resources", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Continental Resources Streamlines Operations with Automation: From fragmented emails to orchestrated operations at Continental Resources", "url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "date": "No displayed publication/update date; accessed 2026-10-01", "review": "Full primary vendor case-study page read. Customer contact is named and quoted; performance remains vendor-published/customer-attributed. Full page opened/read. Workflow and transition from human validation to automated handling at lines 241-255; quoted Sr. IT Manager Justin Cornell at 257-259; results at 260-265; 200+ emails/40+ processes/~5 minutes at 228-237."}, {"title": "Honoring the 2025 UiPath AI25 Winners", "url": "https://www.uipath.com/ai/ai25-awards/winners", "date": "Undated page; references 2025 award cohort; accessed 2026-10-01", "review": "Full page read. No publication/update or metric-measurement date displayed. Full page opened/read; Continental listing at lines 294-302, reports WellView and Databricks integration and high-90% accuracy vs ~70%, near-equal manual-intervention reduction."}, {"title": "UiPath Announces 2025 AI25 Awards Winners", "url": "https://www.uipath.com/newsroom/uipath-announces-2025-ai25-awards-winners", "date": "2025-09-29", "review": "Dated primary UiPath press release reviewed. Used solely to date the award announcement, not the operational metrics. Full press-release content reviewed from issuer's newsroom/search result. Names Continental among 2025 AI25 award winners. Does not disclose workflow details or performance metrics."}, {"title": "UiPath + Databricks: Data Intelligence Meets Agentic Automation", "url": "https://www.uipath.com/platform/agentic-automation/ai-ecosystem/databricks", "date": "Undated page; accessed 2026-10-01", "review": "Original page reviewed; no date or incremental result. Continental customer panel: Databricks Genie Agents and UiPath Agents for email processing."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "Chord Energy", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Chord Energy FY2025 results release", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "date": "2026-02-25", "review": "Full source opened and read. Operations Update, lines 210–216; specifically line 216 (AI deployment, run time metric, workover rig days)."}, {"title": "Chord Energy 1Q26 Earnings Presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2026/05/CHRD-Q1-2026-Earnings-Presentation.pdf", "date": "2026-05-05", "review": "Full 28-page PDF opened; slide 16 read. Slide 16; full extracted slide lines 457–505."}, {"title": "Chord Energy 2Q26 results release", "url": "https://ir.chordenergy.com/2026-08-05-Chord-Energy-Reports-Second-Quarter-2026-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Updates-2026-Outlook", "date": "2026-08-05", "review": "Full source opened and read. Operations Update, lines 156–161; continued AI artificial-lift optimization, with no fresh AI-specific result."}, {"title": "Chord Energy 2Q25 Investor Presentation", "url": "https://www.chordenergy.com/wp-content/uploads/2025/08/CHRD-2Q25-Earnings-Presentation.pdf", "date": "2025-08-06", "review": "Full 24-page PDF opened; lists AI/ML scheduling and artificial-lift optimization, gas-lift algorithms, predictive maintenance, and remote monitoring as continuous-improvement initiatives; no distinct deployed AI result or attribution. Slide 11 / parsed PDF page 10, lines 273–306."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "Devon Energy", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Devon Energy Q1 2026 Earnings Call Corrected Transcript", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-05-06", "review": "Original source opened; relevant passage read. pp. 3-4 (PDF pp. 3-4), CEO Clay Gaspar prepared remarks, lines 108-128; fully read 20-page transcript"}, {"title": "Devon Energy 2025 Sustainability Report", "url": "https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf", "date": "2025 (report; covers calendar 2024 unless noted)", "review": "Original source opened; relevant passage read. printed p. 15 / PDF p. 15, lines 1228-1261; full 90-page PDF read"}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "Diamondback Energy", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Diamondback Energy 2025 Corporate Sustainability Report", "url": "https://ir.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418", "date": "2025 (2024 performance reporting)", "review": "Original source opened; relevant passage read. printed p. 20 / PDF p. 20, lines 1422-1438; full 51-page PDF read"}, {"title": "Diamondback Energy Letter to Stockholders (Q3 2025)", "url": "https://ir.diamondbackenergy.com/node/17156/pdf", "date": "2025-11-03", "review": "Original source opened; relevant passage read. PDF p. 1, lines 61-64; full 3-page PDF read"}, {"title": "Diamondback Energy Letter to Stockholders (Q1 2026)", "url": "https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11", "date": "2026-05-04", "review": "Original source opened; relevant passage read. lines 72-77, especially line 74; full release read"}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "Encino Energy", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Encino Energy streamlines operations with Microsoft Power Platform", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "date": "2025-04-29 (visible on page)", "review": "Full dated official Microsoft customer story reviewed; named Encino leaders quoted. Customer-attributed vendor case study. Full page opened/read. Executive summary at lines 26-31; winter-maintenance flow/results at 42-46; field data and tank telemetry at 47-53; Snowflake/Dataverse/Copilot agent plans and example questions at 64-74; invoice AI confidence review and approval routing at 80-83."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "Expand Energy", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "date": "No publication/update date displayed; accessed 2026-10-01", "review": "Full official Snowflake customer case read; customer executive quotes present. Undated, vendor-authored, estimates/claims not independently verified. Full page opened/read. Profile and 3,700-site scale at lines 16-21, 48-52; type-curve ML and labor estimate at lines 73-80; drilling ML, Cortex questions, monitoring/digital twins and alert escalation in full page lines 85-101; future AI dashboard/ML roadmap lines 110-112."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "ConocoPhillips", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "Case Study: Field Deployments of AI-Based IOCaaS Advancing Artificial Lift and Flow Assurance", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "date": "2026-04-01", "review": "Original full article opened/read, including the Montney deployment, architecture, timeline, outcomes, and authorship. The source gives operator/team attributions but does not provide underlying data or independent validation. Montney case; architecture, implementation, first-winter results and authorship."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}, {"company": "ExxonMobil", "status": "Workflow evidence and its limits checked; use the linked profiles for detail.", "documents_checked": [{"title": "ExxonMobil Guyana Limited leverages drilling automation to set new performance benchmarks in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "date": "2026 (URL path); no exact date displayed in page text", "review": "Original full page opened/read; deployment, named products, functions, 48-km path, autonomy share, and downlink metrics appear in final case-study section. SLB is the supplier and reports the performance claims. Case study final sections: Neuro/DrillOps functions, 48 km, autonomous share and downlinks."}], "coverage_note": "Relevant sources reviewed; no claim that every quarterly document was examined."}], "supplemental": [{"company": "Continental Resources", "title": "Owner-relations agent", "finding": "UiPath promotes a deployed royalty-workflow agent with human approvals. A July 14, 2026 webinar is advertised; it does not establish go-live or each listed task as deployed.", "limit": "Landing page undated, recording not reviewed, no quantified outcome. Land, lease, suspense-release and JIB extensions remain opportunities, not confirmed deployments.", "url": "https://start.uipath.com/How-Built-a-UiPath-Agent-to-Automate-Owner-Relations.html"}], "excluded": [{"title": "Devon pressure calculation is not established AI", "reason": "The specific Snowflake–Whitson engine is physics-based. Retained in the platform comparison without inflating the AI register.", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/devon-energy/"}, {"title": "bp Genie: timing still unverified", "reason": "Substantive customer transcript retained in platform research. Relative search-index dates and a future event listing do not prove an in-window publication/deployment date; not counted as a new dated initiative.", "url": "https://answers.databricks.com/accelerating-maintenance-conversational-ai-bp-uses-genie-databricks-WsDSWkNLmZs"}, {"title": "ExxonMobil / Halliburton / Sekal Guyana", "reason": "March 2026 release supports automated well placement, but does not explicitly establish AI. Keep its metrics separate from SLB Neuro/DrillOps.", "url": "https://www.halliburton.com/en/about-us/press-release/exxonmobil-halliburton-worlds-first-fully-closed-loop-automated-well-placement-guyana"}, {"title": "Historical Shell Databricks case", "reason": "Reviewed source is copyright 2020; outside the 24-month window.", "url": "https://www.databricks.com/fr/wp-content/uploads/2020/12/Shell-Case-Study-FINAL.pdf"}, {"title": "Generic Databricks oilfield examples", "reason": "Command-center and drilling Genie articles describe architectures without a named customer deployment; product examples are not operator initiatives.", "url": "https://www.databricks.com/blog/gen-ai-powered-command-center"}], "gaps": ["Verified dates for current vendor customer stories.", "Named upstream-only economics for Snowflake and Databricks workflows.", "Results after pilot: sustained use, exception burden, failures and recurring cost.", "A complete company-by-quarter source census remains unfinished."], "method": "Three Luna researchers worked in parallel, with an additional cross-check of Continental, Chord, Encino and Expand. Original sources were reviewed, existing workflows matched before integration, and chronology and economic attribution checked. This is a targeted expansion, not an exhaustive market census.", "search_log": [{"target": "Operator deepening", "result": "Continental, Chord, Devon, Diamondback, ConocoPhillips and ExxonMobil original issuer/vendor sources reviewed. Duplicate workflows enriched; distinct practical uses added."}, {"target": "Snowflake", "result": "Expand and Encino enriched; Devon/Whitson retained as engineering/data infrastructure rather than established AI."}, {"target": "Databricks", "result": "Continental relationship confirmed via UiPath; BP maintenance transcript retained with date limits. Historical Shell and unnamed architectures excluded from dated initiative additions."}, {"target": "Aker BP / ConocoPhillips / EQT platform follow-up", "result": "No additional named upstream AI workflow verified in the bounded platform search. An EQT Lakeflow testimonial lacks operating-workflow results; absence here is not absence of activity."}], "added": ["AI-248", "AI-249", "AI-250", "AI-251"], "enriched": ["AI-097", "AI-006", "AI-045", "AI-166", "AI-195", "AI-202", "AI-072"]}, "oct4": {"as_of": "2026-10-04", "stages": [{"id": "AI-006", "stage": "Scaled", "stage_basis": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Resolve stale 'Reference pending review' flag: primary operator-hosted call verifies active scaled deployment. Replace 850-well Q1 scale with 1,000-well Q2 status; do not claim all are gas-lift or Delaware wells. The specific workflow title is gas-lift; Q2 calls this closed-loop AI artificial lift and later names smart gas lift.", "limits": "Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale.", "source_id": "SW8-001"}, {"id": "AI-097", "stage": "Scaled", "stage_basis": "Issuer FY2025 results dated 2026-02-25 say AI/ML reached about 99% of Chord’s rod-lift wells during 2025; early rod-pump run-time improvement was about 25%.", "source_url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "source_date": "2026-02-25", "evidence_origin": "Operator", "limits": "99% denominator is rod-lift wells, not total company wells. The ~25% is an early improvement, not fleet-wide average. Mixed corporate/production cost-savings numbers are not AI-attributed.", "source_id": "SW7-005"}, {"id": "AI-072", "stage": "Operating", "stage_basis": "ExxonMobil’s public digital-technologies page states its proprietary AI drilling-advisory system is currently used in Guyana and supports closed-loop automation without human intervention. The corporate page is undated, so this is the page’s present-tense status, not a date-verified change within the window.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "source_date": "Undated; checked 2026-10-04", "evidence_origin": "Operator", "review_correction": "Keep the original operator workflow and its undated source date. Do not assign the 2026 URL date from the separate SLB supplemental case to ExxonMobil’s operator disclosure or treat it as exact publication date.", "limits": "Page says currently used but gives no deployment date, scale, performance measures, autonomy boundaries, or last-updated date. Current statement alone cannot establish presence by 2026-10-01.", "source_id": "RS003"}, {"id": "AI-235", "stage": "Scaled", "stage_basis": "JPT/SPE highlight dated 2025-10-01 reports closed-loop ML gas-lift optimization deployed across more than 1,300 ExxonMobil/XTO Permian wells.", "source_url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "source_date": "2025-10-01", "evidence_origin": "Technical paper", "limits": "JPT highlight is member-exclusive after the opening summary. This supports reported workflow and scale, not test history, eligibility or active-well denominator; no operational outcome metric is available in the accessible text. The October 2025 highlight refers to earlier technical work; its underlying August 2024 paper predates the core window.", "source_id": "SW4-005"}, {"id": "AI-202", "stage": "Pilot", "stage_basis": "The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "evidence_origin": "Vendor", "limits": "Four-month observation dates are not specified. Production/LOE and no-hydrate-outage claims are source/team-reported and not independently validated; projected uplift is not an achieved outcome.", "source_id": "FU003"}, {"id": "AI-063", "stage": "Not AI", "stage_basis": "EQT/Homer City agreement in principle is for EQT to source and supply natural gas to a planned 4.4-GW power facility for a data-center campus. It is an energy-supply relationship serving AI demand, not an AI workflow operated by EQT.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "source_date": "2025-07-15; page updated 2025-07-21", "evidence_origin": "Counterparty", "review_correction": "Exclude from chart of AI workflow maturity or retain only with Not AI label; preserve the separate commercial stage as agreement-in-principle.", "limits": "Announcement is from project developer/counterparty; definitive supply contract, power operations, and data-center service start are not established by this source.", "source_id": "S42"}, {"id": "AI-095", "stage": "Unclear", "stage_basis": "EQT CEO said during the 2026-02-18 earnings call that the company had AI tools that 'would allow' it to pinpoint service-rebid efforts, while describing aggressive future rebidding and a possible low-single-digit procurement opportunity. The quote establishes tool availability, but not that this particular targeting workflow is live or piloted.", "source_url": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "source_date": "2026-02-18", "evidence_origin": "Operator", "review_correction": "Classify as Unclear rather than Pilot or Operating pending direct disclosure that AI was applied to bids. EQT has tools; target workflow/use extent remain unconfirmed.", "limits": "No named tool, user group, data, rebid count, realized savings, or deployment status. The low-single-digit cost opportunity is prospective and not AI-attributed.", "source_id": "EQTD01"}, {"id": "AI-195", "stage": "Operating", "stage_basis": "Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.", "source_url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "source_date": "2026-03-03", "evidence_origin": "Operator", "review_correction": "Retain the dated primary source as AI-195’s anchor. Distinguish Snowflake workflows described as established/current from Leucipa/Lucy deployment beginning through OSC; not all AI workflows are fully rolled out.", "limits": "No quantified drilling/forecast lift, failure reduction, or alert precision disclosed. AI dashboard development is separately described as active development in the undated Snowflake customer story.", "source_id": "RS134"}, {"id": "AI-248", "stage": "Operating", "stage_basis": "Undated UiPath customer case describes Continental’s agent/robot email intake, validation and engineer data delivery as an operating workflow. The 2025-09-29 award announcement dates recognition, not go-live or case metrics.", "source_url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "source_date": "Undated case; separate 2025 award announcement 2025-09-29", "evidence_origin": "Vendor", "limits": "Vendor/customer case has no deployment date or period for reported 200+ emails/day, five minutes/email, accuracy, or 40+ process re-use. Award page is undated; metrics remain vendor-published/customer-attributed.", "source_id": "SW7-001"}, {"id": "AI-007", "stage": "Operating", "stage_basis": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Use later primary-source operating evidence; retain 'AI agents' only if Q4 2025 deck is reopened to verify that specific term/function. The Q2 source supports operating AI D&C monitoring/benchmarking.", "limits": "No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only.", "source_id": "SW8-001"}, {"id": "AI-011", "stage": "Scaled", "stage_basis": "ADNOC’s 2025-11-03 announcement says AiPSO launched with an initial deployment across eight fields and plans to roll out further through 2027; the existing multi-field footprint is deployed while additional expansion remains planned.", "source_url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "source_date": "2025-11-03", "evidence_origin": "Operator", "limits": "No field-level adoption, achieved production metric, or independent validation; planned 2027 reach is not achieved scale.", "source_id": "S06"}, {"id": "AI-012", "stage": "Scaled", "stage_basis": "SLB’s 2026-01-23 earnings disclosure reports its remote offshore drilling operating arrangement deployed on multiple bp Azerbaijan rigs, with operational result reported in the same earnings update.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration reviewed here, and 'multiple rigs' is not a count of wells or offshore fields.", "source_id": "RS046"}, {"id": "AI-013", "stage": "Operating", "stage_basis": "SLB’s 2026-01-23 earnings update reports Neuro autonomous directional drilling applied on four OQ wells. That is field use, but four wells alone do not establish broad scaling.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration or well-by-well outcome denominator reviewed.", "source_id": "RS046"}, {"id": "AI-031", "stage": "Pilot", "stage_basis": "Repsol’s 2026 report describes a multi-agent employee-workflow platform piloted from 2025 for over four months and updated to 34 agents across three workflows. This remains explicitly pilot-stage despite the breadth of agents/workflows.", "source_url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "source_date": "2026-02-19 (update; pilot began during 2025)", "evidence_origin": "Operator", "limits": "Agent count is a dated snapshot; no mature production availability or outcome attribution beyond reported pilot results.", "source_id": "S19"}, {"id": "AI-033", "stage": "Scaled", "stage_basis": "Equinor’s 2026-01-07 operator update reports multiple implemented AI applications across maintenance, field planning and seismic, with company-reported outcomes covering 2025 and cumulative history since 2020.", "source_url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "source_date": "2026-01-07 (outcomes: 2025 and cumulative since 2020)", "evidence_origin": "Operator", "limits": "Parent total and subcase values have differing periods and must not be summed. Company values are not independently audited.", "source_id": "S20"}, {"id": "AI-042", "stage": "Operating", "stage_basis": "Woodside says AI maintenance-interval tool Maint Intel was tested on Angel and launched at North West Shelf; global rollout was scheduled during 2026. Existing operation on NWS is established, broader rollout remains planned.", "source_url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "source_date": "2026-01-05", "evidence_origin": "Operator", "limits": "No global rollout completion verified. Omit the prior 15% maintenance-hours claim unless its source is traced; reported operation does not establish the scheduled expansion occurred.", "source_id": "RS006"}, {"id": "AI-050", "stage": "Operating", "stage_basis": "PETRONAS’s 2025-11-03 release says its TIRAI assistant went live on 2025-10-18 for MBR2025 block pages in myPROdata.", "source_url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "source_date": "2025-11-03 (go-live 2025-10-18)", "evidence_origin": "Operator", "limits": "Deployment is scoped to specified block pages; no user adoption, answer-quality or time-saved metric reported.", "source_id": "S29"}, {"id": "AI-068", "stage": "Operating", "stage_basis": "SLB’s 2026-07-24 quarterly release describes an integrated closed-loop process-optimization solution in live operation at CNOOC Bohai offshore terminal.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "source_date": "2026-07-24", "evidence_origin": "Vendor", "limits": "Single terminal and vendor-reported status; no operating metrics or CNOOC corroboration reviewed.", "source_id": "S27"}, {"id": "AI-225", "stage": "Scaled", "stage_basis": "ONGC’s 2025 annual report says AI/ML-plus-physics water-injection recommendations started implementation at South Heera in March 2023 and were subsequently implemented in North and Mid Heera wells; the 2025 filing says field-performance monitoring continues. Multi-area implementation supports scaled within Heera, not whole-company scale.", "source_url": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "source_date": "2025-08-07 (disclosed; implementation 2023 onward)", "evidence_origin": "Operator", "limits": "Pilot and first implementation predate the review period; in-window evidence is continued monitoring. No quantified results or company-wide scope disclosed.", "source_id": "SW3-003"}, {"id": "AI-231", "stage": "Operating", "stage_basis": "Presidio’s 2026-06 SEC-filed investor presentation says AI well-surveillance/production intelligence is live and deployed to operators and engineers; full-fleet agent orchestration and a 3–5% uplift are future targets, with reported ~1% YTD progress on an unspecified basis.", "source_url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026069875/ea029480701_ex99-1.htm", "source_date": "2026-06 (issuer presentation)", "evidence_origin": "Operator", "limits": "Company presentation; surveillance is live but agentic orchestration remains planned. The ~1% is company-reported and measurement basis is unspecified; do not treat target uplift as achieved.", "source_id": "SW4-003"}], "outcomes": [{"id": "AI-006", "company": "Devon Energy", "workflow": "Closed-loop smart gas-lift optimization using AI models and physics-based calculations to adjust injection rates at wells.", "metric_group": "Production uplift in pilot", "unit": "% uplift", "value": null, "range_low": 2, "range_high": 3, "period": "2025 pilot; reported on 2026-05-06 earnings call.", "scope": "2025 gas-lift pilot; pilot-well count is not disclosed. Separately, live autonomous lift deployment reached >850 wells in May 2026 and 1,000 in August 2026. The pilot uplift is not the result for that later fleet.", "baseline": "Pilot production comparison method and baseline not specified in the call.", "attribution": "Devon SVP E&P Asset Management directly attributed this range to the 2025 smart-gas-lift pilot.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-05-06", "limit": "Pilot result only. In the same call, Devon said full implementation exceeded pilot uplift but declined to give a number; do not carry pilot range over to the 850+ rollout.", "cost_disclosed": false}, {"id": "AI-097", "company": "Chord Energy", "workflow": "AI-driven machine learning optimizes pumping operations on rod-lift wells.", "metric_group": "Rod-pump run time", "unit": "% longer run time", "value": 25, "period": "Early performance during 2025 rollout; reported 2026-02-25.", "scope": "AI/ML scaled to approximately 99% of Chord rod-lift wells in 2025; early performance result, with runtime cohort details not given.", "baseline": "Prior rod-pump run times; exact duration and comparison construction undisclosed.", "attribution": "Issuer release directly associates early ~25% run-time improvement with the AI-driven ML program.", "source_url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "source_date": "2026-02-25", "limit": "Issuer-reported early runtime outcome; do not interpret as 25% production uplift or failure-rate reduction. Chord separately cites ~1,200 fewer workover days from broader optimization improvements, not AI alone.", "cost_disclosed": false}, {"id": "AI-167", "company": "TotalEnergies / Microsoft", "workflow": "MARGE Copilot Studio agent analyzes and classifies industrial maintenance intervention reports to identify likely fault causes.", "metric_group": "Fault-cause identification", "unit": "% improvement", "value": null, "range_low": 25, "range_high": 45, "period": "Reported 2025-10-22; comparison period not stated.", "scope": "Maintenance report-classification workflow; site/report count and deployment scale not disclosed.", "baseline": "Prior manual analysis/classification process; precise accuracy/quality measure not defined.", "attribution": "Microsoft customer story quoting TotalEnergies' Global IT Services director; provider-hosted, company-attributed result.", "source_url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "source_date": "2025-10-22", "limit": "Improvement in identification of fault causes, not 25-45 percentage-point accuracy gain or maintenance-cost reduction. No methodology or independent audit disclosed.", "cost_disclosed": false}, {"id": "AI-168", "company": "Encino Energy / Ambyint", "workflow": "InfinityPL uses AI surveillance, anomaly detection and autonomous setpoint control for plunger-lift wells.", "metric_group": "Annual production decline rate", "unit": "percentage points lower annual decline rate", "value": 1, "period": "Full-scale deployment outcome reported 2025-08-05; exact measurement window not stated; platform timeframe 2023 to present.", "scope": "Full-scale deployment; Ambyint says 344 wells are on its platform, but does not state how many are in this metric cohort.", "baseline": "Exponential decline curves: manual/control 10% versus auto mode 9%.", "attribution": "Vendor case study quotes Encino production-engineering manager and reports the outcome; underlying well-level calculation not supplied.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "source_date": "2025-08-05", "limit": "A one-percentage-point difference in annual decline rate; do not chart as 1% production uplift. Do not combine with separate pilot normal-arrival metrics.", "cost_disclosed": false}, {"id": "AI-168", "company": "Encino Energy / Ambyint", "workflow": "Autonomous InfinityPL control optimizes plunger-lift wells and extends coverage to low-producing late-life wells.", "metric_group": "Venting events", "unit": "% fewer venting events", "value": 62, "period": "Reported 2025-08-05; exact measurement window not stated; platform timeframe 2023 to present.", "scope": "Encino deployment; well count and venting-event count not stated.", "baseline": "Prior venting-event frequency not given.", "attribution": "Vendor case study attributes the reduction to the InfinityPL deployment.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "source_date": "2025-08-05", "limit": "Vendor-reported; no event totals or measurement protocol. This is emissions-related event count, not pump runtime or production uplift.", "cost_disclosed": false}, {"id": "AI-202", "company": "ConocoPhillips / OPX Ai", "workflow": "Montney IOCaaS monitors AI-optimized wells, hydrate risk, liquid loading and compressor condition using existing SCADA/historian infrastructure.", "metric_group": "Production versus forecast", "unit": "% above forecast", "value": null, "range_low": 3, "range_high": 4, "period": "After four months of operation; source published 2026-04-01; exact evaluation dates not specified.", "scope": "AI-optimized wells in one multiwell-pad Montney asset; well count not stated. Early/partial-year result; coverage was later expanded to additional pads.", "baseline": "Asset team calculation relative to forecast; forecast method, normalization, comparison group and confidence interval not disclosed.", "attribution": "Vendor-authored JPT case study says the Montney asset team calculated the result; vendor authors include OPX Ai executives.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "limit": "Early partial-year vendor-published case. The source also mentions a ~6% expected uplift as more wells onboard; that is a projection and excluded here.", "cost_disclosed": false}, {"id": "AI-202", "company": "ConocoPhillips / OPX Ai", "workflow": "Montney IOCaaS flow-assurance surveillance and exception-driven operations.", "metric_group": "Lease operating costs", "unit": "% LOE reduction", "value": 5, "period": "Four-month evaluation; source published 2026-04-01; exact evaluation dates not specified.", "scope": "ConocoPhillips Montney asset; precise cost pool and well denominator undisclosed.", "baseline": "Prior LOE; accounting definition, nominal dollars and normalized comparison period not supplied.", "attribution": "Vendor-authored article attributes approximately 5% lower LOE to fewer emergency callouts and more efficient chemical usage.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "limit": "Percentage cost claim only; not dollar cash savings, net savings, or ROI. Vendor-authored and not independently audited.", "cost_disclosed": false}, {"id": "AI-226", "company": "ADNOC / SLB", "workflow": "AI-enabled real-time drilling operations center consolidating live rig data and insights for engineers overseeing onshore and offshore drilling.", "metric_group": "Engineering effort", "unit": "% reduction", "value": null, "range_low": 30, "range_high": 40, "period": "Reported at platform deployment, 2026-08-04; no measurement period stated.", "scope": "ADNOC RTOC across more than 120 rigs.", "baseline": "Prior multiple-tool workflows; precise baseline/time study not disclosed.", "attribution": "ADNOC issuer-reported deployment result, with SLB as system provider.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "source_date": "2026-08-04", "limit": "Company reports a 30-40% effort reduction; calculation method, role/sample, and independently measured outcome are not published. Do not convert effort reduction into payroll cash savings.", "cost_disclosed": false}, {"id": "AI-226", "company": "ADNOC / SLB", "workflow": "Centralized RTOC gives drilling engineers live fleet oversight and AI-powered performance insights.", "metric_group": "Rig oversight capacity", "unit": "rigs per engineer (multiple of prior capacity)", "value": null, "range_low": 2, "range_high": 3, "period": "Reported at deployment, 2026-08-04; no measurement period stated.", "scope": "ADNOC fleet deployment of more than 120 rigs.", "baseline": "Prior rig oversight load per engineer not numerically stated.", "attribution": "ADNOC issuer-reported capability/result.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "source_date": "2026-08-04", "limit": "Company says engineers can support two to three times more rigs; this is capacity/coverage, not a measured headcount reduction or cash saving.", "cost_disclosed": false}, {"id": "AI-248", "company": "Continental Resources / UiPath / Databricks", "workflow": "Agents and RPA validate drilling/completion data arriving in unstructured supplier emails, route errors back to suppliers and deliver clean records to engineers.", "metric_group": "Manual validation time", "unit": "minutes saved per email previously requiring manual validation", "value": 5, "period": "Case page undated; Continental is identified as a 2025 UiPath AI25 winner (award page 2025).", "scope": "Case page reports 200+ emails processed daily and one robot across 40+ processes; five minutes applies only to emails previously requiring manual validation. Do not multiply by the full daily volume.", "baseline": "Manual validation took approximately five minutes per email requiring that work; no time study or sample described.", "attribution": "Vendor case study; the cited five-minute figure is vendor-reported and not independently verified.", "source_url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "source_date": "Undated case page; 2025 award corroboration at https://www.uipath.com/ai/ai25-awards/winners", "limit": "Time saved is not cash savings, headcount reduction, or measured annual hours. The fraction of daily emails previously manually validated is undisclosed.", "cost_disclosed": false}], "platforms": [{"company": "Continental Resources", "platform": "Databricks + UiPath", "kind": "Operating agent workflow; current detail undated", "workflow": "Databricks and UiPath support agent-assisted validation of drilling and completion data arriving by email.", "proof": "UiPath reports 200+ emails/day and approximately five minutes saved per formerly manual email.", "limit": "2025 award establishes recognition, not the date of current metrics. No independent performance audit or cash ROI.", "url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "ids": ["AI-248"]}, {"company": "Expand Energy", "platform": "Snowflake / Snowpark / Cortex", "kind": "Named ML use; fuller dashboard in development", "workflow": "Dynamic forecasts, drilling analytics, natural-language questions and SCADA-based site alerts.", "proof": "March 2026 operator release anchors the relationship; undated vendor case estimates 1,000+ labor hours annually in one unit.", "limit": "Keep the estimate separate from M&A platform benefits and from achieved production uplift.", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "ids": ["AI-195"]}, {"company": "Encino Energy", "platform": "Snowflake + Microsoft", "kind": "Data foundation; invoice AI live; agents rolling out", "workflow": "Field data in Snowflake feeds Microsoft apps. AI Builder handles invoice extraction; production-question agents were being built.", "proof": "Dated April 2025 Microsoft story; invoice app reportedly saves hundreds of hours annually for one employee.", "limit": "App-level time saving, not isolated AI value. Snowflake is not the inference engine in this account.", "url": "https://www.microsoft.com/en/customers/story/23198-encino-energy-llc-power-apps", "ids": ["AI-166"]}, {"company": "Devon Energy", "platform": "Snowflake + Whitson", "kind": "Operating physics workflow; AI not established", "workflow": "Two-way sharing of wellbore/time-series data supports daily bottomhole-pressure calculations for 5,000+ wells.", "proof": "Calculations return to Whitson+ and Devon SCADA; named engineering users.", "limit": "Undated story. $50,000 sensor comparison is not $250m of realized savings. Geoscience AI/ML appears as work in progress.", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/devon-energy/", "ids": []}, {"company": "bp", "platform": "Databricks Genie / Unity Catalog", "kind": "Customer-presenter account; date unresolved", "workflow": "Natural-language maintenance questions over curated equipment/work-order data from 60 systems, including 47 SAP instances.", "proof": "Speaker reports questions answered in ~30 seconds versus weeks; MVP reached 400 users, 9,000-user rollout was an intention.", "limit": "Mixed upstream/downstream. Demo used mock data. No exact publication date, maintenance-outcome measure or project cost verified.", "url": "https://answers.databricks.com/accelerating-maintenance-conversational-ai-bp-uses-genie-databricks-WsDSWkNLmZs", "ids": []}, {"company": "Diversified Energy", "kind": "Company-reported operating architecture", "platform": "Microsoft Cloud / Snowflake / ServiceNow", "workflow": "Producer-statement normalization for accounting and revenue distribution; contract terms feed pricing and settlements.", "proof": "Governed data contracts, approval and evidence workflows, RAG agents and human oversight are described in the 2025 Sustainability Report.", "limit": "2025 report published in 2026. No workflow-level model, adoption count or financial outcome is disclosed. Statement automation is not specifically established as AI.", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "ids": ["AI-210"]}, {"company": "EQT", "kind": "Data foundation; AI capability is vendor-described", "platform": "Context Labs DaaS / emissions data framework", "workflow": "Standardized emissions calculations, reporting inputs and measurement-informed gas certification.", "proof": "Issuer report confirms 2024 production-data integration. Vendor identifies its AI-enabled platform; the operator does not specify the AI model.", "limit": "June 2025 report: production-data integration complete in 2024; midstream and offset/verification milestones still planned. No named AI model, achieved price premium or platform-attributed emissions reduction.", "url": "https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf", "ids": []}], "summary": "Six source-linked exhibits now connect the opportunity estimates, company activity, departments, deployment evidence, reported results and data platforms. A targeted three-agent review assessed 20 priority deployment claims and 12 outcome candidates, updated Devon to 1,000 live lift wells, and added context to existing Diversified and EQT workflows. No duplicate initiative profiles were added.", "gaps": [{"title": "Cash returns remain hard to compare", "text": "The outcome review did not establish a comparable AI-project cost and realized cash-return pair. Time saved, fleet scale and technical gains are different evidence types."}, {"title": "Land, JIB and AFE need named deployments", "text": "This targeted review did not verify an additional named operator deployment in these workflows. Anonymous consulting cases and generic product pages remain research leads."}, {"title": "Source dates and scope matter", "text": "Undated vendor cases stay marked. Exxon’s August 2024 technical paper predates the core window; its October 2025 JPT highlight is a later account of that earlier work. The outcome chart omits that historical primary result and an ADNOC response-time capability that is not an achieved average."}, {"title": "Review coverage is explicit", "text": "The maturity exhibit classifies 20 priority profiles, not the entire register. Research history, exclusions and search logs remain available below."}], "research": [{"company": "Diversified Energy Company", "title": "OneDEC: producer-statement normalization and governed finance/operations AI foundation", "existing_id": "AI-210", "department": "Production accounting, marketing, revenue distribution/settlements; enterprise IT and business-process automation", "workflow": "New company-authored detail for the existing OneDEC profile: statement automation digitizes and normalizes producer data across all sales points to reduce re-keying for Production Accounting, Marketing and Revenue Distribution; a centralized contract repository is described as the authoritative source for commercial terms used in pricing and settlements. The report also describes a single historian supporting manage-by-exception surveillance. This is an enrichment of the existing work, not a new duplicate record.", "technology": "OneDEC is described as a cloud-based data-centric operating environment. Microsoft Cloud provides identity/security/monitoring; operational and financial systems feed Snowflake through governed data contracts; ServiceNow enterprise playbooks provide approval/evidence workflows; APIs support portable model/agent use without data duplication. RAG agents use governed company data. Prompts, actions and outputs are versioned and monitored; logging through ServiceNow GRC supports SOX-aligned change control. The report does not specify document AI/OCR model vendors or whether producer-statement digitization itself uses generative AI.", "stage": "Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.", "metric": "No workflow-specific statement volume, extraction accuracy, processing time, exception rate, headcount change or user adoption disclosed. Report describes intended/qualitative effects: improved accuracy, eliminated re-keying, faster/more consistent pricing and settlements, lower manual oversight.", "economics": "No quantified AI/statement-automation savings, costs or payback. ROI is an evaluation gate for each AI use case, not a reported result. Separate company-wide financial performance should not be attributed to this program.", "limits": "Source is a company-authored annual sustainability report rather than a technical deployment postmortem. It reports environmental/operational data for 2025 but publishes in 2026. The workflow list sits under 'AI Efficiency at Scale,' yet no specific model is linked to statement automation or the contract repository. Architecture is company-reported; no independent validation of automation benefits. Financial controls are unusually specific, but no actual JIB or AFE workflow is claimed.", "sources": [{"title": "Diversified Energy Company Demonstrates Proven Sustainability Leadership (2025 Sustainability Report release)", "url": "https://ir.div.energy/news-events/us-press-releases/detail/225/diversified-energy-company-demonstrates-proven-sustainability-leadership", "date": "2026-05-28 (issuer release announcing 2025 report); IR report index separately lists June 17, 2026", "locator": "Official issuer release confirms report publication; report content fully opened/read in linked PDF. OneDEC/Snowflake/ServiceNow architecture: report pp. 62-63; data and controls, HITL, audit, AI governance: p. 63; statement automation, historian surveillance and pricing/settlements: p. 63. Report data period is calendar 2025 (p. 67).", "review_status": "Full issuer-authored report PDF opened/read and cross-checked against official IR release and report index. Metric year (2025) and publication date (2026) kept separate."}, {"title": "Diversified 2025 Sustainability Report", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "date": "Announced 2026-05-28; report index lists 2026-06-17", "locator": "Full 85-page issuer PDF read; especially pp. 62-63 lines 4044-4217 and reporting-period notes p. 67 lines 4256-4274.", "review_status": "Full primary issuer document reviewed. No dated workflow-specific operational metric disclosed."}]}, {"company": "EQT Corporation", "title": "Contextualized emissions data and measurement-to-report digital framework", "existing_id": null, "department": "Production operations, emissions/compliance reporting, commercial gas certification, enterprise data/IT", "workflow": "EQT's 2024 ESG Report says the company is digitizing emissions calculations and building a data framework to track, report and verify GHG emissions, with a goal of linking measurement-informed carbon intensity to differentiated gas transactions. EQT says production operational/asset data integration was completed in 2024; it lists midstream data integration for 2025 and carbon-offset data plus additional monitoring/verification methodology for 2026. Context Labs is the named advanced climate-data partner. This is a named upstream data/controls architecture update; the source does not prove a specific AI model or a fully complete, independently verified end-to-end workflow.", "technology": "EQT's source describes centralized, standardized emissions data gathering, governance and reporting, measurement-informed product certification, and a digital framework built with Context Labs. The report states only production operational/asset-data integration completed in 2024, with future midstream/offset/verification milestones. Context Labs' dated vendor page describes its AI-enabled DaaS carbon-management platform as the supporting data backbone, including measurement-based quantification, digital asset certification and secure reporting; retain vendor attribution for those product-specific claims.", "stage": "Dated EQT 2024 ESG Report published 2025-06-24. EQT reports production data integration complete in 2024 and the framework as still being built/expanded at report publication. Midstream integration targeted for 2025; carbon-offset integration and added verification methodology targeted for 2026. These planned milestones must not be restated as completed without a later source.", "metric": "No reported automation cycle-time or accuracy result for the platform. EQT separately reports 2024 Scope 1/2 emissions and a net-zero claim, but that outcome cannot be attributed to the Context Labs data platform alone. EQT says emissions figures follow stated EPA/location-based/offset boundaries; the company says its 2024 ESG report was not externally assured and net-zero does not include Scope 3 or acquired Equitrans assets in 2024.", "economics": "No system cost, labor savings, price premium, implementation cost or ROI for the data framework disclosed. The goal of measurement-backed differentiated gas and commercial transactions is forward-looking, not an achieved monetization result in these sources.", "limits": "The strongest evidence is the named customer’s own 2024 ESG report, with explicit timing/status and boundary caveats. AI-enabled characterization is in Context Labs vendor copy; EQT describes digitized calculations and data integration without specifying model/AI use, algorithm, source-system list, exception workflow or validation method. Do not infer completed midstream or carbon-offset integration from plans.", "sources": [{"title": "EQT Releases 2024 ESG Report, 'Promises Made, Promises Delivered'", "url": "https://ir.eqt.com/newsroom/news-releases/news-release-details/2025/EQT-Releases-2024-ESG-Report-Promises-Made-Promises-Delivered/default.aspx", "date": "2025-06-24", "locator": "Issuer release dates the report; complete official report PDF reviewed at https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf, p. 34: completed 2024 production-data integration, ongoing framework, 2025 midstream and 2026 offset/verification milestones; report boundary/external assurance caveats p. 2 and p. 34.", "review_status": "Official EQT IR release opened; issuer-authored ESG report PDF fully opened/read. Primary customer evidence."}, {"title": "EQT 2024 ESG Report", "url": "https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf", "date": "2025-06-24", "locator": "Printed p. 34 / PDF page 33; boundaries on p. 2", "review_status": "Issuer report relevant passages reviewed"}, {"title": "EQT Achieves Net Zero Scope 1 & 2 Emissions", "url": "https://contextlabs.com/eqt-net-zero/", "date": "2025-04-04", "locator": "Full vendor customer page reviewed; lines 55-79 identifies the Context Labs platform and claims its role as data backbone for EQT Net Zero Now+; use only to describe vendor-stated product role, not to date specific sub-workflows or attribute emissions outcome to AI.", "review_status": "Dated vendor case/press content reviewed. Secondary partner claim, used alongside EQT's primary report and qualified accordingly."}]}], "raw_reviews": {"stages": {"title": "Chart-ready operating maturity normalization", "as_of": "2026-10-04", "status_values": ["Announced", "Not AI", "Operating", "Pilot", "Scaled", "Unclear"], "classification_rule": "Stage reflects disclosed workflow deployment, not company AI ambition, vendor product launch, data-center demand, or elapsed time. Distinguish source/report date from operating/deployment date; sparse evidence is not evidence of inactivity.", "entries": [{"id": "AI-006", "stage": "Scaled", "stage_basis": "Devon’s operator-hosted Q2 2026 transcript (2026-08-05) says its closed-loop AI system autonomously optimized 1,000 wells in real time, 24/7; management said 200 wells in March and a path to broad deployment across >2,000 additional Permian wells, including first legacy Coterra wells. This supersedes the Q1 count of >850.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Resolve stale 'Reference pending review' flag: primary operator-hosted call verifies active scaled deployment. Replace 850-well Q1 scale with 1,000-well Q2 status; do not claim all are gas-lift or Delaware wells. The specific workflow title is gas-lift; Q2 calls this closed-loop AI artificial lift and later names smart gas lift.", "limits": "Transcript reports management claims; no eligible-well denominator, accuracy, lift mix by well, independent validation, or AI-specific production delta disclosed. Broader >2,000 is opportunity, not deployed scale.", "source_id": "SW8-001"}, {"id": "AI-097", "stage": "Scaled", "stage_basis": "Issuer FY2025 results dated 2026-02-25 say AI/ML reached about 99% of Chord’s rod-lift wells during 2025; early rod-pump run-time improvement was about 25%.", "source_url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend", "source_date": "2026-02-25", "evidence_origin": "Operator", "limits": "99% denominator is rod-lift wells, not total company wells. The ~25% is an early improvement, not fleet-wide average. Mixed corporate/production cost-savings numbers are not AI-attributed.", "source_id": "SW7-005"}, {"id": "AI-072", "stage": "Operating", "stage_basis": "ExxonMobil’s public digital-technologies page states its proprietary AI drilling-advisory system is currently used in Guyana and supports closed-loop automation without human intervention. The corporate page is undated, so this is the page’s present-tense status, not a date-verified change within the window.", "source_url": "https://corporate.exxonmobil.com/who-we-are/technology-and-collaborations/digital-technologies", "source_date": "Undated; checked 2026-10-04", "evidence_origin": "Operator", "review_correction": "Keep the original operator workflow and its undated source date. Do not assign the 2026 URL date from the separate SLB supplemental case to ExxonMobil’s operator disclosure or treat it as exact publication date.", "limits": "Page says currently used but gives no deployment date, scale, performance measures, autonomy boundaries, or last-updated date. Current statement alone cannot establish presence by 2026-10-01.", "source_id": "RS003"}, {"id": "AI-235", "stage": "Scaled", "stage_basis": "JPT/SPE highlight dated 2025-10-01 reports closed-loop ML gas-lift optimization deployed across more than 1,300 ExxonMobil/XTO Permian wells.", "source_url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "source_date": "2025-10-01", "evidence_origin": "Technical paper", "limits": "JPT highlight is member-exclusive after the opening summary. This supports reported workflow and scale, not test history, eligibility or active-well denominator; no operational outcome metric is available in the accessible text. The October 2025 highlight refers to earlier technical work; its underlying August 2024 paper predates the core window.", "source_id": "SW4-005"}, {"id": "AI-202", "stage": "Pilot", "stage_basis": "The 2026-04-01 JPT vendor-contributed technical case describes four months of operation on Montney multiwell pads, with further pads reported and another Canadian asset under consideration. This is an evaluated asset deployment, not demonstrated fleet-wide scale.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "evidence_origin": "Vendor", "limits": "Four-month observation dates are not specified. Production/LOE and no-hydrate-outage claims are source/team-reported and not independently validated; projected uplift is not an achieved outcome.", "source_id": "FU003"}, {"id": "AI-063", "stage": "Not AI", "stage_basis": "EQT/Homer City agreement in principle is for EQT to source and supply natural gas to a planned 4.4-GW power facility for a data-center campus. It is an energy-supply relationship serving AI demand, not an AI workflow operated by EQT.", "source_url": "https://www.homercityredevelopment.com/post/press-release-homer-city-redevelopment-announces-agreement-in-principle-for-eqt-corporation-to-supp", "source_date": "2025-07-15; page updated 2025-07-21", "evidence_origin": "Counterparty", "review_correction": "Exclude from chart of AI workflow maturity or retain only with Not AI label; preserve the separate commercial stage as agreement-in-principle.", "limits": "Announcement is from project developer/counterparty; definitive supply contract, power operations, and data-center service start are not established by this source.", "source_id": "S42"}, {"id": "AI-095", "stage": "Unclear", "stage_basis": "EQT CEO said during the 2026-02-18 earnings call that the company had AI tools that 'would allow' it to pinpoint service-rebid efforts, while describing aggressive future rebidding and a possible low-single-digit procurement opportunity. The quote establishes tool availability, but not that this particular targeting workflow is live or piloted.", "source_url": "https://s205.q4cdn.com/630272887/files/doc_financials/2025/q4/EQT-Q4-2025-Earnings-Call-Transcript.pdf", "source_date": "2026-02-18", "evidence_origin": "Operator", "review_correction": "Classify as Unclear rather than Pilot or Operating pending direct disclosure that AI was applied to bids. EQT has tools; target workflow/use extent remain unconfirmed.", "limits": "No named tool, user group, data, rebid count, realized savings, or deployment status. The low-single-digit cost opportunity is prospective and not AI-attributed.", "source_id": "EQTD01"}, {"id": "AI-195", "stage": "Operating", "stage_basis": "Expand Energy’s 2026-03-03 operator announcement describes current Snowflake platform capabilities including automated ML forecasting, hosting real-time drilling data/ML models, and SCADA-based issue detection with alerts to the Operations Support Center. In the same release Leucipa is only beginning deployment; do not conflate the two systems’ maturity.", "source_url": "https://www.expandenergy.com/2026/03/03/exe-announces-strategic-partnerships-to-strengthen-data-driven-operational/", "source_date": "2026-03-03", "evidence_origin": "Operator", "review_correction": "Retain the dated primary source as AI-195’s anchor. Distinguish Snowflake workflows described as established/current from Leucipa/Lucy deployment beginning through OSC; not all AI workflows are fully rolled out.", "limits": "No quantified drilling/forecast lift, failure reduction, or alert precision disclosed. AI dashboard development is separately described as active development in the undated Snowflake customer story.", "source_id": "RS134"}, {"id": "AI-248", "stage": "Operating", "stage_basis": "Undated UiPath customer case describes Continental’s agent/robot email intake, validation and engineer data delivery as an operating workflow. The 2025-09-29 award announcement dates recognition, not go-live or case metrics.", "source_url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "source_date": "Undated case; separate 2025 award announcement 2025-09-29", "evidence_origin": "Vendor", "limits": "Vendor/customer case has no deployment date or period for reported 200+ emails/day, five minutes/email, accuracy, or 40+ process re-use. Award page is undated; metrics remain vendor-published/customer-attributed.", "source_id": "SW7-001"}, {"id": "AI-007", "stage": "Operating", "stage_basis": "Devon’s operator-hosted 2026-08-05 call describes AI combining live drilling/completions data, elevating performance gaps, and teams acting on those alerts; management presents this as part of current operating performance. This verifies the broader live D&C analytics workflow, not every earlier presentation’s 'agent' label.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q2/Q2-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-08-05", "evidence_origin": "Operator", "review_correction": "Use later primary-source operating evidence; retain 'AI agents' only if Q4 2025 deck is reopened to verify that specific term/function. The Q2 source supports operating AI D&C monitoring/benchmarking.", "limits": "No discrete agent, rig coverage, response SLA, autonomous decision, or metric attributed solely to AI; D&C output/cost changes are not AI-only.", "source_id": "SW8-001"}, {"id": "AI-011", "stage": "Scaled", "stage_basis": "ADNOC’s 2025-11-03 announcement says AiPSO launched with an initial deployment across eight fields and plans to roll out further through 2027; the existing multi-field footprint is deployed while additional expansion remains planned.", "source_url": "https://www.adnoc.ae/en/news-and-media/press-releases/2025/adnoc-and-slb-launch-transformative-ai-powered-solution-to-boost-upstream-productivity", "source_date": "2025-11-03", "evidence_origin": "Operator", "limits": "No field-level adoption, achieved production metric, or independent validation; planned 2027 reach is not achieved scale.", "source_id": "S06"}, {"id": "AI-012", "stage": "Scaled", "stage_basis": "SLB’s 2026-01-23 earnings disclosure reports its remote offshore drilling operating arrangement deployed on multiple bp Azerbaijan rigs, with operational result reported in the same earnings update.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration reviewed here, and 'multiple rigs' is not a count of wells or offshore fields.", "source_id": "RS046"}, {"id": "AI-013", "stage": "Operating", "stage_basis": "SLB’s 2026-01-23 earnings update reports Neuro autonomous directional drilling applied on four OQ wells. That is field use, but four wells alone do not establish broad scaling.", "source_url": "https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae", "source_date": "2026-01-23", "evidence_origin": "Vendor", "limits": "Provider-reported; no operator corroboration or well-by-well outcome denominator reviewed.", "source_id": "RS046"}, {"id": "AI-031", "stage": "Pilot", "stage_basis": "Repsol’s 2026 report describes a multi-agent employee-workflow platform piloted from 2025 for over four months and updated to 34 agents across three workflows. This remains explicitly pilot-stage despite the breadth of agents/workflows.", "source_url": "https://www.repsol.com/en/press-room/press-releases/2025/repsol-launches-intelligent-multi-agent-platform-consolidating-global-ai-leader-energy-sector/index.cshtml", "source_date": "2026-02-19 (update; pilot began during 2025)", "evidence_origin": "Operator", "limits": "Agent count is a dated snapshot; no mature production availability or outcome attribution beyond reported pilot results.", "source_id": "S19"}, {"id": "AI-033", "stage": "Scaled", "stage_basis": "Equinor’s 2026-01-07 operator update reports multiple implemented AI applications across maintenance, field planning and seismic, with company-reported outcomes covering 2025 and cumulative history since 2020.", "source_url": "https://www.equinor.com/news/20260107-artificial-intelligence-saved-equinor-usd-130-million", "source_date": "2026-01-07 (outcomes: 2025 and cumulative since 2020)", "evidence_origin": "Operator", "limits": "Parent total and subcase values have differing periods and must not be summed. Company values are not independently audited.", "source_id": "S20"}, {"id": "AI-042", "stage": "Operating", "stage_basis": "Woodside says AI maintenance-interval tool Maint Intel was tested on Angel and launched at North West Shelf; global rollout was scheduled during 2026. Existing operation on NWS is established, broader rollout remains planned.", "source_url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "source_date": "2026-01-05", "evidence_origin": "Operator", "limits": "No global rollout completion verified. Omit the prior 15% maintenance-hours claim unless its source is traced; reported operation does not establish the scheduled expansion occurred.", "source_id": "RS006"}, {"id": "AI-050", "stage": "Operating", "stage_basis": "PETRONAS’s 2025-11-03 release says its TIRAI assistant went live on 2025-10-18 for MBR2025 block pages in myPROdata.", "source_url": "https://www.petronas.com/myprodata/news/tirai-your-new-petronas-myprodata-intelligence-assistance-live", "source_date": "2025-11-03 (go-live 2025-10-18)", "evidence_origin": "Operator", "limits": "Deployment is scoped to specified block pages; no user adoption, answer-quality or time-saved metric reported.", "source_id": "S29"}, {"id": "AI-068", "stage": "Operating", "stage_basis": "SLB’s 2026-07-24 quarterly release describes an integrated closed-loop process-optimization solution in live operation at CNOOC Bohai offshore terminal.", "source_url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "source_date": "2026-07-24", "evidence_origin": "Vendor", "limits": "Single terminal and vendor-reported status; no operating metrics or CNOOC corroboration reviewed.", "source_id": "S27"}, {"id": "AI-225", "stage": "Scaled", "stage_basis": "ONGC’s 2025 annual report says AI/ML-plus-physics water-injection recommendations started implementation at South Heera in March 2023 and were subsequently implemented in North and Mid Heera wells; the 2025 filing says field-performance monitoring continues. Multi-area implementation supports scaled within Heera, not whole-company scale.", "source_url": "https://nsearchives.nseindia.com/corporate/ONGC_07082025205917_covernotice_merged.pdf", "source_date": "2025-08-07 (disclosed; implementation 2023 onward)", "evidence_origin": "Operator", "limits": "Pilot and first implementation predate the review period; in-window evidence is continued monitoring. No quantified results or company-wide scope disclosed.", "source_id": "SW3-003"}, {"id": "AI-231", "stage": "Operating", "stage_basis": "Presidio’s 2026-06 SEC-filed investor presentation says AI well-surveillance/production intelligence is live and deployed to operators and engineers; full-fleet agent orchestration and a 3–5% uplift are future targets, with reported ~1% YTD progress on an unspecified basis.", "source_url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026069875/ea029480701_ex99-1.htm", "source_date": "2026-06 (issuer presentation)", "evidence_origin": "Operator", "limits": "Company presentation; surveillance is live but agentic orchestration remains planned. The ~1% is company-reported and measurement basis is unspecified; do not treat target uplift as achieved.", "source_id": "SW4-003"}], "recommendations": ["For Devon AI-006 replace stale 'Reference pending review'; anchor the stage and chronology to the operator-hosted Q2 call (2026-08-05): 1,000 live autonomous wells, with >2,000 additional Permian opportunity not counted as deployed.", "EQT AI-063 is Not AI for an AI-operations chart: it is a gas supply agreement for an AI/HPC campus; retain commercial agreement-in-principle stage separately.", "EQT AI-095 remains Unclear: AI tools are acknowledged, but the transcript does not establish that they are being applied to the prospective service-rebid workflow.", "Keep Exxon AI-072's current-use statement Operating but source date Undated; do not imply dated in-window rollout or map the separate SLB case's URL year onto Exxon’s disclosure.", "For AI-195 preserve 2026-03-03 operator evidence; treat its Snowflake vendor story as undated additional context, and separate Snowflake-hosted ML from Leucipa deployment beginning through the Operations Support Center."]}, "outcomes": {"cases": [{"id": "AI-226", "company": "ADNOC / SLB", "workflow": "AI-enabled real-time drilling operations center consolidating live rig data and insights for engineers overseeing onshore and offshore drilling.", "metric_group": "Engineering effort", "unit": "% reduction", "value": null, "range_low": 30, "range_high": 40, "period": "Reported at platform deployment, 2026-08-04; no measurement period stated.", "scope": "ADNOC RTOC across more than 120 rigs.", "baseline": "Prior multiple-tool workflows; precise baseline/time study not disclosed.", "attribution": "ADNOC issuer-reported deployment result, with SLB as system provider.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "source_date": "2026-08-04", "limit": "Company reports a 30-40% effort reduction; calculation method, role/sample, and independently measured outcome are not published. Do not convert effort reduction into payroll cash savings.", "cost_disclosed": false}, {"id": "AI-226", "company": "ADNOC / SLB", "workflow": "Centralized RTOC gives drilling engineers live fleet oversight and AI-powered performance insights.", "metric_group": "Rig oversight capacity", "unit": "rigs per engineer (multiple of prior capacity)", "value": null, "range_low": 2, "range_high": 3, "period": "Reported at deployment, 2026-08-04; no measurement period stated.", "scope": "ADNOC fleet deployment of more than 120 rigs.", "baseline": "Prior rig oversight load per engineer not numerically stated.", "attribution": "ADNOC issuer-reported capability/result.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "source_date": "2026-08-04", "limit": "Company says engineers can support two to three times more rigs; this is capacity/coverage, not a measured headcount reduction or cash saving.", "cost_disclosed": false}, {"id": "AI-226", "company": "ADNOC / SLB", "workflow": "Real-time rig-data analysis identifies drilling issues earlier to support intervention.", "metric_group": "incident_response_time_reduction", "unit": "hours faster", "value": null, "range_low": 4, "range_high": 12, "period": "Reported at deployment, 2026-08-04; no measurement period stated.", "scope": "ADNOC onshore/offshore drilling operations; fleet-wide denominator not given.", "baseline": "Prior response time not given; source reports the amount of time responses may be shortened.", "attribution": "ADNOC issuer-reported expected/achievable reduction, not a verified fleet average.", "source_url": "https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance", "source_date": "2026-08-04", "limit": "Release says the center can reduce incident response times by 4-12 hours. No case count, average, or counterfactual is supplied.", "cost_disclosed": false}, {"id": "AI-006", "company": "Devon Energy", "workflow": "Closed-loop smart gas-lift optimization using AI models and physics-based calculations to adjust injection rates at wells.", "metric_group": "Production uplift in pilot", "unit": "% uplift", "value": null, "range_low": 2, "range_high": 3, "period": "2025 pilot; reported on 2026-05-06 earnings call.", "scope": "Pilot wells; pilot well count and basin-level denominator not stated. Separate from later Delaware full implementation (>850 wells), for which management gave no quantified uplift.", "baseline": "Pilot production comparison method and baseline not specified in the call.", "attribution": "Devon SVP E&P Asset Management directly attributed this range to the 2025 smart-gas-lift pilot.", "source_url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "source_date": "2026-05-06", "limit": "Pilot result only. In the same call, Devon said full implementation exceeded pilot uplift but declined to give a number; do not carry pilot range over to the 850+ rollout.", "cost_disclosed": false}, {"id": "AI-235", "company": "ExxonMobil / XTO Energy", "workflow": "Closed-loop gas-lift workflow conducts multirate tests, acquires well data, calculates optimal injection setpoints, and applies machine-learning recommendations for wells without downhole pressure gauges.", "metric_group": "oil_production_uplift", "unit": "% average uplift", "value": 2.0, "period": "Operational result reported in SPE paper presented 2024-08-20 to 2024-08-22; paper published 2024-08-19.", "scope": "More than 1,300 ExxonMobil Permian wells; rollout deployment, not an isolated pilot.", "baseline": "Incremental uplift tracked using a model-based approach based on well productivity and steady-state well tests; detailed uncertainty/confidence interval not stated in abstract.", "attribution": "Primary SPE abstract by ExxonMobil, XTO, and ExxonMobil Global Services authors reports approximately 2.0% average oil production uplift.", "source_url": "https://onepetro.org/SPEALCE/proceedings-abstract/24ALCE/24ALCE/D031S008R001/550379", "source_date": "2024-08-19", "limit": "Averaged operator-reported result; sample period, distribution and confidence interval are not disclosed in abstract. No major OPEX/CAPEX expenditure is asserted qualitatively, but project/recurring cost and ROI are not quantified.", "cost_disclosed": false}, {"id": "AI-097", "company": "Chord Energy", "workflow": "AI-driven machine learning optimizes pumping operations on rod-lift wells.", "metric_group": "Rod-pump run time", "unit": "% longer run time", "value": 25, "period": "Early performance during 2025 rollout; reported 2026-02-25.", "scope": "AI/ML scaled to approximately 99% of Chord rod-lift wells in 2025; early performance result, with runtime cohort details not given.", "baseline": "Prior rod-pump run times; exact duration and comparison construction undisclosed.", "attribution": "Issuer release directly associates early ~25% run-time improvement with the AI-driven ML program.", "source_url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "source_date": "2026-02-25", "limit": "Issuer-reported early runtime outcome; do not interpret as 25% production uplift or failure-rate reduction. Chord separately cites ~1,200 fewer workover days from broader optimization improvements, not AI alone.", "cost_disclosed": false}, {"id": "AI-202", "company": "ConocoPhillips / OPX Ai", "workflow": "Montney IOCaaS monitors AI-optimized wells, hydrate risk, liquid loading and compressor condition using existing SCADA/historian infrastructure.", "metric_group": "Production versus forecast", "unit": "% above forecast", "value": null, "range_low": 3, "range_high": 4, "period": "After four months of operation; source published 2026-04-01; exact evaluation dates not specified.", "scope": "AI-optimized wells in one multiwell-pad Montney asset; well count not stated. Early/partial-year result; coverage was later expanded to additional pads.", "baseline": "Asset team calculation relative to forecast; forecast method, normalization, comparison group and confidence interval not disclosed.", "attribution": "Vendor-authored JPT case study says the Montney asset team calculated the result; vendor authors include OPX Ai executives.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "limit": "Early partial-year vendor-published case. The source also mentions a ~6% expected uplift as more wells onboard; that is a projection and excluded here.", "cost_disclosed": false}, {"id": "AI-202", "company": "ConocoPhillips / OPX Ai", "workflow": "Montney IOCaaS flow-assurance surveillance and exception-driven operations.", "metric_group": "Lease operating costs", "unit": "% LOE reduction", "value": 5, "period": "Four-month evaluation; source published 2026-04-01; exact evaluation dates not specified.", "scope": "ConocoPhillips Montney asset; precise cost pool and well denominator undisclosed.", "baseline": "Prior LOE; accounting definition, nominal dollars and normalized comparison period not supplied.", "attribution": "Vendor-authored article attributes approximately 5% lower LOE to fewer emergency callouts and more efficient chemical usage.", "source_url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "source_date": "2026-04-01", "limit": "Percentage cost claim only; not dollar cash savings, net savings, or ROI. Vendor-authored and not independently audited.", "cost_disclosed": false}, {"id": "AI-168", "company": "Encino Energy / Ambyint", "workflow": "InfinityPL uses AI surveillance, anomaly detection and autonomous setpoint control for plunger-lift wells.", "metric_group": "Annual production decline rate", "unit": "percentage points lower annual decline rate", "value": 1, "period": "Full-scale deployment outcome reported 2025-08-05; exact measurement window not stated; platform timeframe 2023 to present.", "scope": "Full-scale deployment; Ambyint says 344 wells are on its platform, but does not state how many are in this metric cohort.", "baseline": "Exponential decline curves: manual/control 10% versus auto mode 9%.", "attribution": "Vendor case study quotes Encino production-engineering manager and reports the outcome; underlying well-level calculation not supplied.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "source_date": "2025-08-05", "limit": "A one-percentage-point difference in annual decline rate; do not chart as 1% production uplift. Do not combine with separate pilot normal-arrival metrics.", "cost_disclosed": false}, {"id": "AI-168", "company": "Encino Energy / Ambyint", "workflow": "Autonomous InfinityPL control optimizes plunger-lift wells and extends coverage to low-producing late-life wells.", "metric_group": "Venting events", "unit": "% fewer venting events", "value": 62, "period": "Reported 2025-08-05; exact measurement window not stated; platform timeframe 2023 to present.", "scope": "Encino deployment; well count and venting-event count not stated.", "baseline": "Prior venting-event frequency not given.", "attribution": "Vendor case study attributes the reduction to the InfinityPL deployment.", "source_url": "https://www.ambyint.com/case-studies/enhancing-efficiency-and-sustainability-ambyints-infinitypl-drives-measurable-gains-for-encino-energy", "source_date": "2025-08-05", "limit": "Vendor-reported; no event totals or measurement protocol. This is emissions-related event count, not pump runtime or production uplift.", "cost_disclosed": false}, {"id": "AI-248", "company": "Continental Resources / UiPath / Databricks", "workflow": "Agents and RPA validate drilling/completion data arriving in unstructured supplier emails, route errors back to suppliers and deliver clean records to engineers.", "metric_group": "Manual validation time", "unit": "minutes saved per email previously requiring manual validation", "value": 5, "period": "Case page undated; Continental is identified as a 2025 UiPath AI25 winner (award page 2025).", "scope": "Case page reports 200+ emails processed daily and one robot across 40+ processes; five minutes applies only to emails previously requiring manual validation. Do not multiply by the full daily volume.", "baseline": "Manual validation took approximately five minutes per email requiring that work; no time study or sample described.", "attribution": "Vendor case study; the cited five-minute figure is vendor-reported and not independently verified.", "source_url": "https://www.uipath.com/resources/automation-case-studies/continental-resources-streamlines-operations-with-automation", "source_date": "Undated case page; 2025 award corroboration at https://www.uipath.com/ai/ai25-awards/winners", "limit": "Time saved is not cash savings, headcount reduction, or measured annual hours. The fraction of daily emails previously manually validated is undisclosed.", "cost_disclosed": false}, {"id": "AI-167", "company": "TotalEnergies / Microsoft", "workflow": "MARGE Copilot Studio agent analyzes and classifies industrial maintenance intervention reports to identify likely fault causes.", "metric_group": "Fault-cause identification", "unit": "% improvement", "value": null, "range_low": 25, "range_high": 45, "period": "Reported 2025-10-22; comparison period not stated.", "scope": "Maintenance report-classification workflow; site/report count and deployment scale not disclosed.", "baseline": "Prior manual analysis/classification process; precise accuracy/quality measure not defined.", "attribution": "Microsoft customer story quoting TotalEnergies' Global IT Services director; provider-hosted, company-attributed result.", "source_url": "https://www.microsoft.com/en/customers/story/25505-totalenergies-agents", "source_date": "2025-10-22", "limit": "Improvement in identification of fault causes, not 25-45 percentage-point accuracy gain or maintenance-cost reduction. No methodology or independent audit disclosed.", "cost_disclosed": false}], "findings": ["The most chartable recent operator-reported outcomes are production uplift for Devon's 2025 gas-lift pilot (2-3%), ExxonMobil/XTO's 1,300+ well gas-lift program (~2% average oil uplift), and ConocoPhillips Montney's early production above forecast (3-4%). These have different baselines and scopes; label them clearly and do not treat their percentages as strictly like-for-like.", "Keep output uplift separate from runtime: Chord's ~25% improvement is longer rod-pump run time, not a 25% oil increase. Keep Encino's 10%-to-9% annual decline change expressed as one percentage point lower decline, not as production uplift.", "ADNOC's 30-40% engineering-effort reduction and 2-3x rig capacity per engineer are capacity claims; they do not evidence payroll cash savings. Continental's five minutes is per email previously requiring manual validation; unknown share of total emails means no annual-hours extrapolation is supportable.", "No selected operational AI case discloses attributable implementation plus recurring cost and a project-level realized cash return or payback. Chord's issuer materials cite broad continuous-improvement savings, but those bundle multiple initiatives and cannot be allocated to AI without further evidence."], "search_log": ["Inspected all 251 register entries and deduplicated selected cases against existing IDs AI-006, AI-097, AI-167, AI-168, AI-202, AI-226, AI-235, and AI-248.", "Opened/read ADNOC's original 2026-08-04 issuer release for the 120+ rig RTOC and effort, oversight-capacity, response-time, and downtime claims.", "Opened/read Devon's original issuer-hosted Q1 2026 call transcript, including Q&A: separates 2025 pilot's 2-3% uplift from the unquantified 850+ well full implementation.", "Opened/read original ExxonMobil-authored SPE OnePetro abstract SPE-219553 for deployment and average uplift (paper presented August 2024). Also checked later JPT reporting; used the primary abstract's approximate 2.0% rather than the secondary staged 1.87%/2.2% figures.", "Opened/read original Chord issuer release dated 2026-02-25; retained rod-pump run time only and excluded broad optimization workover-day figure from AI-specific outcomes.", "Opened/read original 2026-04-01 JPT/SPE IOCaaS article for ConocoPhillips Montney; both production and LOE metrics are vendor-authored claims with disclosed limits.", "Opened/read Ambyint's original Encino case study dated 2025-08-05; kept full-scale decline-rate and venting outcomes distinct from pilot normal-arrival metrics.", "Opened/read UiPath's original Continental case and 2025 AI25 award page; used per-email time only, did not extrapolate based on total email volume.", "Opened/read Microsoft's TotalEnergies customer story dated 2025-10-22; distinguished MARGE fault-cause identification improvement from BuyerCompanion's 10% procurement savings target (not realized outcome).", "Checked Coterra/Halliburton 2025 autonomous-frac release: 17% stage-efficiency improvement is a named operational outcome, but the source establishes automation rather than AI; excluded it from AI outcome cases and retained as an exclusion/finding rather than mislabeling it.", "Searched for incremental project economics. No selected case had project-level AI implementation/recurring cost plus realized cash savings or payback. Chord's published category savings are mixed-initiative and not AI-attributable."], "exclusions": [{"id": "AI-201", "reason": "Chevron Kaybob 5% LOE and ~6% output versus forecast are described as first-year historical results from an older deployment; source provides no specific result year, and they are less time-comparable to recent in-window implementations than the selected recent cases."}, {"id": "AI-109", "reason": "Halliburton/Coterra reports 17% initial-rollout stage-efficiency improvement and then wider deployment, but the source proves automated fracturing, not AI. Avoid presenting it as AI-derived outcome."}, {"id": "AI-168", "reason": "Excluded the pilot's +26% normal-arrival metric and 7% improved decline versus manually optimized wells from selected chart rows because the page does not give enough definition/denominators to reconcile the metric. It also reports full-scale normal-arrival +14% and +23% vs manual; those two comparisons are not interchangeable."}, {"id": "AI-248", "reason": "Excluded accuracy rising from ~70% to the high 90s from numeric charts because exact high-end value, sample size, evaluation set and measurement protocol are not disclosed."}, {"id": "AI-167", "reason": "Excluded BuyerCompanion's estimated 10% procurement saving because the source phrases it as a future/expected target ('should allow'), not realized benefit."}]}, "gaps": {"scope": "Targeted audit for named upstream oil-and-gas deployments or dated enrichments in land/title, AFE/JIB, revenue/accounting, procurement, and operator Snowflake/Databricks architecture. Window: 2024-10-01 through 2026-10-04. Baseline: 251 profiles in work/energy-ai-site/dist/assets/research-data.json. No site files edited.", "cases": [{"company": "Diversified Energy Company", "title": "OneDEC: producer-statement normalization and governed finance/operations AI foundation", "existing_id": "AI-210", "department": "Production accounting, marketing, revenue distribution/settlements; enterprise IT and business-process automation", "workflow": "New company-authored detail for the existing OneDEC profile: statement automation digitizes and normalizes producer data across all sales points to reduce re-keying for Production Accounting, Marketing and Revenue Distribution; a centralized contract repository is described as the authoritative source for commercial terms used in pricing and settlements. The report also describes a single historian supporting manage-by-exception surveillance. This is an enrichment of the existing work, not a new duplicate record.", "technology": "OneDEC is described as a cloud-based data-centric operating environment. Microsoft Cloud provides identity/security/monitoring; operational and financial systems feed Snowflake through governed data contracts; ServiceNow enterprise playbooks provide approval/evidence workflows; APIs support portable model/agent use without data duplication. RAG agents use governed company data. Prompts, actions and outputs are versioned and monitored; logging through ServiceNow GRC supports SOX-aligned change control. The report does not specify document AI/OCR model vendors or whether producer-statement digitization itself uses generative AI.", "stage": "Issuer's 2025 Sustainability Report, published May 28, 2026, presents the OneDEC architecture and statement automation as current capabilities. It describes AI use cases as ROI-, risk- and governance-reviewed and describes RAG agents/human oversight as the operating model. Treat individual workstreams as live/current per company report; rollout dates and adoption counts for statement automation are not supplied. The report's AI controls/governance claims are distinct from proof that every cited automation uses an AI model.", "metric": "No workflow-specific statement volume, extraction accuracy, processing time, exception rate, headcount change or user adoption disclosed. Report describes intended/qualitative effects: improved accuracy, eliminated re-keying, faster/more consistent pricing and settlements, lower manual oversight.", "economics": "No quantified AI/statement-automation savings, costs or payback. ROI is an evaluation gate for each AI use case, not a reported result. Separate company-wide financial performance should not be attributed to this program.", "limits": "Source is a company-authored annual sustainability report rather than a technical deployment postmortem. It reports environmental/operational data for 2025 but publishes in 2026. The workflow list sits under 'AI Efficiency at Scale,' yet no specific model is linked to statement automation or the contract repository. Architecture is company-reported; no independent validation of automation benefits. Financial controls are unusually specific, but no actual JIB or AFE workflow is claimed.", "sources": [{"title": "Diversified Energy Company Demonstrates Proven Sustainability Leadership (2025 Sustainability Report release)", "url": "https://ir.div.energy/news-events/us-press-releases/detail/225/diversified-energy-company-demonstrates-proven-sustainability-leadership", "date": "2026-05-28 (issuer release announcing 2025 report); IR report index separately lists June 17, 2026", "locator": "Official issuer release confirms report publication; report content fully opened/read in linked PDF. OneDEC/Snowflake/ServiceNow architecture: report pp. 62-63; data and controls, HITL, audit, AI governance: p. 63; statement automation, historian surveillance and pricing/settlements: p. 63. Report data period is calendar 2025 (p. 67).", "review_status": "Full issuer-authored report PDF opened/read and cross-checked against official IR release and report index. Metric year (2025) and publication date (2026) kept separate."}, {"title": "Diversified 2025 Sustainability Report", "url": "https://www.div.energy/wp-content/uploads/2026/05/DEC_SR_2025_Final.pdf", "date": "Announced 2026-05-28; report index lists 2026-06-17", "locator": "Full 85-page issuer PDF read; especially pp. 62-63 lines 4044-4217 and reporting-period notes p. 67 lines 4256-4274.", "review_status": "Full primary issuer document reviewed. No dated workflow-specific operational metric disclosed."}]}, {"company": "EQT Corporation", "title": "Contextualized emissions data and measurement-to-report digital framework", "existing_id": null, "department": "Production operations, emissions/compliance reporting, commercial gas certification, enterprise data/IT", "workflow": "EQT's 2024 ESG Report says the company is digitizing emissions calculations and building a data framework to track, report and verify GHG emissions, with a goal of linking measurement-informed carbon intensity to differentiated gas transactions. EQT says production operational/asset data integration was completed in 2024; it lists midstream data integration for 2025 and carbon-offset data plus additional monitoring/verification methodology for 2026. Context Labs is the named advanced climate-data partner. This is a named upstream data/controls architecture update; the source does not prove a specific AI model or a fully complete, independently verified end-to-end workflow.", "technology": "EQT's source describes centralized, standardized emissions data gathering, governance and reporting, measurement-informed product certification, and a digital framework built with Context Labs. The report states only production operational/asset-data integration completed in 2024, with future midstream/offset/verification milestones. Context Labs' dated vendor page describes its AI-enabled DaaS carbon-management platform as the supporting data backbone, including measurement-based quantification, digital asset certification and secure reporting; retain vendor attribution for those product-specific claims.", "stage": "Dated EQT 2024 ESG Report published 2025-06-24. EQT reports production data integration complete in 2024 and the framework as still being built/expanded at report publication. Midstream integration targeted for 2025; carbon-offset integration and added verification methodology targeted for 2026. These planned milestones must not be restated as completed without a later source.", "metric": "No reported automation cycle-time or accuracy result for the platform. EQT separately reports 2024 Scope 1/2 emissions and a net-zero claim, but that outcome cannot be attributed to the Context Labs data platform alone. EQT says emissions figures follow stated EPA/location-based/offset boundaries; the company says its 2024 ESG report was not externally assured and net-zero does not include Scope 3 or acquired Equitrans assets in 2024.", "economics": "No system cost, labor savings, price premium, implementation cost or ROI for the data framework disclosed. The goal of measurement-backed differentiated gas and commercial transactions is forward-looking, not an achieved monetization result in these sources.", "limits": "The strongest evidence is the named customer’s own 2024 ESG report, with explicit timing/status and boundary caveats. AI-enabled characterization is in Context Labs vendor copy; EQT describes digitized calculations and data integration without specifying model/AI use, algorithm, source-system list, exception workflow or validation method. Do not infer completed midstream or carbon-offset integration from plans.", "sources": [{"title": "EQT Releases 2024 ESG Report, 'Promises Made, Promises Delivered'", "url": "https://ir.eqt.com/newsroom/news-releases/news-release-details/2025/EQT-Releases-2024-ESG-Report-Promises-Made-Promises-Delivered/default.aspx", "date": "2025-06-24", "locator": "Issuer release dates the report; complete official report PDF reviewed at https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf, p. 34: completed 2024 production-data integration, ongoing framework, 2025 midstream and 2026 offset/verification milestones; report boundary/external assurance caveats p. 2 and p. 34.", "review_status": "Official EQT IR release opened; issuer-authored ESG report PDF fully opened/read. Primary customer evidence."}, {"title": "EQT 2024 ESG Report", "url": "https://esg.eqt.com/content/EQT-2024-ESG-Report.pdf", "date": "2025-06-24", "locator": "Printed p. 34 / PDF page 33; boundaries on p. 2", "review_status": "Issuer report relevant passages reviewed"}, {"title": "EQT Achieves Net Zero Scope 1 & 2 Emissions", "url": "https://contextlabs.com/eqt-net-zero/", "date": "2025-04-04", "locator": "Full vendor customer page reviewed; lines 55-79 identifies the Context Labs platform and claims its role as data backbone for EQT Net Zero Now+; use only to describe vendor-stated product role, not to date specific sub-workflows or attribute emissions outcome to AI.", "review_status": "Dated vendor case/press content reviewed. Secondary partner claim, used alongside EQT's primary report and qualified accordingly."}]}], "search_log": [{"date": "2026-10-04", "query_or_scope": "Named upstream land/title AI, lease abstraction, owner relations, revenue accounting, JIB/AFE and procurement automation; targeted searches for operator-authored / vendor customer stories in window.", "sources_reviewed": ["KPMG client story: Tackling rapid-growth hurdles with AI (https://kpmg.com/us/en/how-we-work/client-stories/energy-cognitive-contract-management.html), full page read", "Enverus Raisa Energy automatic well matching story (https://www.enverus.com/customer-story/automatic-well-matching-accelerates-deal-evaluation/), full page read", "Diversified Energy official investor relations report and company-authored 2025 Sustainability Report, full document read", "EQT official 2024 ESG Report and 2025 publication release, full relevant report section read"], "result": "Two strongest qualified updates: Diversified (specific revenue/statement automation controls and Snowflake architecture; enriches AI-210) and EQT (dated customer-authored emissions data/reporting architecture, named Context Labs integration, staged rollout). No named, dated new AFE/JIB or land/title AI deployment with reviewed full source met the bar in this bounded pass."}, {"date": "2026-10-04", "query_or_scope": "Snowflake/Databricks architecture updates at upstream operators; AI controls and finance/revenue data processing.", "sources_reviewed": ["EQT 2024 ESG Report p. 34; June 2025 EQT publication announcement", "Diversified Energy 2025 Sustainability Report pp. 62-63; official issuer release and report index"], "result": "Positive evidence distinguishes vendor/AI platform foundation from named end-user workflow; detailed human oversight, data contracts, ServiceNow approval evidence and SOX-aligned change controls are disclosed by Diversified, while EQT provides concrete completed-versus-planned source integration milestones."}], "excluded": [{"company": "Unnamed U.S. oil-and-gas lease acquirer", "workflow": "KPMG cognitive contract management for more than 10,000 acquired lease contracts using OCR/NLP/ML and exemption clusters; claims 97% fewer reviews and hours-to-minutes individual lease review.", "reason": "Strong named workflow details but client is explicitly anonymous, so cannot add a named operator case. Full page says no client identity or publication date; exclude under named-operator and timing requirements.", "source": "https://kpmg.com/us/en/how-we-work/client-stories/energy-cognitive-contract-management.html"}, {"company": "Unnamed oil-and-gas AP client (Integritty) / unnamed global oil-and-gas AP client (OpenText VIM)", "workflow": "Dated vendor writeups describe SAP AI Core/Datasphere non-PO invoice coding or OpenText VIM and SAP/Ariba AP automation.", "reason": "Customer identity and system scope are withheld; one full-page open failed in this review. These do not support a named operator deployment or independently reviewable economics.", "sources": ["https://integritty.com/transforming-accounts-payable-automation-for-an-oil-gas-company", "https://integritty.com/modernizing-accounts-payable-with-opentext-vim"]}, {"company": "Raisa Energy", "workflow": "Enverus EnergyLink Premium automatic well-API matching to revenue and JIB statements; vendor story says deal evaluation fell from 20 hours to 5 hours across 15,000+ non-operated wells.", "reason": "Named and workflow-specific, but source page shows no publication date and describes API/network automation rather than an AI/ML workflow; search index suggests older than the requested window. Good adjacent non-AI process automation example, not an in-window AI deployment candidate.", "source": "https://www.enverus.com/customer-story/automatic-well-matching-accelerates-deal-evaluation/"}, {"company": "BW Offshore", "workflow": "Custom joint-venture billing application built with Genus; monthly close/reporting automation.", "reason": "Specific named JIB-adjacent workflow but the reviewed case does not establish AI/ML and shows no in-window date; excluded from AI-specific case additions.", "source": "https://genus2.squarespace.com/case-study/bw-offshore-joint-venture-billing-with-genus"}, {"company": "Independent E&P operators (unidentified), ClarityArc and iFactory examples", "workflow": "Vendor solution pages market AI automation for JIB, AFE, land obligations and royalty calculations; one uses anonymous testimonials/examples.", "reason": "Generic solution descriptions or anonymous customers without named, dated operator deployment evidence. Do not treat vendor product capability as a customer case.", "sources": ["https://clarityarc.com/solutions/jib-automation/", "https://ifactoryapp.com/industries/oil-and-gas/ai-driven-robotic-process-automation-rpa-in-oil-and-gas-back-office"]}]}}}, "oct14_review": {"as_of": "2026-10-14", "findings": [{"id": "O14-CNX", "record": "AI-288", "company": "CNX Resources", "title": "Safety cameras that flag unsafe conditions", "department": "HSE & emissions", "stage": "operating", "status": "Company describes current use; start date unknown", "source_id": "OCT14-CNX-SAFETY", "date": "Undated", "url": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "publisher": "CNX Resources", "grade": "B", "attribution": "operator", "source_title": "CNX Workplace Health and Safety", "locator": "Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections", "summary": "CNX says it uses AI video analytics to identify unsafe conditions, coach employees in real time and inform training. HSE++ also analyzes earlier incidents and safety observations to find trends.", "mechanism": "Video analytics flags unsafe conditions for coaching and process improvement. Separately, HSE++ predictive analytics uses historical incidents and prior safety observations. The source does not name a model, camera provider or implementation stack.", "measure": "No AI-attributed incident reduction, false-positive rate, site or camera count, implementation cost or net return is disclosed.", "operating_period": "The current company page describes use, but gives no publication date or deployment start. Its 2025 safety milestones do not date the AI workflow.", "limit": "Company-reported use, without independent validation or a quantified effect. Do not attribute station safety milestones to AI. Fleet telematics on the same page is not identified as AI. The undated page cannot establish an in-window launch.", "question": "Which unsafe conditions trigger useful interventions, how many alerts are rejected, and did incident rates change on comparable sites?", "related_ids": ["AI-288"], "assessment": "Safety AI is a practical workflow to investigate at CNX. The disclosure tells us what the tools support, but not whether they reduced incidents or review work.", "sources": [{"title": "CNX Workplace Health and Safety", "url": "https://www.cnx.com/sustainability-radical-transparency/people/workplace-health-and-safety/", "date": "Undated", "publisher": "CNX Resources", "locator": "Artificial Intelligence section; separate Measuring Performance and Safety Milestones sections", "access": "Relevant original source reopened 14 October 2026. Source claims are not independently validated."}]}, {"id": "O14-AKER", "record": "AI-289", "company": "Aker BP", "title": "Agents for risk assessment and alarm reviews", "department": "HSE & emissions", "stage": "announced", "status": "Implementation announced; operating results not established", "source_id": "OCT14-AKER-ABB", "date": "2026-05-21", "url": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "publisher": "Cognite", "grade": "C", "attribution": "vendor", "source_title": "Cognite and ABB collaborate on industrial agentic applications", "locator": "Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits", "summary": "Cognite and ABB name Aker BP as the first customer signed on to implement an agentic offering for risk assessment and alarm rationalization.", "mechanism": "The announced collaboration would add Cognite Industrial AI and Data capabilities to ABB Ability SafetyInsight and AlarmInsight. The release describes agents coordinating work across systems to support operational risk assessment and alarm rationalization. It does not establish live control authority or autonomous safety decisions.", "measure": "No achieved result is disclosed for this offering. Faster workflow claims and hundreds of agents are ambitions. Aker BP production efficiency and future production targets are company context, not AI-attributed outcomes.", "operating_period": "Published 21 May 2026. Aker BP signed on to implement; completion date, site scope and sustained live use are not established.", "limit": "Vendor announcement, not proof of scaled operation. Existing root-cause-analysis benefits in AI-049 must not be transferred to this workflow. AI-131 covers Microsoft Copilot and is a separate initiative. No measured safety benefit, cost or net return is disclosed.", "question": "Which risk and alarm reviews are live, what actions can agents take, and where are human approval and safety-system boundaries?", "related_ids": ["AI-289", "AI-049"], "assessment": "This moves Aker BP’s agent ambitions toward a specific industrial job. The next evidence needed is a live workflow, its authority limits and an observed result.", "sources": [{"title": "Cognite and ABB collaborate on industrial agentic applications", "url": "https://www.cognite.com/en/company/newsroom/cognite-and-abb-collaborate-to-integrate-agentic-ai-into-industrial-applications-to-deliver", "date": "2026-05-21", "publisher": "Cognite", "locator": "Published date; Aker BP first-customer paragraph; ABB Ability SafetyInsight and AlarmInsight; intended workflow benefits", "access": "Relevant original source reopened 14 October 2026. Source claims are not independently validated."}]}, {"id": "O14-EARTH", "record": null, "company": "Anonymous operator in Oman", "title": "One fracture-model output failed; the workflow was narrowed", "department": "Exploration & geoscience", "status": "Vendor account of a failed output; anonymous customer", "date": "2026-09-10", "url": "https://www.earthscan.io/v2/case-studies/the-generalization-cliff-what-horizontal-wells-did-to-our-fracture-model", "publisher": "EarthScan", "source_title": "The Generalization Cliff: Turning a Broken Azimuth Metric Into a Deployment Protocol", "locator": "The one line of measurement; output-specific tolerances; the hand-off; scope and comparator limitations", "summary": "EarthScan reports that fracture-direction estimates became unreliable on horizontal wells. The team removed that output from automation and kept human interpretation, while using fracture location and depth as an aid.", "mechanism": "The vendor describes a blind evaluation on five horizontal wells in an Oman carbonate dataset. Its team and the operator’s interpreter agreed acceptance bands separately for depth, dip and azimuth. The source has ambiguous descriptions of training and transfer; it does not support a clean matched model comparison.", "measure": "Reported horizontal azimuth mean absolute error: 65.4 degrees, outside the interpreter’s 5–15 degree tolerance bands. Vertical reference was about 10 degrees. Vertical depth, dip and F1 comparators were not from a matched controlled run. No review hours, cash savings or net return are disclosed.", "operating_period": "Case published 10 September 2026; the blind test was in July 2023, outside the initial screening window. The publication describes subsequent use as a preprocessing aid, without a dated deployment population.", "limit": "Anonymous customer, vendor-authored account, five horizontal test wells and confidential data/code. The test was in July 2023; September 2026 is the publication date. This shows a failed output and narrower automation, not an abandoned program. It is not included in the named-company adoption or numerical-result counts.", "question": "Does each output meet the engineer’s tolerance on each well type, and how much interpretation work remains after automation?", "related_ids": [], "sources": [{"title": "The Generalization Cliff: Turning a Broken Azimuth Metric Into a Deployment Protocol", "url": "https://www.earthscan.io/v2/case-studies/the-generalization-cliff-what-horizontal-wells-did-to-our-fracture-model", "date": "2026-09-10", "publisher": "EarthScan", "locator": "The one line of measurement; output-specific tolerances; the hand-off; scope and comparator limitations", "access": "Relevant original source reopened 14 October 2026. Source claims are not independently validated."}]}, {"id": "O14-EXPAND", "record": "AI-195", "company": "Expand Energy", "title": "Forecasting, site data and integration at Expand", "department": "Drilling & completions", "status": "Existing evidence made easier to find; no new project", "date": "Undated", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "publisher": "Snowflake", "source_title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "locator": "Digging deeper into operational efficiency; drilling and site monitoring; M&A integration; Energizing the future", "summary": "The Snowflake case also covers ML well forecasts, natural-language drilling questions, SCADA-linked site data and M&A data integration. These details already appear in AI-195’s research dossier.", "mechanism": "Snowpark ML supports production/type-curve forecasting. Drilling ML and operational alerts use centralized data; Cortex Analyst supports natural-language queries. The case describes site data for 3,700 locations. A full drilling dashboard is under development and broader MLOps maturity is on the roadmap.", "measure": "The vendor estimates more than 1,000 annual labor hours in one business unit. Its headline reports 40% less time on data-related M&A tasks; a customer quote also mentions cost. These are not demonstrated AI-only cash savings, and methodology and observation periods are unspecified.", "operating_period": "Undated vendor page, already reviewed 6 October 2026. It does not establish a new 2026 launch. Current workflows, the developing dashboard and the platform roadmap have different stages.", "limit": "Same source and substantive claims already in AI-195. No additional adoption or numerical-result count. Site data coverage is not proof of automatic control at all 3,700 locations. Keep this platform separate from Leucipa in AI-194.", "question": "Which business unit uses the forecasts, how was the labor estimate made, and what incremental benefit comes from ML rather than data consolidation?", "related_ids": ["AI-195"], "sources": [{"title": "Striking Data Gold: Expand Energy Taps into Snowflake to Scale Smarter", "url": "https://www.snowflake.com/en/customers/all-customers/case-study/expand-energy/", "date": "Undated", "publisher": "Snowflake", "locator": "Digging deeper into operational efficiency; drilling and site monitoring; M&A integration; Energizing the future", "access": "Relevant original source reopened 14 October 2026. Source claims are not independently validated."}]}], "method": "Three targeted Luna research tracks, followed by editor checks of original sources and existing dossiers. This is not an exhaustive review of all earnings calls or private deployments.", "count_note": "Two named-company workflow profiles added: CNX current-use disclosure and Aker BP implementation announcement. The anonymous failure is a separate note. Expand is existing evidence; no duplicate project or result is added."}, "oct14b_review": {"as_of": "2026-10-14", "findings": [{"id": "O14B-XINJIANG", "record": "AI-290", "company": "PetroChina", "title": "A model that diagnoses rod-pumped wells in Xinjiang", "department": "Production & field operations", "stage": "scaled", "grade": "B", "attribution": "operator", "date": "2026-04-15", "url": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "source_id": "OCT14B-CNPC-XINJIANG", "publisher": "CNPC / China Petroleum News", "source_title": "CNPC: industrial large model for rod-pumped-well production optimization", "source_type": "Company news article", "locator": "Article dated 15 April 2026; first four body paragraphs", "status": "Company reports scaled use on more than 3,800 wells", "summary": "CNPC says a production-diagnostics model is used on more than 3,800 rod-pumped wells in Xinjiang. Staff receive an abnormal-condition diagnosis, the evidence behind it and suggested actions.", "mechanism": "A one-billion-parameter Transformer model with contrastive learning uses production data from 78,000 rod-pumped wells across Daqing, Liaohe, Changqing and Xinjiang. Models are fine-tuned for operating scenarios. The first deployed scenario diagnoses well conditions, colors a health display and supplies evidence and recommendations. Online liquid measurement and parameter recommendations are expanding scenarios; broader embodied-agent research is future work.", "measure": "Company reports average condition-diagnosis accuracy above 90%, abnormal-well discovery moving from days to minutes, and wells managed per person increasing from 3.5 to 11.5. It also reports usable data rising from 40% to above 95%. Validation definitions, samples, observation periods and staffing populations are not supplied.", "operating_period": "15 April 2026 disclosure: more than 3,800 Xinjiang wells in use. It describes coverage of light-oil rod-pumped wells at named plants including plants 1 and 2 and Baikouquan. The 78,000-well figure concerns data from four oilfields, not the live rollout.", "limit": "Company-reported results without independent validation or complete cost/return data. Do not treat training-data population as deployment reach, multiply staffing ratios into cash savings, or assume automatic control. A Transformer industrial model is not established as a text LLM or a company communication brain.", "question": "How does diagnosis hold up across lift conditions, and which recommendations lead to useful field interventions after false alerts and support work?", "related_ids": ["AI-290", "AI-081", "AI-236"], "sources": [{"title": "CNPC: industrial large model for rod-pumped-well production optimization", "url": "https://news.cnpc.com.cn/system/2026/04/15/030190061.shtml", "date": "2026-04-15", "publisher": "CNPC / China Petroleum News", "locator": "Article dated 15 April 2026; first four body paragraphs", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}, {"id": "O14B-CNX", "record": "AI-291", "company": "CNX Resources", "title": "Finding methane-release precursors and prioritizing repairs", "department": "HSE & emissions", "stage": "operating", "grade": "B", "attribution": "operator", "date": "Undated", "url": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "source_id": "OCT14B-CNX-METHANE", "publisher": "CNX Resources", "source_title": "CNX Commitment to Methane Reduction", "source_type": "Company website", "locator": "2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys", "status": "Company describes current use; exact start date unknown", "summary": "CNX describes ML on well and equipment signals to identify conditions preceding methane releases. ML-assisted camera and aerial analysis helps prioritize sources, while operating data is reconciled with the emissions inventory.", "mechanism": "Well/equipment performance data is used to identify release precursors. OGI camera video and aerial sensor data help identify sources and prioritize high-emitting components. Integrated operating and emissions-inventory data surfaces anomalies. The surrounding monitoring program includes remote equipment shutdown and in-house leak-tracking/work-order software; the source does not establish autonomous AI shutdown.", "measure": "CNX describes qualitative success in Q1 2026. No AI-specific event count, avoided methane volume, precision/recall, repair-time change, asset population, implementation cost or net return is disclosed. Earlier company-wide emissions reductions are not AI results.", "operating_period": "Undated current company page, with explicit 2026 program and Q1 2026 progress. No model launch date or deployment population is given.", "limit": "This is a verified October 9 research backlog item, separate from AI-288 safety cameras. The company’s 2020–24 emissions reductions and its 2025 reporting-baseline reset cannot establish AI-attributable improvement. Remote control capability is not evidence of autonomous AI authority.", "question": "Which release precursors are detected reliably, how many interventions prevent a release, and how much inspection and repair work remains?", "related_ids": ["AI-291", "AI-288"], "sources": [{"title": "CNX Commitment to Methane Reduction", "url": "https://www.cnx.com/sustainability-radical-transparency/environment/commitment-to-methane-reduction/", "date": "Undated", "publisher": "CNX Resources", "locator": "2026 Performance; Innovation Spotlight: Continuous Monitoring; LDAR and advanced surveys", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}, {"id": "O14B-HG", "record": "AI-292", "company": "HG Oil Holdings", "title": "An invoice assistant with human exception review", "department": "Finance, investment & strategy", "stage": "operating", "grade": "C", "attribution": "vendor", "date": "Undated", "url": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "source_id": "OCT14B-HG-INVOICES", "publisher": "Preisser Solutions", "source_title": "HG Oil Holdings: AI invoice processing", "source_type": "Vendor customer case", "locator": "Before; What we built; Operational handoffs; Outcomes", "status": "Undated vendor-described use; benefits not independently verified", "summary": "Preisser describes an invoice assistant for HG Oil that extracts and codes items, flags anomalies and routes approvals. Low-confidence extractions go to people for review.", "mechanism": "Staff upload invoices; the vendor describes extraction of vendor details, line items, totals and GL codes, expense categorization, anomaly flags and approval routing by thresholds and cost centers. A human exception queue handles low-confidence extractions. A separate vendor product page describes human payment approval and names HG as its reference; that control is not stated in the named case alone.", "measure": "Vendor reports 75% less manual invoice handling, more than 40 staff hours/week redirected and an avoided administrative hire. Its baseline is also described as more than 40 hours/week, but the page does not provide a common measured period or exact time denominator. These are reported labor/capacity claims, not audited cash savings.", "operating_period": "Both vendor case and operator About page are undated. The operator identifies itself as a Kansas producer; its company-wide well count is not the invoice-system population. The disclosure cannot establish an in-window deployment date.", "limit": "Vendor-only AI deployment and benefit claims; HG’s own website corroborates operator identity, not this automation. No invoice volume, error/exception rate, staffing denominator, matched time study, costs or independently validated return. Do not calculate dollar ROI or date the deployment from search metadata.", "question": "How many invoices pass without correction, what review effort remains, and does released capacity reduce costs or improve the close?", "related_ids": ["AI-292"], "sources": [{"title": "HG Oil Holdings: AI invoice processing", "url": "https://preissersolutions.com/case-studies/hg-oil-ai-invoice-processing", "date": "Undated", "publisher": "Preisser Solutions", "locator": "Before; What we built; Operational handoffs; Outcomes", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}, {"id": "O14B-SANTOS", "record": "AI-139", "company": "Santos", "title": "Production AI, virtual pressure measurement and drilling controls", "department": "Production & field operations", "date": "2026-05-26", "url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "publisher": "Santos", "source_title": "Santos 2026 Investor Briefing Day", "locator": "Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross", "status": "Specific workflows now reviewed; achieved benefits and targets separated", "summary": "Santos’s deck has more specific evidence than our earlier excerpt. It separates production-optimization AI, a pressure model replacing physical gauges, rig-connected drilling automation and stimulation logistics.", "mechanism": "The pressure model estimates bottom-hole pressure instead of relying on physical downhole gauges. A separate drilling-automation column describes rig-control integration monitored from an operating center. The production-optimization AI column reports delivered benefits. The fourth column concerns bulk proppant/fluid handling and is not identified as AI.", "measure": "Production-optimization AI: $28m benefits delivered in 2025 and over $64m potential annual benefit by 2029. Pressure model: up to about $70m potential capex reduction. The two potential dollar figures are gross. Separate columns claim 20% penetration-rate improvement and 80% US frac-installation productivity improvement; these must not be assigned to the pressure model.", "operating_period": "26 May 2026 investor briefing; the deck reports 2025 production-AI benefits and 2029 potential. The stimulation column compares 2026 productivity with a 2023 baseline. No well/rig population or model-specific measurement period is supplied.", "limit": "The pressure-model figure is potential, not achieved savings. No pressure-model accuracy, complete project costs or matched net return. The production-AI program may overlap AI-083/Xecta, so its benefit remains outside numerical-result counts pending scope reconciliation. Drilling integration belongs to its own column; stimulation logistics is adjacent technology.", "question": "Where is virtual pressure measurement accepted as a replacement for gauges, and which costs and operating limits determine the realized benefit?", "related_ids": ["AI-139", "AI-083"], "sources": [{"title": "Santos 2026 Investor Briefing Day", "url": "https://www.santos.com/wp-content/uploads/2026/05/Santos-2026-Investor-Briefing-Day_ASX.pdf", "date": "2026-05-26", "publisher": "Santos", "locator": "Release cover; slide 72 / PDF page 74, four separate columns; footnote 1: Gross", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}, {"id": "O14B-KPMG-GLOBAL", "record": null, "company": "KPMG International", "title": "Technology priorities, with an oil-and-gas subgroup", "department": "Industry research", "date": "2026 edition", "url": "https://assets.kpmg.com/content/dam/kpmgsites/ie/pdf/insights/energy/ie-kpmg-global-tech-report-2026-energy-resources-chemicals.pdf.coredownload.inline.pdf", "publisher": "KPMG International", "source_title": "KPMG Global tech report 2026: Energy, Natural Resources and Chemicals", "locator": "Printed pages 6–7: priorities and AI strategy; page 16: methodology", "status": "Survey of larger companies; priorities are not realized returns", "summary": "The report includes 58 oil-and-gas technology leaders within a 258-person energy/extractives sample. It helps compare priorities, but does not measure whether a particular AI project paid off.", "mechanism": "Respondents represent organizations above US$1bn annual revenue across 22 countries and territories. The broader sample also includes mining, chemicals, utilities and renewables. The oil-and-gas subgroup and broader energy findings have different populations.", "measure": "79% of the oil-and-gas subgroup prioritizes AI/automation; the broader energy/extractives figure is 69%, versus 77% across sectors. The report says IT leads AI strategy at 58% of the broader energy group. These are self-reports, not operating measurements.", "operating_period": "2026 report edition; exact publication day and survey fieldwork dates were not established in this review.", "limit": "Small oil-and-gas subgroup; companies below US$1bn are outside the stated population. AI/automation includes multiple technologies. Expected future scaling is not achieved deployment. Do not apply these percentages to our project register or infer net ROI.", "question": "Do the priorities translate into sustained use and measured outcomes on comparable assets?", "related_ids": [], "report_id": "O14B-KPMG-GLOBAL", "sources": [{"title": "KPMG Global tech report 2026: Energy, Natural Resources and Chemicals", "url": "https://assets.kpmg.com/content/dam/kpmgsites/ie/pdf/insights/energy/ie-kpmg-global-tech-report-2026-energy-resources-chemicals.pdf.coredownload.inline.pdf", "date": "2026 edition", "publisher": "KPMG International", "locator": "Printed pages 6–7: priorities and AI strategy; page 16: methodology", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}, {"id": "O14B-KPMG-US", "record": null, "company": "KPMG US", "title": "US energy companies’ reported technology maturity", "department": "Industry research", "date": "2026 edition", "url": "https://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2026/how-energy-can-scale-advanced-tech-safely.pdf", "publisher": "KPMG US", "source_title": "How energy can scale advanced tech safely", "locator": "PDF page 3 / printed page 1: population; PDF page 5 / printed page 3: maturity comparisons", "status": "Broader-energy survey; no upstream-only deployment rate", "summary": "KPMG’s US energy report adds a broader view of the gap between technology ambition and implementation. It covers 100 US energy-business executives within a global survey.", "mechanism": "The report describes 2,500 executives across 27 countries, including 100 from US energy businesses. Its maturity comparisons concern multiple technologies and energy subsectors. It is a separate publication from the global energy/extractives report.", "measure": "52% of energy respondents describe advanced AI/automation implementation, versus 57% elsewhere. For data/analytics the comparison is 45% versus 60%. These are reported implementation stages, not audited business results.", "operating_period": "2026 report edition; exact publication day not established.", "limit": "No separate upstream or oil-and-gas denominator is supplied for these comparisons. Do not merge the US sample with the 258-person global energy sample; overlap is not established. Technology maturity does not establish AI-only cash returns.", "question": "Which data and integration bottlenecks prevent useful systems from becoming part of everyday work?", "related_ids": [], "report_id": "O14B-KPMG-US", "sources": [{"title": "How energy can scale advanced tech safely", "url": "https://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2026/how-energy-can-scale-advanced-tech-safely.pdf", "date": "2026 edition", "publisher": "KPMG US", "locator": "PDF page 3 / printed page 1: population; PDF page 5 / printed page 3: maturity comparisons", "access": "Relevant original text reviewed 14 October 2026. CNPC encoding recovered; Santos column attribution checked on rendered original slide. Source review is not independent validation."}]}], "method": "Three targeted Luna tracks followed by editor checks of original sources and existing dossiers. A bounded pass, not a census of every disclosure.", "count_note": "Three workflow profiles added, Santos’s existing profile expanded, and two industry reports added. CNX methane was already in the research backlog. Undated vendor claims and potential benefits keep their qualifications."}, "oct14c_review": {"as_of": "2026-10-14", "findings": [{"id": "O14C-ARAMCO", "record": "AI-293", "company": "Aramco", "title": "Connecting rock samples to a geological knowledge platform", "department": "Exploration & geoscience", "stage": "operating", "grade": "B", "attribution": "operator", "date": "2026-08-25", "source_id": "OCT14C-ARAMCO-CORE", "publisher": "Aramco", "source_type": "Company magazine article", "source_title": "Understanding the subsurface: inside Aramco’s Core Laboratories Center", "url": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "locator": "25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration", "status": "Company describes use; additional ML work is underway", "summary": "Aramco describes AI-supported core analysis and a shared geological knowledge platform. Samples, images, lab results, interpretations and production outcomes are connected for geoscientists.", "mechanism": "Digitized rock imaging feeds AI pattern recognition and geological knowledge graphs. Teams use the linked information to assess reservoir quality and development opportunities. The same article says additional ML work on integrated datasets is underway.", "measure": "No AI-specific accuracy, turnaround improvement, cost or financial result is disclosed.", "operating_period": "Published 25 August 2026. The center’s 150+ employees and archive of more than one million feet of samples are facility context, not AI deployment or validation denominators.", "limit": "Issuer account with mixed technology maturity. Routine lab turnaround is not an AI-attributed result. No company-wide communications brain, executive oracle, autonomous field control or independently validated benefit is established.", "question": "Which geological judgments improve when teams can connect the sample, its analysis and the well’s later performance?", "related_ids": ["AI-293", "AI-281", "AI-282"], "sources": [{"title": "Understanding the subsurface: inside Aramco’s Core Laboratories Center", "url": "https://www.aramco.com/en/news-media/elements-magazine/2026/understanding-the-subsurface-inside-aramco-core-laboratories-center", "date": "2026-08-25", "publisher": "Aramco", "locator": "25 August 2026 byline; Utilizing advanced technologies; Shaping the future of exploration", "access": "Published 25 August 2026. The center’s 150+ employees and archive of more than one million feet of samples are facility context, not AI deployment or validation denominators. Original-source review is not independent validation."}]}, {"id": "O14C-EXXON", "record": "AI-072", "company": "ExxonMobil", "title": "Autonomous drilling in Guyana, with supplier-reported results", "department": "Drilling & completions", "stage": "operating", "grade": "C", "attribution": "vendor", "date": "Undated (2026 URL)", "source_id": "SW7-014", "publisher": "SLB", "source_type": "Vendor customer case", "source_title": "ExxonMobil Guyana: drilling automation in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "locator": "Results: Autonomous drilling and rig automation drive record-setting performance in Guyana", "status": "Existing operating profile; supplier-reported results now visible", "summary": "SLB’s named ExxonMobil case describes Neuro directional control and DrillOps surface-equipment automation, with shore teams monitoring the work.", "mechanism": "Neuro updates trajectory and requests downlink commands. DrillOps controls pumps, drawworks and top drive, adjusts drilling parameters and responds to drilling dysfunction. Operations are monitored from an onshore center.", "measure": "SLB reports more than 48 km of complex well paths, more than 93% of operations executed autonomously, 51% fewer controlled-ROP downlinks and 75% fewer off-bottom downlinks since implementation.", "operating_period": "Undated case in the supplier’s 2026 URL directory. Exact start, observation period, well count and definition of the operations denominator are not supplied.", "limit": "Supplier-reported operational measures, not independently audited performance or cash ROI. This substantiates the Neuro/DrillOps case, not the separate pending March 2026 first-well claim or its different vendor attribution. Existing source history is retained.", "question": "How much well-delivery time changed after automation, and how often did shore teams intervene?", "related_ids": ["AI-072"], "sources": [{"title": "ExxonMobil Guyana: drilling automation in deepwater operations", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/exxonmobil-leverages-drilling-automation-to-set-new-performance-benchmarks-in-deepwater-operations", "date": "Undated (2026 URL)", "publisher": "SLB", "locator": "Results: Autonomous drilling and rig automation drive record-setting performance in Guyana", "access": "Undated case in the supplier’s 2026 URL directory. Exact start, observation period, well count and definition of the operations denominator are not supplied. Original-source review is not independent validation."}]}, {"id": "O14C-ECOPETROL", "record": "AI-148", "company": "Ecopetrol", "title": "An operating energy desk, with a dispatch model still being built", "department": "IT, data & enterprise transformation", "stage": "operating", "grade": "A", "attribution": "operator", "date": "2026 (2025 annual report)", "source_id": "OCT14C-ECOPETROL-CADE", "publisher": "Ecopetrol", "source_type": "Operator annual report", "source_title": "Ecopetrol Integrated Management Report 2025", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "status": "CADE described in use; Rubiales dispatch-model MVP under construction in 2025", "summary": "Ecopetrol’s annual report describes CADE consolidating live operating information and supporting electricity-supply decisions with advanced analytics and generative AI.", "mechanism": "A Rubiales dispatch-model proof of concept was developed in the first half of 2025. Work on an MVP began in the second half, using electrical signals to generate operating setpoints. That model’s construction is separate from the broader CADE platform already described in use.", "measure": "No measured energy, emissions, reliability, cost or production benefit is disclosed for these AI components.", "operating_period": "2025 operating and development activity, described in the 2026 annual-report edition. Exact publication day and MVP commissioning date are not established.", "limit": "The platform’s operating status does not establish live use of the Rubiales MVP. The name “autonomous” does not establish AI control authority. Model, approval and override details are not given.", "question": "Are Rubiales recommendations now used, who approves a setpoint, and how is the result compared with the previous dispatch method?", "related_ids": ["AI-148"], "sources": [{"title": "Ecopetrol Integrated Management Report 2025", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 (2025 annual report)", "publisher": "Ecopetrol", "locator": "PDF page 90 / printed page 89: Centro Autónomo de Despacho Energético (CADE)", "access": "2025 operating and development activity, described in the 2026 annual-report edition. Exact publication day and MVP commissioning date are not established. Original-source review is not independent validation."}]}, {"id": "O14C-PETRONAS", "record": "AI-143", "company": "PETRONAS", "title": "Deployed exploration tools and a separate equipment-development agreement", "department": "Exploration & geoscience", "stage": "announced", "grade": "B", "attribution": "operator", "date": "2026-07-08", "source_id": "OCT14C-PETRONAS-TRICIPTA", "publisher": "PETRONAS", "source_type": "Company release", "source_title": "PETRONAS advances TriCipta AI through the third joint development agreement", "url": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "locator": "8 July 2026 release: third JDA scope and final two paragraphs on earlier tools", "status": "Earlier exploration tools deployed; new surface-equipment work remains development", "summary": "PETRONAS says AI.SEEK and Global Exploration Basin Screening & Analysis (GEB) have been deployed. The third TriCipta agreement develops a different set of surface-equipment tools.", "mechanism": "The earlier tools unify geoscience information for geospatial analysis. IBM Malaysia and Tridiagonal.AI are partners in the new surface-equipment optimization work, aimed at production and maintenance decisions.", "measure": "No measured exploration, production, maintenance, time or financial result is supplied. Deployment is a stage disclosure, not a quantified benefit.", "operating_period": "Company release dated 8 July 2026. Earlier exploration tools came from 2025 agreements; the third agreement was newly signed.", "limit": "This is already mentioned in AI-143; the update makes the split clearer. Do not transfer the earlier tools’ deployment status to the new equipment workflow or count another project from the same announcement. Asset, user and model-validation populations are undisclosed.", "question": "What do teams use AI.SEEK and GEB to decide, and which surface-equipment tools have since reached live use?", "related_ids": ["AI-143"], "sources": [{"title": "PETRONAS advances TriCipta AI through the third joint development agreement", "url": "https://www.petronas.com/media/media-releases/petronas-advances-tricipta-ai-through-third-joint-development-agreement-ibm", "date": "2026-07-08", "publisher": "PETRONAS", "locator": "8 July 2026 release: third JDA scope and final two paragraphs on earlier tools", "access": "Company release dated 8 July 2026. Earlier exploration tools came from 2025 agreements; the third agreement was newly signed. Original-source review is not independent validation."}]}, {"id": "O14C-CHORD", "record": "AI-097", "company": "Chord Energy", "title": "Rod-lift AI sits inside a broader artificial-lift improvement program", "department": "Production & field operations", "date": "2026-02-25", "source_id": "OCT14C-CHORD-LIFT", "publisher": "Chord Energy", "source_type": "Operator earnings release", "source_title": "Chord fourth-quarter and full-year 2025 results", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "locator": "Operations Update / Production and LOE; two consecutive artificial-lift paragraphs", "status": "Existing scaled AI workflow; broader operating results kept separate", "summary": "Chord reports ML on about 99% of rod-lift wells and an early roughly 25% improvement in rod-pump run times. Those figures were already in the register.", "mechanism": "The release places rod-lift ML alongside a wider artificial-lift optimization effort. It does not describe the model, vendor or approval/control interface.", "measure": "The broader program reports over 50% improvement in ESP replacement cycle times, over 25% improvement in failure rates since early 2025, and about 1,200 fewer workover rig days in 2025. These are not separately isolated ML effects.", "operating_period": "2025 results, disclosed 25 February 2026. ESP comparisons start at the beginning of 2025; an exact matched AI evaluation period is not supplied.", "limit": "Keep the rod-lift result separate from ESP and workover outcomes. Do not attribute the whole continuous-improvement program or its free-cash-flow benefit to AI. No additional result profile or cash ROI is counted.", "question": "What part of fewer failures and workovers comes from ML rather than equipment changes, maintenance and field practices?", "related_ids": ["AI-097"], "sources": [{"title": "Chord fourth-quarter and full-year 2025 results", "url": "https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results,-Issues-2026-Outlook-and-Declares-Base-Dividend", "date": "2026-02-25", "publisher": "Chord Energy", "locator": "Operations Update / Production and LOE; two consecutive artificial-lift paragraphs", "access": "2025 results, disclosed 25 February 2026. ESP comparisons start at the beginning of 2025; an exact matched AI evaluation period is not supplied. Original-source review is not independent validation."}]}, {"id": "O14C-TOTAL", "record": null, "company": "TotalEnergies", "title": "A 15-year agreement for a shared well-planning environment", "department": "Drilling & completions", "date": "2026-10-08", "publisher": "SLB", "source_title": "SLB awarded digital drilling contract by TotalEnergies", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-awarded-digital-drilling-contract-totalenergies", "locator": "8 October 2026 release; agreement description and Key Points", "status": "New digital-planning agreement; no AI deployment established", "summary": "SLB announced a 15-year DrillPlan agreement for TotalEnergies’s global upstream portfolio. The work connects subsurface information with well design and engineering.", "mechanism": "Offset-well data, trajectories, casing, completion inputs and planning assumptions are brought into a common environment while retaining customer workflows and internal digital expertise.", "measure": "No rollout population, achieved operating benefit, contract value or return is disclosed.", "operating_period": "Agreement announced 8 October 2026; implementation and completion dates are not given.", "limit": "The release does not identify an AI component. It belongs in operating-technology context, not the AI-adoption count. A 15-year term does not prove the whole portfolio is already live.", "question": "Does a common planning environment make it easier to compare alternatives and preserve the assumptions behind a well decision?", "related_ids": [], "sources": [{"title": "SLB awarded digital drilling contract by TotalEnergies", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-awarded-digital-drilling-contract-totalenergies", "date": "2026-10-08", "publisher": "SLB", "locator": "8 October 2026 release; agreement description and Key Points", "access": "Agreement announced 8 October 2026; implementation and completion dates are not given. Original-source review is not independent validation."}]}, {"id": "O14C-FORECAST", "record": null, "company": "Petroleum Research", "title": "A more complicated model did not improve these forecasts", "department": "Technical research", "date": "2026-01-21", "publisher": "Petroleum Research / Elsevier", "source_title": "Full-field petroleum production forecasting: Is deep learning all you need?", "url": "https://www.sciencedirect.com/science/article/pii/S2096249526000074", "locator": "Publisher abstract; online publication date 21 January 2026", "status": "Comparative study; not a named deployment failure", "report_id": "GAPE004", "summary": "In a comparison of four production datasets and three forecast horizons, kernel-based models outperformed neural architectures. More complexity did not deliver an accuracy advantage in these tests.", "mechanism": "The study compares ML proxies for field-scale production under operating scenarios. Its abstract also reports that spatial data coverage appeared more important for generalization than dataset size.", "measure": "No company-wide deployment result, cancellation rate or project cash return is established.", "operating_period": "Published online 21 January 2026. We reviewed the publisher’s indexed abstract; direct full-article access was unavailable.", "limit": "The finding applies to the tested inputs, datasets and horizons. It does not establish that deep learning is ineffective across petroleum tasks. This is technical benchmarking, not a failed operator rollout.", "question": "Does the proposed model outperform a simpler baseline on the operating decisions and conditions that matter?", "related_ids": [], "sources": [{"title": "Full-field petroleum production forecasting: Is deep learning all you need?", "url": "https://www.sciencedirect.com/science/article/pii/S2096249526000074", "date": "2026-01-21", "publisher": "Petroleum Research / Elsevier", "locator": "Publisher abstract; online publication date 21 January 2026", "access": "Published online 21 January 2026. We reviewed the publisher’s indexed abstract; direct full-article access was unavailable. Original-source review is not independent validation."}]}, {"id": "O14C-SURFACE", "record": null, "company": "Energy and AI", "title": "A good prediction is only part of a reliable control system", "department": "Technical research", "date": "2026-09", "publisher": "Energy and AI / Elsevier", "source_title": "AI and ML for production prediction and optimization of oil and gas surface networks", "url": "https://www.sciencedirect.com/science/article/pii/S2666546826000984", "locator": "Introduction and review method; sections 4.1, 4.2 and conclusions", "status": "Literature review; not a field incident or failure-rate study", "report_id": "O14C-SURFACE", "summary": "A surface-network review finds gaps in uncertainty, governance, transfer across fields and evidence for safe closed-loop use beyond pilots. Sensor drift and rare operating conditions can undermine an otherwise accurate model.", "mechanism": "The authors reviewed prediction, monitoring and control literature from January 2015 through March 2026. They argue for physically defensible recommendations, oversight and protected actuation.", "measure": "This review does not measure incident frequency, failure prevalence or the performance of a named operator’s deployed system.", "operating_period": "September 2026 journal edition; literature search through March 2026. Relevant publisher text was available through the web index.", "limit": "Literature synthesis is not an independent field audit. Its gaps do not show that a particular company’s controls are unsafe or that every deployment remains a pilot.", "question": "What detects drift, handles unfamiliar conditions and prevents a recommendation from becoming an unsafe action?", "related_ids": [], "sources": [{"title": "AI and ML for production prediction and optimization of oil and gas surface networks", "url": "https://www.sciencedirect.com/science/article/pii/S2666546826000984", "date": "2026-09", "publisher": "Energy and AI / Elsevier", "locator": "Introduction and review method; sections 4.1, 4.2 and conclusions", "access": "September 2026 journal edition; literature search through March 2026. Relevant publisher text was available through the web index. Original-source review is not independent validation."}]}], "method": "Three research tracks and editor review of original disclosures, the existing register and source access. Bounded research, not complete global coverage.", "count_note": "One new Aramco workflow; four existing profiles clarified, including supplier-reported Exxon results already in the history; one digital-planning note and two technical publications. No named failed deployment was verified in this pass."}, "watch_2026_10_14_pm": {"as_of": "2026-10-14", "run": "2026-10-14-pm", "checked_at_utc": "2026-10-15", "findings": [{"id": "W14PM-INPEX", "record": "AI-294", "company": "INPEX", "title": "Employees build agents for research, rules and management meetings", "department": "IT, data & enterprise transformation", "stage": "scaled", "grade": "C", "attribution": "vendor", "date": "2026-09-01", "source_id": "WATCH14PM-INPEX", "publisher": "Microsoft / named INPEX speakers", "source_type": "Vendor customer story", "source_title": "INPEX transforms business processes with Microsoft 365 Copilot and AI agents", "url": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "locator": "Research tool; Building a foundation for change; Looking beyond estimated annual value", "status": "Shared employee agents in use; active-agent population unknown", "summary": "INPEX uses Copilot for research and shared agents for company rules, regulatory searches and contract knowledge. A management-meeting agent checks uploaded materials and suggests questions.", "mechanism": "SharePoint, OneDrive, Office and Teams provide work context. Employees can create simpler agents; advanced Studio agents using external connectors or triggers require a cost-and-access approval process. Training videos and departmental ambassadors support use.", "measure": "Microsoft reports 1,450 paid licenses, over 1,000 agents and 130 research requests, most completed the same day versus previously weeks. A survey found 44% perceived less task time. INPEX’s estimated annual value of over JPY2bn covers company AI use; it is not Copilot-only cash savings.", "operating_period": "Published 1 September 2026; Copilot became the core platform in 2025. Exact measurement windows, active-agent count, survey size and value calculation are not supplied.", "limit": "Vendor story with named customer speakers. Report throughput does not establish report accuracy. License/agent counts do not measure sustained use. The company-wide value estimate may overlap other AI work, including Ichthys, and has no disclosed net-return method.", "question": "Which agents do people keep using, and does better research or meeting preparation change a decision?", "related_ids": ["AI-294", "AI-240"], "sources": [{"title": "INPEX transforms business processes with Microsoft 365 Copilot and AI agents", "url": "https://www.microsoft.com/en/customers/story/27096-inpex-corporation-microsoft-365-copilot", "date": "2026-09-01", "publisher": "Microsoft / named INPEX speakers", "locator": "Research tool; Building a foundation for change; Looking beyond estimated annual value", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-WILLIAMS", "record": "AI-224", "company": "Williams", "title": "A field technician turns equipment reports into turbine alerts", "department": "IT, data & enterprise transformation", "stage": "scaled", "grade": "C", "attribution": "vendor", "date": "2026-09-09", "source_id": "WATCH14PM-WILLIAMS", "publisher": "Microsoft / named Williams speakers", "source_type": "Vendor customer story", "source_title": "Williams transforms frontline work with Copilot, gaining 90% adoption", "url": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "locator": "Bringing AI to the frontline first; Accelerating operations with Copilot", "status": "Employee-built alert workflow described in use", "summary": "A Williams technician built an anomaly detector that reads emailed equipment reports, compares them with operating baselines and flags changes in temperature, oil pressure or vibration.", "mechanism": "The Microsoft story names Copilot Cowork and Outlook distribution-list reports. Staff also create agents for procedures and safety paperwork. The technician receives alerts and can investigate; automatic equipment control is not described.", "measure": "More than 90% organizational AI adoption is repeated from earlier coverage. No detector-specific uptime, false-alert, safety or financial result is supplied.", "operating_period": "Published 9 September 2026; the existing profile retains Williams’s June company-wide adoption disclosure. The detector’s start date and asset count are unspecified.", "limit": "Midstream operating example, not an upstream production result. Vendor account without disclosed detector validation, alert threshold, escalation protocol or before/after equipment comparison. Organization-wide adoption is not this detector’s user share.", "question": "Do alerts catch faults earlier than normal monitoring, and how many alerts require unnecessary investigation?", "related_ids": ["AI-224"], "sources": [{"title": "Williams transforms frontline work with Copilot, gaining 90% adoption", "url": "https://www.microsoft.com/en/customers/story/27209-williams-microsoft-365-copilot", "date": "2026-09-09", "publisher": "Microsoft / named Williams speakers", "locator": "Bringing AI to the frontline first; Accelerating operations with Copilot", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-NOV", "record": "AI-158", "company": "NOV", "title": "Egyptian drilling teams expand an edge-based warning system", "department": "Drilling & completions", "stage": "operating", "grade": "C", "attribution": "vendor", "date": "2026-03-11", "source_id": "WATCH14PM-NOV", "publisher": "NOV", "source_type": "Vendor deployment article", "source_title": "NOV’s Drilling Beliefs & Analytics advances digital operations in Egypt", "url": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "locator": "11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example", "status": "Vendor reports expansion from two rigs to a double-digit fleet", "summary": "NOV says Drilling Beliefs & Analytics expanded from a two-rig trial to double-digit rigs in Egypt’s Western Desert. An RTOC engineer monitors several rigs, with crews acting on warnings.", "mechanism": "Bayesian belief indices use drilling parameters to flag dysfunction. Edge processing supports use with limited connectivity. An overpull warning led staff to correct drill-off procedures; a separate autozero warning led a crew to reset bit weight.", "measure": "NOV estimates $120,000 avoided in the stuck-pipe case. A separate autozero case reports 8.2 hours and about $26,000 saved. These are distinct from the earlier $75,000 bit-balling estimate in AI-158; they are not added together.", "operating_period": "Article published 11 March 2026. Egypt introduction dates to 2023; the source does not date each intervention or disclose exact rig/well counts.", "limit": "Operator names remain anonymous. Event-level vendor estimates lack an independent counterfactual or full implementation costs. No autonomous control or fleet-wide cash ROI is established.", "question": "How often are warnings actionable, and does the fleet show lower downtime after accounting for differences between wells?", "related_ids": ["AI-158"], "sources": [{"title": "NOV’s Drilling Beliefs & Analytics advances digital operations in Egypt", "url": "https://www.nov.com/news/novs-drilling-beliefs-and-analytics-advances-digital-operations-in-egypt", "date": "2026-03-11", "publisher": "NOV", "locator": "11 March 2026 article: trial expansion, stuck-pipe example and Bit Weight Autozero example", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-ENI", "record": null, "company": "Eni", "title": "Closed-loop drilling and pressure management on one Indonesian well", "department": "Drilling & completions", "date": "2026-07-09", "publisher": "Halliburton", "source_title": "Eni and Halliburton achieve industry first with closed-loop rig automation in deepwater Indonesia", "url": "https://www.halliburton.com/en/about-us/press-release/eni-halliburton-achieve-industry-first-closed-loop-rig-automation-indonesia", "locator": "9 July 2026 release: deployment, integrated controls and reported efficiency", "status": "Operating automation; AI component not specified", "summary": "Halliburton describes LOGIX coordinating drilling, tripping and managed-pressure drilling on an Eni deepwater exploration well offshore Indonesia.", "mechanism": "Rig equipment, automated well placement, downhole hydraulics and pressure controls operate in a connected closed loop. Human approval, handoff and safety-interlock arrangements are not detailed.", "measure": "Supplier reports over 15% better efficiency, without defining the efficiency measure, baseline or absolute time/cost effect.", "operating_period": "9 July 2026 announcement; one exploration well supports the reported result, not a fleet-wide rollout.", "limit": "The release does not identify the AI/ML elements. Retained as operating technology, outside the AI-adoption and result counts.", "question": "Which decisions are automated, and how much useful well-delivery time changes under comparable pressure conditions?", "related_ids": [], "sources": [{"title": "Eni and Halliburton achieve industry first with closed-loop rig automation in deepwater Indonesia", "url": "https://www.halliburton.com/en/about-us/press-release/eni-halliburton-achieve-industry-first-closed-loop-rig-automation-indonesia", "date": "2026-07-09", "publisher": "Halliburton", "locator": "9 July 2026 release: deployment, integrated controls and reported efficiency", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-ENCAP", "record": null, "company": "EnCap / Quantica", "title": "An energy investor backs sites, power and connectivity for AI demand", "department": "Finance, investment & strategy", "date": "2025-07-28", "publisher": "EnCap Investments", "source_title": "EnCap launches Quantica Infrastructure", "url": "https://www.encapinvestments.com/news/encap-investments-lp-launches-quantica-infrastructure-deliver-integrated-ai-and-hyperscale", "locator": "28 July 2025 launch release; platform scope and existing pipeline", "status": "Infrastructure platform launch; not operator AI adoption", "summary": "EnCap’s Energy Transition arm backed Quantica to combine land, renewable and grid power, and network connectivity for data-center sites.", "mechanism": "Integrated site development serves hyperscale and AI-compute customers. The source describes an existing project pipeline without quantifying it.", "measure": "No completed-project, deployed-load, AI productivity or financial result is reported. Leadership’s prior 15GW experience is not Quantica’s delivered capacity.", "operating_period": "Platform launch announced 28 July 2025; rollout and commissioning dates are unspecified.", "limit": "AI is a customer-demand driver here, not a technology shown improving a portfolio operator’s work. Keep this capital-allocation evidence separate from AI adoption.", "question": "Which sites have secured power and customers, and how much announced capacity becomes operating load?", "related_ids": [], "sources": [{"title": "EnCap launches Quantica Infrastructure", "url": "https://www.encapinvestments.com/news/encap-investments-lp-launches-quantica-infrastructure-deliver-integrated-ai-and-hyperscale", "date": "2025-07-28", "publisher": "EnCap Investments", "locator": "28 July 2025 launch release; platform scope and existing pipeline", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-NGP", "record": null, "company": "NGP / Cloverleaf", "title": "Powered-site development is a separate kind of AI exposure", "department": "Finance, investment & strategy", "date": "Undated", "publisher": "NGP Energy Capital", "source_title": "NGP portfolio profile: Cloverleaf Infrastructure", "url": "https://ngpenergy.com/", "locator": "Portfolio Profile / Cloverleaf Infrastructure", "status": "Sponsor reports development and sales; commissioning not established", "summary": "NGP identifies Cloverleaf as a portfolio company since 2024, developing powered sites for large data centers. Its current profile reports over 5GW developed and sold.", "mechanism": "Site selection, power access and community work support data-center customers. The CEO describes AI demand, market saturation and local collaboration as business conditions.", "measure": "Over 5GW developed and sold over the last two years, according to the undated sponsor page. This is not commissioned capacity, active AI load or an AI-use result.", "operating_period": "Undated page reviewed 14 October 2026. Its relative two-year period cannot be anchored to a known publication date.", "limit": "No site/customer count, utilization, sale economics or AI-specific revenue allocation. The sponsor relationship does not establish AI adoption by NGP’s upstream holdings.", "question": "What distinguishes a sold development project from an energized data center serving customers?", "related_ids": [], "sources": [{"title": "NGP portfolio profile: Cloverleaf Infrastructure", "url": "https://ngpenergy.com/", "date": "Undated", "publisher": "NGP Energy Capital", "locator": "Portfolio Profile / Cloverleaf Infrastructure", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-BCG", "record": null, "company": "BCG", "title": "Plans for agent authority are ahead of reported controls", "department": "Industry research", "date": "2026-09-30", "publisher": "Boston Consulting Group", "source_title": "The Formula for Agentic AI Value / Applied AI Index 2026", "url": "https://www.bcg.com/publications/2026/the-formula-for-agentic-ai-value", "locator": "Key takeaways; study scope; FAQ on agent autonomy", "status": "Cross-sector survey; not an oil-and-gas adoption rate", "report_id": "WATCH14PM-BCG", "summary": "BCG’s survey finds that 42% expect agents to have decision authority by 2030, while 5% report having all the relevant controls today.", "mechanism": "The article covers over 1,300 senior leaders across more than 20 sectors; its companion release specifies 1,330. It argues for organizational ownership, governed actions and business-outcome measurement.", "measure": "42% is a future intention; 5% is a present control-readiness claim. They are not the same measure or an estimate of unsafe deployments.", "operating_period": "Published 30 September 2026. Survey fieldwork dates and an oil-and-gas denominator were not established in the reviewed material.", "limit": "Self-reported cross-sector evidence and maturity classifications, not causal proof or audited project ROI. The reviewed article highlights power/utilities and chemicals but does not provide an upstream-only comparison.", "question": "Before granting an agent authority, who owns its decision, what limits its actions, and how can a person reverse them?", "related_ids": [], "sources": [{"title": "The Formula for Agentic AI Value / Applied AI Index 2026", "url": "https://www.bcg.com/publications/2026/the-formula-for-agentic-ai-value", "date": "2026-09-30", "publisher": "Boston Consulting Group", "locator": "Key takeaways; study scope; FAQ on agent autonomy", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}, {"id": "W14PM-BX", "record": null, "company": "Blackstone", "title": "Portfolio CEOs report more deployment, without an energy breakdown", "department": "Industry research", "date": "2026-08-05", "publisher": "Blackstone", "source_title": "Pattern Recognition / AI Adoption Is Accelerating", "url": "https://www.blackstone.com/pattern-recognition/", "locator": "5 August 2026 section; 2Q26 / 2Q25 CEO Survey footnotes", "status": "Private-capital portfolio survey; cross-sector context", "report_id": "WATCH14PM-BX", "summary": "Blackstone says nearly half of surveyed portfolio CEOs report AI in production or fully deployed, up from under 30% a year earlier. Eighty percent plan to increase spending.", "mechanism": "The 2Q26 survey covered 117 portfolio companies, including 73 US CEOs, largely in private equity and credit. The 2Q25 comparison covered 87 companies; respondents changed.", "measure": "Deployment stage and planned spending are reported; productivity, cash returns and energy-company adoption are not measured here.", "operating_period": "Published 5 August 2026. The 2Q26 survey ran 9–30 June 2026; 2Q25 ran 5–19 June 2025.", "limit": "Sponsor-reported, nonrepresentative subset with changing composition and no energy subgroup. Blackstone’s own token consumption is a separate internal-use measure, not a portfolio operating result.", "question": "Which portfolio workflows are in use, and what changes in operating or financial results after deployment?", "related_ids": [], "sources": [{"title": "Pattern Recognition / AI Adoption Is Accelerating", "url": "https://www.blackstone.com/pattern-recognition/", "date": "2026-08-05", "publisher": "Blackstone", "locator": "5 August 2026 section; 2Q26 / 2Q25 CEO Survey footnotes", "access": "Relevant original source read during the 14 October evening review; source claims are not independently validated."}]}], "method": "Three bounded Luna tracks plus editor verification. New-to-this-report evidence, not necessarily news released today. No new named failed or abandoned deployment was substantiated.", "count_note": "INPEX’s existing company-brain research now has a full project profile; Williams and NOV profiles have more operating detail. Automation and AI-serving infrastructure remain separate context. Two survey publications added, and one duplicate technical paper merged into its established entry."}, "watch_2026_10_15_am": {"as_of": "2026-10-15", "run": "2026-10-15-am", "label": "15 October 2026 / Morning research review", "headline": "The useful unit of comparison is the operating workflow.", "intro": "CNOOC connects field data to diagnostics. Suncor optimizes an asset network. Petrobras automates purchasing. The work and measurement matter more than whether a company calls the system AI.", "count_note": "CNOOC’s existing profile gains a field-level example; Suncor and Petrobras gain separate operating-technology notes. No new project, result or ROI count. These are older disclosures surfaced or expanded today.", "method": "Three bounded Luna tracks followed by editor checks of original PDFs and full-corpus duplicates. No substantiated disclosure newer than the prior publication, or new named failed deployment, was found. This is not exhaustive coverage.", "findings": [{"record": "AI-145", "related_ids": ["AI-145"], "date": "2026-04-09", "publisher": "CNOOC Limited", "source_id": "WATCH15AM-CNOOC", "source_type": "Company publication", "id": "W15AM-CNOOC", "company": "CNOOC", "department": "Production & field operations", "title": "One oilfield connects reservoir, production and equipment data", "source_title": "2025 Environmental, Social and Governance Report", "url": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "locator": "Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026", "status": "Operating integrated digital field; AI contribution not isolated", "summary": "CNOOC says Qinhuangdao 32-6 in Bohai uses more than 26,000 data-acquisition points in one platform for reservoirs, injection, production and equipment. It diagnoses production abnormalities and generates optimization plans.", "mechanism": "More than 26,000 acquisition points feed the platform. The company places this field example within its broader AI+ and data-efficiency program. The field passage describes connected instrumentation, diagnostics and plan generation, but does not identify the model or how much of the improvement comes from AI. Automatic plan generation does not establish automatic execution or operator approval.", "measure": "CNOOC reports 30% less time for reservoir-potential studies, more than 90% early-warning accuracy for critical equipment, and more than RMB 10 million lower annual operating and maintenance costs. These are integrated-program results, not isolated AI benefits or net ROI.", "operating_period": "2025 reporting period; report released 9 April 2026. Start date, observation periods and deployed user count are not specified. More than 26,000 points measures instrumentation, not people or wells.", "limit": "Company-reported results. Study-time baseline, warning-accuracy denominator, cost calculation, implementation cost and independent validation are absent. No disclosed execution authority or safety boundary. This expands the already-named Qinhuangdao component of AI-145; it is not a second project or a new counted numerical-result profile. AI-082 remains separately unresolved.", "question": "Which plans do engineers accept, what do false alarms cost, and how much of the reported saving survives integration and support costs?", "excerpt": "This platform enables intelligent diagnosis of production abnormalities and automatic generation of optimisation solutions.", "sources": [{"title": "2025 Environmental, Social and Governance Report", "url": "https://www.cnoocltd.com/english/sustainability/environmentalsocialandgovernancereports/material/2025ESGen.pdf", "date": "2026-04-09", "publisher": "CNOOC Limited", "locator": "Printed page 23 / PDF page 28, Qinhuangdao 32-6 case; report index dates publication 9 April 2026", "access": "Parent retrieved original PDF and checked the relevant page on 15 October 2026; results remain source-reported."}]}, {"record": null, "related_ids": ["AI-154", "AI-099"], "date": "2026-08-05", "publisher": "Suncor Energy", "source_id": "WATCH15AM-SUNCOR", "source_type": "Official earnings-call transcript", "id": "W15AM-SUNCOR", "company": "Suncor", "department": "Production & field operations", "title": "A single model helps decide where the molecules should go", "source_title": "Second Quarter 2026 Financial Results Call", "url": "https://www.suncor.com/-/media/project/suncor/files/investor-centre/transcripts/2026-q2-suncor-energy-financial-results-call-transcript-en.pdf?created=20260818150055&modified=20260828195420", "locator": "PDF page 10 (page index 9), Q&A lines 331–358, Dave Oldreive and Rich Kruger; call dated 5 August 2026", "status": "Operating linear-programming model; management calls it AI", "summary": "Suncor describes one linear-programming model connecting its oil-sands operations, Edmonton refinery, domestic markets and global export sales. During an upset, teams can change the crude slate, move diluent and route intermediate streams.", "mechanism": "The model supports regional choices across connected assets and markets. Dave Oldreive calls it a big AI tool; the disclosed method is linear programming. No machine learning or LLM is established. Management describes integrated teams maximizing value, but does not specify who approves individual model recommendations.", "measure": "The stated 10,000–15,000 barrels/day is intermediate-stream routing capability. It is not incremental AI production, savings or measured uplift. Record throughput and sales are not assigned to the optimizer.", "operating_period": "Present-tense operating description on the 5 August 2026 call, discussing Q2 adjustments. This passage was already in our earnings-call ledger; it is now easier to find in the operating-technology view.", "limit": "No architecture beyond linear programming, counterfactual, model-specific benefit, implementation cost or detailed control boundary. Keep it outside new AI-adoption and result counts. Suncor haulage and truck dispatch remain separate existing records.", "question": "Can a team optimize the whole asset network rather than improving one unit at the expense of another?", "excerpt": "We run a single, what we call linear programming model to optimize that.", "sources": [{"title": "Second Quarter 2026 Financial Results Call", "url": "https://www.suncor.com/-/media/project/suncor/files/investor-centre/transcripts/2026-q2-suncor-energy-financial-results-call-transcript-en.pdf?created=20260818150055&modified=20260828195420", "date": "2026-08-05", "publisher": "Suncor Energy", "locator": "PDF page 10 (page index 9), Q&A lines 331–358, Dave Oldreive and Rich Kruger; call dated 5 August 2026", "access": "Parent read the relevant original issuer transcript passage on 15 October 2026; results remain source-reported."}]}, {"record": null, "related_ids": ["AI-146", "AI-219", "AI-220", "AI-221"], "date": "2025-11-27", "publisher": "Petrobras", "source_id": "WATCH15AM-PETROBRAS", "source_type": "Investor presentation", "id": "W15AM-PETROBRAS", "company": "Petrobras", "department": "Supply chain & procurement", "title": "Purchasing automation has its own value case", "source_title": "2026–2030 Business Plan", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "locator": "Printed slide 158 / PDF page 158, purchase-request automation box; separate from the AI and digital-twin boxes", "status": "Reported purchasing automation; AI component not established", "summary": "Petrobras reports automating purchase requests to speed up purchasing and avoid costs. The same slide also describes AI projects, but this purchasing box does not identify AI.", "mechanism": "The slide names request automation without explaining the software, process steps, business-unit coverage, approval rules or exception handling. We retain it as purchasing automation, separate from the neighboring well, forecasting, flare and control-room AI examples.", "measure": "Petrobras reports 190,000 items automated per year and average avoided costs of R$111 million per year. Avoided cost is not the same as cash collected or net savings after implementation and support costs.", "operating_period": "Business plan published 27 November 2025. The slide gives annual figures without a measurement period, start date or rollout denominator.", "limit": "Company-reported values without a disclosed baseline, cost method, implementation cost or independent validation. The slide does not establish whether the estimate is realized or forecast. Do not count this as AI adoption or add it to the separate AI benefits.", "question": "Which purchasing steps disappear, which approvals still need people, and how is avoided cost calculated?", "excerpt": "Automation of requests, speeding up purchases and avoiding costs.", "sources": [{"title": "2026–2030 Business Plan", "url": "https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0", "date": "2025-11-27", "publisher": "Petrobras", "locator": "Printed slide 158 / PDF page 158, purchase-request automation box; separate from the AI and digital-twin boxes", "access": "Parent retrieved original PDF and checked the relevant page on 15 October 2026; results remain source-reported."}]}]}, "watch_2026_10_15_extra": {"as_of": "2026-10-15", "run": "2026-10-15-extra", "label": "15 October 2026 / Follow-up research", "headline": "What does the engineer still have to do?", "intro": "An assistant that helps with analysis, a tested interface to reservoir systems, and a controller that executes checked changes do different jobs. These examples make that distinction easier to see.", "count_note": "Four source-linked notes added. GEORGE remains reported use pending technical corroboration; the Petrobras assistant is research, Exxon’s ESP pilot is physics-based control, and Aethon is transformation context. No new adoption, numerical-result or ROI count.", "method": "Three bounded Luna tracks plus editor source checks and full-corpus deduplication. No post-cutoff news or new named failed deployment was substantiated. Older disclosures and access limits are labeled.", "archive_label": "More company context and earlier research", "findings": [{"record": null, "related_ids": [], "department": "Production & field operations", "id": "W15X-GEORGE", "company": "Chevron", "title": "GEORGE helps production engineers work across thousands of wells", "source_id": "WATCH15X-GEORGE", "date": "2026-09-04", "publisher": "SPE / Journal of Petroleum Technology", "source_type": "Original conference reporting of named operator remarks", "source_title": "Maximizing Every Barrel: Artificial Lift Embraces Automation", "url": "https://jpt.spe.org/maximizing-every-barrel-artificial-lift-embraces-automation", "locator": "Chevron section: Peter Nesom remarks following the image caption; 25 August 2026 plenary, reported 4 September", "status": "Reported company use; issuer or technical corroboration still needed", "summary": "SPE reports that Chevron’s Peter Nesom described GEORGE, a production-engineering assistant used for analysis, optimization and automation tasks on more than 5,000 wells.", "mechanism": "The report describes an assistant working with production engineers across several problems. It does not specify its models, tools, data connections, approval steps or control-system permissions.", "measure": "More than 5,000 wells is reported use scope. No production uplift, time saving, cost or net return is supplied.", "operating_period": "Remarks at the 25 August 2026 plenary; SPE report published 4 September. Launch date and active-well denominator unknown.", "limit": "Original conference reporting of named operator remarks, not an operator-authored deployment paper. The event program corroborates speaker and session only. Keep as reported use awaiting technical corroboration, outside additional adoption/result counts. Do not infer an LLM architecture, autonomous field control or a company-wide knowledge system.", "question": "Which engineering tasks does GEORGE complete, what must engineers approve, and what changes in field performance?", "excerpt": "used for over 5,000 wells", "sources": [{"title": "Maximizing Every Barrel: Artificial Lift Embraces Automation", "url": "https://jpt.spe.org/maximizing-every-barrel-artificial-lift-embraces-automation", "date": "2026-09-04", "publisher": "SPE / Journal of Petroleum Technology", "locator": "Chevron section: Peter Nesom remarks following the image caption; 25 August 2026 plenary, reported 4 September", "access": "Original SPE/JPT report read by editor 15 October 2026; named operator remarks reported by the publisher, not issuer technical validation."}, {"title": "SPE Artificial Lift 2026 opening plenary program", "url": "https://www.spe-events.org/artificiallift/attend/special-sessions/opening-plenary-session", "date": "2026-08-25 (event date; page publication date unknown)", "publisher": "Society of Petroleum Engineers", "locator": "Opening plenary: Peter Nesom, Chevron Global Upstream Production Engineering CoE", "access": "Official event page read; corroborates speaker identity and session, not GEORGE scope or performance."}]}, {"record": null, "related_ids": [], "department": "Reservoir engineering & field development", "id": "W15X-PETROBRAS", "company": "Petrobras", "title": "A reservoir assistant is tested as an interface to technical systems", "source_id": "WATCH15X-PETROBRAS", "date": "2025-10-21", "publisher": "Petrobras and Tecgraf / PUC-Rio authors; Offshore Technology Conference", "source_type": "Operator-coauthored conference abstract", "source_title": "Virtual Assistant Using Generative Pre-Trained Transformers on Reservoir Management Context", "url": "https://onepetro.org/OTCBRASIL/proceedings-abstract/25OTCB/25OTCB/D011S005R007/792383", "locator": "OTC-36185-MS; publisher-indexed author list and abstract, test-scenario paragraphs", "status": "Research prototype and test scenarios; operating rollout not established", "summary": "A Petrobras-coauthored paper proposes a GPT-based assistant for reservoir-management systems, intended to let engineers use natural-language requests instead of navigating technical interfaces.", "mechanism": "The abstract describes model training and tests of domain-specific requests, including attempts to avoid sensitive or irrelevant responses. It does not identify the model provider, connected applications, tool permissions or production deployment.", "measure": "The authors report favorable test behavior and training loss approaching zero. No held-out task-success denominator, active users, field result or economics is disclosed. Training loss is not proof of reliable operating decisions.", "operating_period": "Published 21 October 2025, OTC Brasil. Test dates and subsequent rollout unknown.", "limit": "Publisher-indexed original abstract reviewed; full paper not retrieved. Retained as research, outside operating adoption and outcome counts. No evidence of permission to change reservoir plans or execute downstream actions.", "question": "Does it perform useful tasks reliably on unseen requests, preserve access rights and leave an auditable record?", "excerpt": "this work proposes an intelligent virtual assistant specialized in Reservoir Management systems.", "sources": [{"title": "Virtual Assistant Using Generative Pre-Trained Transformers on Reservoir Management Context", "url": "https://onepetro.org/OTCBRASIL/proceedings-abstract/25OTCB/25OTCB/D011S005R007/792383", "date": "2025-10-21", "publisher": "Petrobras and Tecgraf / PUC-Rio authors; Offshore Technology Conference", "locator": "OTC-36185-MS; publisher-indexed author list and abstract, test-scenario paragraphs", "access": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved."}]}, {"record": null, "related_ids": ["AI-235"], "department": "Production & field operations", "id": "W15X-EXXON-ESP", "company": "ExxonMobil", "title": "ESP recommendations pass safety checks before reaching the controller", "source_id": "WATCH15X-EXXON-ESP", "date": "2025-08-18", "publisher": "ExxonMobil authors / Society of Petroleum Engineers", "source_type": "Operator-authored conference abstract", "source_title": "Closed-Loop ESP Optimization to Maximize Asset Value", "url": "https://onepetro.org/SPEESP/proceedings-abstract/25ESP/25ESP/D031S006R001/788064", "locator": "SPE-225261-MS; abstract, system checks and 26-well pilot paragraphs", "status": "26-well pilot in 2025; broader 2026 rollout reported separately", "summary": "ExxonMobil authors describe a closed-loop ESP pilot on 26 Midland wells. Commercial physics-based software recommends changes; internal checks test facility constraints and pump limits before execution.", "mechanism": "Recommendations that pass the checks can be written automatically to field devices from different controller makers, without a person approving every change. The abstract does not identify machine learning; an AI keyword does not establish the deployed method.", "measure": "The authors describe less surveillance and execution work, without numerical time, production or financial results. No complete cost or net return is supplied.", "operating_period": "Paper published 18 August 2025; pilot dates unspecified. A September 2026 conference report describes wider Permian rollout underway, without a completed-fleet count.", "limit": "Original publisher-indexed abstract reviewed; full paper not retrieved. It establishes physics-based control and a validation gate, not an LLM controlling pumps. This is separate from AI-235’s gas-lift ML and its oil-uplift figures.", "question": "What happens when data is stale, a recommendation fails a check, or operating conditions move outside the validated range?", "excerpt": "The closed-loop optimization system was piloted on 26 ESP wells in our Midland assets.", "sources": [{"title": "Closed-Loop ESP Optimization to Maximize Asset Value", "url": "https://onepetro.org/SPEESP/proceedings-abstract/25ESP/25ESP/D031S006R001/788064", "date": "2025-08-18", "publisher": "ExxonMobil authors / Society of Petroleum Engineers", "locator": "SPE-225261-MS; abstract, system checks and 26-well pilot paragraphs", "access": "Original publisher-indexed abstract read by editor 15 October 2026; direct page retrieval failed and the full paper was not retrieved."}, {"title": "Maximizing Every Barrel: Artificial Lift Embraces Automation", "url": "https://jpt.spe.org/maximizing-every-barrel-artificial-lift-embraces-automation", "date": "2026-09-04", "publisher": "SPE / Journal of Petroleum Technology", "locator": "ExxonMobil / Mark Agnew remarks, closed-loop ESP rollout and facility constraints paragraphs", "access": "Original SPE/JPT conference report read. Later rollout is reported underway; no completed-fleet count or quantified ESP benefit. Separate from the 2025 pilot abstract."}]}, {"record": null, "related_ids": [], "department": "IT, data & enterprise transformation", "id": "W15X-AETHON", "company": "Aethon Energy", "title": "Aethon’s CIO describes the foundations for better data use", "source_id": "WATCH15X-AETHON", "date": "2026-02-20 to 2026-02-26", "publisher": "Dallas Business Journal / DallasCIO advertising supplement", "source_type": "Named executive profile in advertising supplement", "source_title": "Dallas CIO ORBIE special advertising supplement", "url": "https://media.bizj.us/view/img/13143357/low-res-dallasorbiesciospecial-section2026-8951090d-b07f-4f22-bfba-90c5a075a999.pdf", "locator": "PDF page 10 / printed page 10, Kang Chen profile; supplement dated 20–26 February 2026", "status": "Executive-reported modernization; no AI deployment disclosed", "summary": "Aethon CIO Kang Chen reports a full cloud migration and technology overhaul, along with new data-quality, analytics and workforce-training initiatives.", "mechanism": "The profile also describes cybersecurity benchmarking. It names no cloud provider, data domains, field applications or AI model.", "measure": "No quantitative operating or financial result is supplied.", "operating_period": "20–26 February 2026 issue. Individual project dates and completion of the data initiatives are unknown.", "limit": "Named executive self-report in an advertising supplement, not independent technical validation. Cloud migration and upskilling do not prove AI readiness or deployment. Keep as company transformation context, outside AI adoption and outcome counts.", "question": "Which field and business data can teams actually reuse, and what applications have they put into daily work?", "excerpt": "fully migrating all IT systems to the cloud.", "sources": [{"title": "Dallas CIO ORBIE special advertising supplement", "url": "https://media.bizj.us/view/img/13143357/low-res-dallasorbiesciospecial-section2026-8951090d-b07f-4f22-bfba-90c5a075a999.pdf", "date": "2026-02-20 to 2026-02-26", "publisher": "Dallas Business Journal / DallasCIO advertising supplement", "locator": "PDF page 10 / printed page 10, Kang Chen profile; supplement dated 20–26 February 2026", "access": "Original advertising-supplement PDF text read by editor 15 October 2026; named executive self-report, not independent validation."}]}]}, "watch_2026_10_15_pm": {"as_of": "2026-10-15", "run": "2026-10-15-pm", "label": "15 October 2026 / Evening research", "headline": "What changes when the model meets the operation?", "intro": "Connecting a model to field work takes more than a promising prediction. These sources show the next questions: what can it control, what context does it need, and who can check its work?", "count_note": "Four source-linked notes and two publications added. Initial phases, planned pilots and research stay separate from operating adoption. Project, numerical-result and ROI counts are unchanged.", "method": "Three bounded Luna research tracks, followed by editor checks against original sources and full-corpus deduplication. These are newly found disclosures, not verified post-cutoff news. No new named failed deployment was substantiated; access and date limits remain visible.", "archive_label": "More company context and earlier research", "findings": [{"id": "W15PM-REPSOL", "record": null, "related_ids": ["AI-031", "AI-032"], "company": "Repsol", "title": "Repsol completes an initial phase of an autonomous glycol-plant optimizer", "date": "2026-02-19", "publisher": "Repsol", "source_type": "Original issuer annual report", "source_title": "2025 Group Annual Financial Report", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/ref/2026/ref19022026-2025-group-annual-financial-report.pdf", "locator": "PDF page 195 / printed Management Report page 45, section 5.5: AI-based autonomous optimizers", "status": "Initial phase completed; operating use not established", "summary": "Repsol says it completed the initial phase of its first Autonomous Optimizer with Control, developed internally for one section of a glycol plant.", "mechanism": "The report calls the optimizer AI-based. It does not identify the plant, model, controlled variables, operating limits, approval steps or fallback.", "measure": "No measured plant result, cost or net return is disclosed.", "operating_period": "2025 annual report, released 19 February 2026. Project and commissioning dates are not supplied.", "limit": "Completing an initial phase does not establish routine autonomous operation or a plant-wide rollout. Keep separate from Repsol’s multiagent, workforce and legal programs.", "question": "What did the initial phase demonstrate, and what must operators approve before it can control the process?", "department": "Production & field operations", "excerpt": "end of the initial phase of the first Autonomous Optimizer with Control (OAC)", "source_id": "WATCH15PM-REPSOL", "report_id": null, "sources": [{"title": "2025 Group Annual Financial Report", "url": "https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/ref/2026/ref19022026-2025-group-annual-financial-report.pdf", "date": "2026-02-19", "publisher": "Repsol", "locator": "PDF page 195 / printed Management Report page 45, section 5.5: AI-based autonomous optimizers", "access": "Editor retrieved the original PDF, read the relevant text and inspected the rendered page on 15 October 2026."}]}, {"id": "W15PM-CORVA-OLIA", "record": null, "related_ids": [], "company": "Corva / Oliasoft", "title": "Well designs and drilling data are being connected in both directions", "date": "2026-10-12", "publisher": "Corva and Oliasoft", "source_type": "Joint vendor announcement", "source_title": "Corva and Oliasoft Create a Continuous Learning Loop Between Well Design and Drilling", "url": "https://www.corva.ai/blog/corva-and-oliasoft-create-a-continuous-learning-loop-between-well-design-and-drilling", "locator": "Joint release: engineering-to-operations, operations-to-engineering, and customer-pilot paragraphs", "status": "Integration developed; customer pilots being prepared", "summary": "The vendors have developed an integration between Oliasoft WellDesign and Corva Real-Time AI and say they are preparing customer pilots.", "mechanism": "Design trajectories, casing and pressure assumptions can inform drilling workflows. As-drilled trajectories, operational data and nonproductive-time events return to planning and model calibration.", "measure": "No pilot customer, well or rig count, measured benefit, implementation cost or ROI is disclosed.", "operating_period": "Announced 12 October 2026; pilot start dates are unspecified. Newly found evidence, not a disclosure after the previous update.", "limit": "Model feedback is not evidence of automatic rig control. Two vendor versions of the same release do not provide independent validation; action permissions are unspecified.", "question": "Can the pilot show fewer repeat design errors or faster decisions, with engineers able to review the changes?", "department": "Drilling & completions", "excerpt": "preparing customer pilots", "source_id": "WATCH15PM-CORVA-OLIA", "report_id": null, "sources": [{"title": "Corva and Oliasoft Create a Continuous Learning Loop Between Well Design and Drilling", "url": "https://www.corva.ai/blog/corva-and-oliasoft-create-a-continuous-learning-loop-between-well-design-and-drilling", "date": "2026-10-12", "publisher": "Corva and Oliasoft", "locator": "Joint release: engineering-to-operations, operations-to-engineering, and customer-pilot paragraphs", "access": "Editor read both vendors’ original releases on 15 October 2026; neither supplies customer-pilot outcomes."}, {"title": "Corva and Oliasoft Create a Continuous Learning Loop Between Well Design and Drilling", "url": "https://oliasoft.com/articles/corva-and-oliasoft-create-a-continuous-learning-loop-between-well-design-and-drilling", "date": "2026-10-12", "publisher": "Oliasoft", "locator": "Joint release: engineering-to-operations, operations-to-engineering, and customer-pilot paragraphs", "access": "Original partner release read; same joint announcement, not independent validation."}]}, {"id": "W15PM-FORECAST", "record": null, "related_ids": [], "company": "UNICAMP / UNISIM", "title": "A forecast can look accurate and still miss the operating change that matters", "date": "2026-10 (issue month; exact online date unknown)", "publisher": "Geoenergy Science and Engineering / Elsevier", "source_type": "Technical journal article", "source_title": "Data-driven production forecast for oil and gas reservoirs - Advantages and limitations", "url": "https://www.sciencedirect.com/science/article/pii/S2949891026002137", "locator": "Volume 265, article 214569; DOI 10.1016/j.geoen.2026.214569; publisher abstract and indexed sections, no PDF page locator", "status": "Synthetic-reservoir benchmark; no operator deployment", "summary": "A UNISIM-IV study finds that production-history-only forecasts miss changes driven by operating events, interactions between wells and constraints.", "mechanism": "Its BiLSTM illustration omits operating context from the baseline. The authors emphasize input design and constraints over a more elaborate architecture.", "measure": "Aggregate forecast error can hide important operating errors. Numerical effect sizes and test splits were not verified.", "operating_period": "October 2026 issue; exact online publication date unknown. It cannot be dated after the previous update.", "limit": "Publisher abstract and indexed section text only; full paper unavailable. This is a synthetic benchmark, not a named failed field deployment or an economic result.", "question": "Does the forecast see interventions and constraints, and how does it behave when the operating regime changes?", "department": "Reservoir engineering & field development", "excerpt": "History-only inputs tend to fail to capture operational regime changes", "source_id": "WATCH15PM-FORECAST", "report_id": "WATCH15PM-FORECAST", "sources": [{"title": "Data-driven production forecast for oil and gas reservoirs - Advantages and limitations", "url": "https://www.sciencedirect.com/science/article/pii/S2949891026002137", "date": "2026-10 (issue month; exact online date unknown)", "publisher": "Geoenergy Science and Engineering / Elsevier", "locator": "Volume 265, article 214569; DOI 10.1016/j.geoen.2026.214569; publisher abstract and indexed sections, no PDF page locator", "access": "Editor independently read publisher-indexed abstract and section text; direct retrieval failed, and no full paper or numerical tables were obtained."}]}, {"id": "W15PM-OFGEM", "record": null, "related_ids": [], "company": "Ofgem AI Reg Lab", "title": "Keep an inferred asset record distinguishable from an original record", "date": "2026-04-29", "publisher": "Ofgem", "source_type": "Official regulator lab summary", "source_title": "AI Reg Lab: High-level findings for licensees and stakeholders — February 2026", "url": "https://www.ofgem.gov.uk/sites/default/files/2026-04/AI-Reg-Lab-Februray-2026.pdf", "locator": "PDF page 1: publication date; page 2: pilot and lab scope; page 3: themes; page 4: disclaimer", "status": "One anonymized energy pilot discussed; no published validation result", "summary": "A February lab examined a GenAI pilot that structures legacy asset records, fills gaps and builds an asset hierarchy.", "mechanism": "Participants called for separating authoritative records from AI inferences, testing across asset conditions, showing uncertainty and preserving change histories and human oversight.", "measure": "No asset count, measured reliability improvement, cost or ROI is published.", "operating_period": "Lab held 25 February 2026; summary published 29 April 2026.", "limit": "An anonymized discussion, not a system audit, observed failure, regulatory endorsement or compliance advice. Energy-infrastructure context; no petroleum operator is identified. Findings reflect participants’ views, not Ofgem policy.", "question": "Can an engineer tell what came from a record, what was inferred, and who accepted the change?", "department": "IT, data & enterprise transformation", "excerpt": "clear identification and distinguishability between authoritative data and AI-generated inferences", "source_id": "WATCH15PM-OFGEM", "report_id": "WATCH15PM-OFGEM", "sources": [{"title": "AI Reg Lab: High-level findings for licensees and stakeholders — February 2026", "url": "https://www.ofgem.gov.uk/sites/default/files/2026-04/AI-Reg-Lab-Februray-2026.pdf", "date": "2026-04-29", "publisher": "Ofgem", "locator": "PDF page 1: publication date; page 2: pilot and lab scope; page 3: themes; page 4: disclaimer", "access": "Editor read the original four-page PDF on 15 October 2026. © Crown copyright 2026; document reused under the Open Government Licence."}]}]}, "watch_2026_10_16_am": {"as_of": "2026-10-16", "run": "2026-10-16-am", "label": "16 October 2026 / Morning research", "headline": "A prediction can be accurate and still not help the decision.", "intro": "A reported case from TotalEnergies shows why a model can be dropped even when it predicts reasonably well. Two deployment updates also separate robots and better instruments from claims about AI.", "count_note": "Three source-linked notes added. Dropping one use case is different from ending a whole pilot. Robotics and imaging remain operating-technology context; AI-adoption, numerical-result and ROI counts are unchanged.", "method": "Three bounded Luna tracks plus editor verification and full-corpus deduplication. No qualifying news after the previous update was established. The account of the dropped use case still needs confirmation from the company or the original presentation; source dates and access limits are shown.", "archive_label": "More company context and earlier research", "findings": [{"id": "W16AM-TOTAL-DRUM", "record": null, "related_ids": ["AI-207"], "company": "TotalEnergies", "title": "TotalEnergies reportedly dropped one drum-cycling use case", "date": "2026-06-11", "publisher": "Control Global / Honeywell Users Group coverage", "source_type": "Original conference reporting of named operator remarks", "source_title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "url": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "locator": "Drum cycling prediction section; Thomas Klein remarks at Honeywell Users Group Americas 2026", "status": "One use case reportedly dropped; company confirmation still needed", "summary": "Control Global quotes TotalEnergies engineer Thomas Klein saying a Port Arthur drum-level model achieved 75% prediction accuracy, but the drum-cycling use case was not retained.", "mechanism": "Drum switching depended on conditions outside the model’s scope. The work provided advice, not autonomous control.", "measure": "The report does not define the accuracy denominator or test set, or quantify cost or loss. Other pilot work continued; the entire program was not reported abandoned.", "operating_period": "Drum work started July 2025; remarks reported June 2026. Exact decision date unknown.", "limit": "Original conference reporting, labeled Sponsored. We could not retrieve the presentation or a later company statement. The earlier company release confirms the broader pilot, not the decision to drop this use case. Keep these figures separate from its five-event result.", "question": "Did the model help operators choose when to switch drums, rather than merely predict drum level?", "department": "Refining, LNG & process operations", "excerpt": "so this use case wasn’t retained", "source_id": "WATCH16AM-TOTAL-DRUM", "sources": [{"title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "url": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "date": "2026-06-11", "publisher": "Control Global / Honeywell Users Group coverage", "locator": "Drum cycling prediction section; Thomas Klein remarks at Honeywell Users Group Americas 2026", "access": "Editor read the original conference report and both partners’ earlier releases. The decision to drop this use case rests on reported named remarks, not a company-authored outcome report."}, {"title": "TotalEnergies and Honeywell pilot AI-assisted control room", "url": "https://corporate.totalenergies.us/news/totalenergies-and-honeywell-pilot-ai-assisted-control-room-accelerated-shift-industrial", "date": "2025-11-11", "publisher": "TotalEnergies", "locator": "Initial DCU pilot and preliminary five-event result paragraphs", "access": "Original issuer release read. Establishes the existing pilot and operator corrective action; it predates and does not corroborate reported drum-use-case non-retention."}]}, {"id": "W16AM-HP", "record": null, "related_ids": ["AI-116"], "company": "Helmerich & Payne", "title": "ExxonMobil has two FlexRobotics systems deployed; seven more are planned", "date": "2026-10-07", "publisher": "Helmerich & Payne", "source_type": "Original company announcement", "source_title": "H&P & ExxonMobil Expand Deployment of FlexRobotics Technology", "url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HP--ExxonMobil-Expand-Deployment-of-FlexRobotics-Technology/default.aspx", "locator": "Opening release paragraphs: first two systems, seven additional units over next 12 months; ExxonMobil quotation", "status": "Two systems deployed; seven additional systems planned", "summary": "H&P reports two FlexRobotics systems deployed on its rigs for ExxonMobil. It expects seven more over the following twelve months, for nine in total.", "mechanism": "The robots automate repetitive rig-floor activities. The announcement does not identify an AI/ML mechanism, task-level approval, supervision or stop authority.", "measure": "Two is the deployed count; nine is the intended total. No quantified safety, time, cost or net return is disclosed.", "operating_period": "Published 7 October 2026. Initial deployment dates, basin and rig identities are not supplied.", "limit": "An ExxonMobil executive supports expansion intent; the joint account is not independent outcome validation. It does not establish installation of the separate AI camera add-on in AI-116. Keep as robotics context, outside AI-adoption counts.", "question": "What changes in crew exposure and connection performance after deployment, under comparable rig conditions?", "department": "Drilling & completions", "excerpt": "expects to deploy seven additional FlexRobotics systems", "source_id": "WATCH16AM-HP", "sources": [{"title": "H&P & ExxonMobil Expand Deployment of FlexRobotics Technology", "url": "https://ir.hpinc.com/news/press-releases/news-details/2026/HP--ExxonMobil-Expand-Deployment-of-FlexRobotics-Technology/default.aspx", "date": "2026-10-07", "publisher": "Helmerich & Payne", "locator": "Opening release paragraphs: first two systems, seven additional units over next 12 months; ExxonMobil quotation", "access": "Original H&P release read by editor; customer quotation corroborates intent, not quantified results."}]}, {"id": "W16AM-RETINA", "record": null, "related_ids": [], "company": "SLB", "title": "Earlier formation images are now reported from more than 50 wells", "date": "2026-10-15", "publisher": "SLB", "source_type": "Original company announcement", "source_title": "Retina At-Bit Imaging System Surpasses 50-Well Milestone in First Year of Deployment", "url": "https://www.slb.com/newsroom/features/2026/2026-1015-retina", "locator": "Published date; Global adoption and Global deployment sections", "status": "Commercial deployment milestone; vendor-reported", "summary": "SLB says Retina has imaged more than 50 wells across 18 locations during its first year of commercial deployment.", "mechanism": "The instrument acquires formation images at the bit while drilling, at roughly 2 mm resolution, to inform well placement and development decisions.", "measure": "Deployment and imaging specifications are not measured recovery, drilling-time savings or ROI. No comparative economics is disclosed.", "operating_period": "Published 15 October 2026; individual deployment dates and customer identities unknown. Publication time cannot be placed after the previous update.", "limit": "No AI/ML mechanism or automatic control is identified. Keep as instrumentation context; do not transfer OQ Oman’s separate drilling-tool results to Retina.", "question": "Does earlier imaging change a placement decision or avoid a cost, compared with the previous workflow?", "department": "Drilling & completions", "excerpt": "more than 50 wells across 18 locations worldwide", "source_id": "WATCH16AM-RETINA", "sources": [{"title": "Retina At-Bit Imaging System Surpasses 50-Well Milestone in First Year of Deployment", "url": "https://www.slb.com/newsroom/features/2026/2026-1015-retina", "date": "2026-10-15", "publisher": "SLB", "locator": "Published date; Global adoption and Global deployment sections", "access": "Editor read the original SLB feature; performance and scope remain vendor-reported."}]}]}, "watch_2026_10_16_pm": {"as_of": "2026-10-16", "run": "2026-10-16-pm", "label": "16 October 2026 / Evening research", "headline": "PTTEP describes where AI is already in use.", "intro": "Its CEO points to drilling-risk monitoring, alongside automated planning and production optimization. A separate IEA report examines the skills and data barriers that keep energy companies from scaling AI.", "count_note": "Two source-linked notes and one industry publication added. PTTEP’s broad operating account is not split into new deployment profiles. AI-adoption, measured-result and ROI counts are unchanged.", "method": "Three bounded Luna research tracks plus editor verification and full-corpus deduplication. Both retained sources predate the previous update; exact publication dates were verified by the editor. No new failed deployment was substantiated.", "archive_label": "More company context and earlier research", "findings": [{"id": "W16PM-PTTEP", "record": null, "related_ids": ["AI-067", "AI-245"], "company": "PTTEP", "title": "PTTEP’s CEO describes AI in drilling, alongside planning and production systems", "date": "2026-04-10", "publisher": "PTTEP", "source_type": "Original company meeting minutes", "source_title": "Minutes of the 2026 Annual General Shareholders’ Meeting", "url": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-minutes-en.pdf", "locator": "PDF page 13 / printed page 12 of 25, response to question 1.7; cover letter on PDF page 1 confirms publication date", "status": "Company reports operating use; individual assets and scale unspecified", "summary": "At its March AGM, PTTEP described deployed AI and automation across subsurface, drilling and production work, including AI that monitors drilling data for early risk detection.", "mechanism": "The other examples are automated planning of well locations and wellhead platforms, and systems optimizing production across many wells. Their specific AI methods are not identified.", "measure": "Management says benefits are reaching financial performance, but gives no isolated AI result, baseline, cost or ROI.", "operating_period": "Meeting held 30 March 2026; minutes posted 10 April, confirmed by the cover letter. Individual deployment dates are unknown.", "limit": "No fields, vendors, well count, model performance or approval/control boundaries are disclosed. The separate USD 328m new-business budget is not AI-only. This broad account does not establish the scope or success of the later offshore advisory pilot in AI-067.", "question": "Which fields use these systems, what do people approve, and what operating result survives a matched comparison?", "department": "Production & field operations", "excerpt": "continuously deployed AI and automation technologies across its core business processes", "source_id": "WATCH16PM-PTTEP", "sources": [{"title": "Minutes of the 2026 Annual General Shareholders’ Meeting", "url": "https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-minutes-en.pdf", "date": "2026-04-10", "publisher": "PTTEP", "locator": "PDF page 13 / printed page 12 of 25, response to question 1.7; cover letter on PDF page 1 confirms publication date", "access": "Editor read the official translated minutes and inspected the rendered passage. Company self-report, not independent outcome validation."}]}, {"id": "W16PM-IEA", "record": null, "related_ids": [], "company": "International Energy Agency", "title": "Skills and data remain barriers to using AI in energy operations", "date": "2026-04-16", "publisher": "International Energy Agency", "source_type": "Original industry research report", "source_title": "Key Questions on Energy and AI", "url": "https://www.iea.org/reports/key-questions-on-energy-and-ai", "locator": "Official report page: Published 16 April 2026; PDF page 10: demand outlook; PDF/printed page 84: Figure 9.4; pages 80–85: adoption barriers", "status": "System outlook and energy-sector survey; not an operator deployment", "summary": "The IEA’s 2026 follow-up to Energy and AI separates data-centre electricity demand from using AI to improve energy operations. Its company survey identifies internal digital skills as the largest adoption barrier.", "mechanism": "Chapter 9 also discusses fragmented data, data protection and equipment digitalisation. Figure 9.4 draws on the IEA Industry Employment Survey 2025.", "measure": "All data centres used an estimated 485 TWh in 2025; the report projects roughly 950 TWh in 2030. Those are not oil-and-gas productivity results or AI-only electricity demand.", "operating_period": "Published 16 April 2026, verified on the official page. Survey vintage: 2025; forecast horizon: 2030.", "limit": "Broad energy-sector evidence; the reviewed chapter supplies no survey respondent count or upstream-only subgroup. Projections are not realized benefits or ROI. This is a distinct follow-up to the 2025 report INS001, not a replacement.", "question": "Before buying another model, which missing skills, data connections or equipment would prevent people from using it?", "department": "Industry research", "excerpt": "the lack of digital skills is the single largest barrier", "source_id": "WATCH16PM-IEA", "sources": [{"title": "Key Questions on Energy and AI", "url": "https://www.iea.org/reports/key-questions-on-energy-and-ai", "date": "2026-04-16", "publisher": "International Energy Agency", "locator": "Official report page: Published 16 April 2026; PDF page 10: demand outlook; PDF/printed page 84: Figure 9.4; pages 80–85: adoption barriers", "access": "Editor read the official page and relevant full-PDF passages, including the rendered Figure 9.4. IEA report text is CC BY 4.0; Annex A has a separate noncommercial share-alike licence."}, {"title": "Key Questions on Energy and AI — full report", "url": "https://iea.blob.core.windows.net/assets/3179f7f8-01f6-4dd6-bffa-c9f7b73f1dc9/KeyQuestionsonEnergyandAI.pdf", "date": "2026-04-16", "publisher": "International Energy Agency", "locator": "PDF page 10; PDF/printed page 84, Figure 9.4; adoption discussion pages 80–85", "access": "Relevant original PDF passages read and chart page visually checked. Survey respondent count not supplied in the reviewed chapter; no upstream-specific failure or adoption rate inferred."}], "report_id": "WATCH16PM-IEA"}], "featured_count": 2}, "watch_2026_10_17_am": {"as_of": "2026-10-17", "run": "2026-10-17-am", "label": "17 October 2026 / Morning research", "headline": "CNRL reports savings from optimizing hydrogen with AI.", "intro": "CNRL names a hydrogen savings claim. A Baker Hughes case separates advisory use from its autonomous product branding. PDO’s agreement shows the data foundation it wanted to build, with rollout still unverified.", "count_note": "Three source notes, an expansion of existing AI-279, and a source recheck for Cenovus AI-061. No new adoption or measured-result profiles are counted. No new failed deployment was verified.", "method": "Three bounded Luna tracks, editor checks of originals and corpus deduplication. These are missed or clarified sources, not announcements since the last update. Original publication and operating dates remain unknown where the sources do not establish them.", "archive_label": "More company context and earlier research", "featured_count": 3, "findings": [{"id": "W17AM-CNRL", "record": "AI-279", "related_ids": ["AI-279", "AI-098"], "company": "Canadian Natural Resources", "title": "CNRL puts an annual savings figure on hydrogen optimization", "date": "2025-11-07", "publisher": "Canadian Natural Resources", "source_type": "Operator investor presentation", "source_title": "Investor Open House · 7 November 2025", "url": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "locator": "PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention", "status": "Operating use reported; site and start date unspecified", "summary": "CNRL says AI helps optimize hydrogen in its oil-sands upgrading operations.", "mechanism": "The presentation names the task but does not explain the model, inputs, vendor or approval process.", "measure": "The company reports about CAD 8 million in annual savings. This is a gross savings claim, without a disclosed all-cost ROI.", "operating_period": "Presentation dated 7 November 2025; exact posting and deployment dates unknown. File path is dated December 2025.", "limit": "No site, baseline, measurement period or attribution method is given. The CAD 107 million total on the slide mixes savings and revenue from several projects; it is not an AI total. Existing AI-279 reliability results are separate.", "question": "How much hydrogen was avoided, against which baseline, and what does the system cost to run?", "department": "Refining, LNG & process operations", "excerpt": "Optimization of hydrogen through use of AI — Saving ~$8 million per year.", "source_id": "WATCH17AM-CNRL", "sources": [{"title": "Investor Open House · 7 November 2025", "url": "https://www.cnrl.com/content/uploads/2025/12/V_CNQ-Investor-Open-House_Nov-7-2025.pdf#page=50", "date": "2025-11-07", "publisher": "Canadian Natural Resources", "locator": "PDF page / printed slide 50: hydrogen; slide 51: separate reliability program; PDF page 97: currency convention", "access": "Editor read original source; company/vendor claims are not independently validated."}]}, {"id": "W17AM-KANTORI", "record": null, "related_ids": [], "company": "Baker Hughes", "title": "An “autonomous” drilling product was used in advisory mode", "date": "2026-04", "publisher": "Baker Hughes", "source_type": "Vendor case sheet", "source_title": "Kantori drilling advisory case · UAE", "url": "https://dam.bakerhughes.com/m/6c9e1d4e04184778/original/i-Trak-service-maximized-ROP-saved-22hrs-UAE-cs.pdf#page=1", "locator": "Single-page case sheet; Solution, Results and footer revision 04/2026", "status": "Single-well advisory case; deployment date unknown", "summary": "Baker Hughes describes an advisory-mode UAE carbonate-drilling case supported from its remote operations centre.", "mechanism": "AI assisted torque-and-drag work. Despite the product name, the case explicitly says advisory mode; who approved parameter changes is not stated.", "measure": "The vendor reports 22 alerts, 22 hours saved against a previous well, 5% higher penetration rate and no stuck pipe or fluid losses.", "operating_period": "PDF revised April 2026, with a 2023 copyright. Original publication and field-work dates are unknown; this may be older operating context. The September 2026 webpage date was not independently verified.", "limit": "The customer and well are unnamed. A prior-well comparison does not isolate AI from the wider automation service, crew or geology. Costs, accepted recommendations and repeatability are undisclosed; no ROI is established. The deployment date is unknown; the PDF revision does not establish a 2026 deployment.", "question": "Which recommendations changed drilling practice, and did the time saving repeat on comparable wells?", "department": "Drilling & completions", "excerpt": "Deployed Kantori autonomous directional drilling in advisory mode", "source_id": "WATCH17AM-KANTORI", "sources": [{"title": "Kantori drilling advisory case · UAE", "url": "https://dam.bakerhughes.com/m/6c9e1d4e04184778/original/i-Trak-service-maximized-ROP-saved-22hrs-UAE-cs.pdf#page=1", "date": "2026-04", "publisher": "Baker Hughes", "locator": "Single-page case sheet; Solution, Results and footer revision 04/2026", "access": "Original PDF read and rendered. HTML case page returned 403 to editor. Revision date is not a field-deployment date."}]}, {"id": "W17AM-PDO", "record": null, "related_ids": [], "company": "Petroleum Development Oman", "title": "PDO contracted for a shared industrial data and workflow system", "date": "2025-09-08", "publisher": "KONGSBERG / NTB", "source_type": "Issuer press release", "source_title": "Kongsberg Digital and PDO digital-twin agreement", "url": "https://kommunikasjon.ntb.no/pressemelding/18645626/kongsberg-digital-and-petroleum-development-oman-sign-agreement-for-deployment-of-ai-powered-digital-twin?lang=en&publisherId=17849150", "locator": "Release timestamp 8 September 2025; scope and deployment-ambition paragraphs", "status": "Agreement announced; operating rollout unverified", "summary": "PDO selected Kongsberg Digital’s Kognitwin to connect operational and enterprise data across operations and maintenance.", "mechanism": "The release calls it AI-integrated, but does not identify an AI task, model or authority to act.", "measure": "Safety, production and cost improvements are goals. No measured benefit, contract value or ROI is disclosed.", "operating_period": "Announced 8 September 2025, with an ambition to start deploying the twins that year. No later completion evidence was verified in this pass.", "limit": "No fields, facilities, users or workflow counts are identified. The company-description paragraph about gas operations does not define deployment scope. A passed target date alone does not establish delay or failure.", "question": "Which facilities are live, which tasks use AI, and what do people still approve?", "department": "IT, data & enterprise transformation", "excerpt": "Kognitwin unifies and contextualises operational (OT) and enterprise (IT) data", "source_id": "WATCH17AM-PDO", "sources": [{"title": "Kongsberg Digital and PDO digital-twin agreement", "url": "https://kommunikasjon.ntb.no/pressemelding/18645626/kongsberg-digital-and-petroleum-development-oman-sign-agreement-for-deployment-of-ai-powered-digital-twin?lang=en&publisherId=17849150", "date": "2025-09-08", "publisher": "KONGSBERG / NTB", "locator": "Release timestamp 8 September 2025; scope and deployment-ambition paragraphs", "access": "Editor read original source; company/vendor claims are not independently validated."}]}]}, "watch_2026_10_17_pm": {"as_of": "2026-10-17", "run": "2026-10-17-pm", "label": "17 October 2026 / Evening research", "headline": "INPEX changed its processes alongside its core software.", "intro": "INPEX describes changing its core business system and working processes. A separate reservoir-planning study shows why a promising model result still needs field validation.", "count_note": "Two supplementary notes and one technical publication added. The library now contains 65 publications. Adoption, operating-deployment and counted-result profile totals are unchanged.", "method": "Three Luna tracks used eight search clauses each, followed by editor checks and full-corpus deduplication. Neither retained source is newer than the previous update. No new named failed deployment was verified.", "archive_label": "More company context and earlier research", "featured_count": 2, "findings": [{"id": "W17PM-INPEX", "record": null, "related_ids": ["AI-294"], "company": "INPEX", "title": "INPEX changed its core system and the processes around it", "date": "2026-06-24", "publisher": "INPEX / SAP Japan / ABeam Consulting", "source_type": "Joint company announcement", "source_title": "INPEX commences operations of SAP Cloud ERP", "url": "https://www.inpex.com/english/news/upload/20260624.pdf#page=2", "locator": "PDF pages 1–2; Japan go-live and Fit-to-Standard approach on page 2", "status": "ERP operating in Japan since January 2026; no AI use case established", "summary": "INPEX reports that SAP Cloud ERP Private began full operation in Japan in January 2026.", "mechanism": "With ABeam’s support, it aligned business processes to SAP standards instead of filling gaps with custom add-ons.", "measure": "The release says development was shortened, without giving elapsed time or a comparison. Reporting and workload improvements are objectives; no measured savings or ROI are disclosed.", "operating_period": "Joint release published 24 June 2026; Japan go-live January 2026. Other countries’ rollout dates are not supplied.", "limit": "AI appears as compatibility and a future efficiency objective, not a deployed function. Users, departments, costs and approval rules are unspecified. No connection to the separate Copilot/management-meeting agents in AI-294 is established.", "question": "Which processes changed, and can the separate AI tools access those records with appropriate permissions?", "department": "IT, data & enterprise transformation", "excerpt": "full-scale operation to commence in Japan in January 2026", "source_id": "WATCH17PM-INPEX", "sources": [{"title": "INPEX commences operations of SAP Cloud ERP", "url": "https://www.inpex.com/english/news/upload/20260624.pdf#page=2", "date": "2026-06-24", "publisher": "INPEX / SAP Japan / ABeam Consulting", "locator": "PDF pages 1–2; Japan go-live and Fit-to-Standard approach on page 2", "access": "Editor read the official three-page PDF and inspected page 2. ERP use is established; AI operation and Copilot integration are not."}]}, {"id": "W17PM-FDP", "record": null, "related_ids": [], "company": "OriGen.AI / SPE", "title": "A better modeled development plan is a hypothesis to test in the field", "date": "2025-11-03", "publisher": "SPE / OnePetro; authors affiliated with OriGen.AI", "source_type": "Supplier-authored conference modeling study", "source_title": "Field Development Plan Optimization Under Uncertainty Using Physics-Informed Neural Networks and Generative AI: A Real-Field Case Study in North America", "url": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/792786", "locator": "SPE-229333-MS, abstract; publisher metadata: Published November 03 2025; conference 3–6 November 2025", "status": "Modeling study; execution of the optimized plan not established", "summary": "An OriGen.AI-authored study models development choices for one unnamed North American reservoir represented by 10 completed wells and four reservoir units.", "mechanism": "The abstract describes generative AI proposing well locations and trajectories, physics-informed production surrogates, and metaheuristic optimization. It does not establish an LLM or chat agent.", "measure": "Authors report about 24% higher expected NPV and roughly 3% surrogate error. These are modeled comparisons, not realized production, cash flow or ROI.", "operating_period": "Published 3 November 2025, according to the publisher’s indexed metadata. Field-data and test dates are not supplied.", "limit": "Editor checked the publisher-indexed abstract; direct page access failed and the full paper was unavailable. Error definition, test split, economic baseline, costs, human review and operator identity remain unverified. One field does not establish repeatability.", "question": "Would the preferred plan survive an independent simulator, economic assumptions and engineering review before capital is committed?", "department": "Reservoir engineering & field development", "excerpt": "expected net present value", "source_id": "WATCH17PM-FDP", "sources": [{"title": "Field Development Plan Optimization Under Uncertainty Using Physics-Informed Neural Networks and Generative AI: A Real-Field Case Study in North America", "url": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/792786", "date": "2025-11-03", "publisher": "SPE / OnePetro; authors affiliated with OriGen.AI", "locator": "SPE-229333-MS, abstract; publisher metadata: Published November 03 2025; conference 3–6 November 2025", "access": "Publisher-indexed metadata and complete abstract inspected by editor. Direct OnePetro page failed; full paper, figures and validation protocol not reviewed."}], "report_id": "WATCH17PM-FDP"}]}, "watch_2026_10_18_am": {"as_of": "2026-10-18", "run": "2026-10-18-am", "label": "18 October 2026 / Morning research", "headline": "KOC is testing a drilling assistant against its own procedures.", "intro": "SANAD gives engineers a way to find technical guidance, with people checking the answers. Two other accounts show how energy data products and field connectivity are being built around specific jobs.", "count_note": "One distinct pilot profile added: AI-295. The register now has 295 profiles, including 209 company-adoption profiles. Operating/scaled and counted-result totals remain unchanged. Two supplementary technology notes are separate from adoption counts.", "method": "Three bounded Luna tracks used eight search clauses each, followed by editor checks of original sources and full-corpus deduplication. Retained sources predate the previous update. No new named failed deployment was verified.", "archive_label": "More company context and earlier research", "featured_count": 3, "findings": [{"id": "W18AM-KOC", "record": "AI-295", "related_ids": ["AI-295", "AI-085"], "company": "Kuwait Oil Company", "title": "SANAD helps drilling engineers find the procedure they need", "date": "2026-09-29", "publisher": "Kuwait Oil Company / KUNA interview carried by Kuwait Times", "source_type": "Operator implementation account and attributed engineer interview", "source_title": "KOC SANAD drilling knowledge assistant", "url": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "locator": "Official Digest: PDF pages 28–31 / printed 26–29; September interview: field-test and decision-responsibility paragraphs", "status": "Pilot and validation; later field test reported, wider rollout unverified", "summary": "KOC describes SANAD, a drilling knowledge assistant developed with SLB and piloted by its North & West Kuwait Drilling Group.", "mechanism": "The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision.", "measure": "A later engineer interview reports one Lower Burgan well finished 13 days early and SOP-related savings of about KD 97,000 per well. It does not isolate SANAD’s contribution; neither figure is counted as an AI result or ROI.", "operating_period": "Development reportedly began March 2025. KOC describes a 2025–26 pilot; its April–June 2026 issue has no verified posting day. The interview was published 29 September 2026. Field-test dates and well ID are not supplied.", "limit": "No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085.", "question": "Can engineers trace each answer to the right procedure, and does it improve decisions after review time and errors are included?", "department": "Drilling & completions", "source_id": "WATCH18AM-KOC-INTERVIEW", "excerpt": "Engineering judgment and responsibility remain with the specialist", "sources": [{"id": "WATCH18AM-KOC-DIGEST", "title": "SANAD AI: The Generative AI Drilling Assistant", "url": "https://www.kockw.com/sites/EN/Periodicals/Digest%20April%20-%20June%202026.pdf#page=28", "date": "2026-Q2", "publisher": "Kuwait Oil Company", "locator": "PDF pages 28–31 / printed 26–29; 14 SOPs on PDF 30; validation, ownership and roadmap on PDF 31", "access": "Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.", "excerpt": "Every output was validated"}, {"id": "WATCH18AM-KOC-INTERVIEW", "title": "KOC’s Sanad uses AI to unlock drilling knowledge", "url": "https://kuwaittimes.com/article/50684/kuwait/other-news/kocs-sanad-uses-ai-to-unlock-drilling-knowledge/", "date": "2026-09-29", "publisher": "KUNA interview, republished by Kuwait Times", "locator": "Timestamp; lower-Burgan field test, SOP-related savings, RAG description and human responsibility", "access": "Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.", "excerpt": "Engineering judgment and responsibility remain with the specialist"}]}, {"id": "W18AM-SPGLOBAL", "record": null, "related_ids": [], "company": "S&P Global Energy", "title": "Energy data is becoming something customers’ agents can query", "date": "2026-09-25", "publisher": "Databricks / S&P Global Energy customer account", "source_type": "Vendor-published technical customer account", "source_title": "From Data to Dialogue: S&P Global Energy’s Genie and MCP architecture", "url": "https://www.databricks.com/blog/data-dialogue-how-sp-global-energy-made-its-structured-data-estate-conversational-databricks", "locator": "Dated customer account; agent curation, permissions, proxy architecture and benchmark sections", "status": "Built architecture described; launch date and user population undisclosed", "summary": "S&P Global Energy describes giving internal and customer agents conversational access to structured energy datasets.", "mechanism": "Domain experts curate narrow dataset agents. A reusable MCP interface combines them across commodities. The account says table permissions and audit controls carry through, and experts can test answers against benchmarks.", "measure": "A customer executive reports setup shrinking from development cycles measured in months to days. No sample, measured accuracy, active-user count, cost or ROI is supplied.", "operating_period": "Published 25 September 2026. Implementation and external-customer launch dates are not disclosed.", "limit": "This is a vendor-published customer account, not an independent quality or security assessment. Benchmarks are described without scores. Access to research data does not establish authority to trade, approve investment or control field equipment.", "question": "Who owns the data definitions, and which benchmark and permission checks must pass before a new agent can use them?", "department": "IT, data & enterprise transformation", "source_id": "WATCH18AM-SPGLOBAL", "excerpt": "now takes days", "sources": [{"id": "WATCH18AM-SPGLOBAL", "title": "From Data to Dialogue: S&P Global Energy’s Genie and MCP architecture", "url": "https://www.databricks.com/blog/data-dialogue-how-sp-global-energy-made-its-structured-data-estate-conversational-databricks", "date": "2026-09-25", "publisher": "Databricks / S&P Global Energy", "locator": "Date; per-dataset curation; Unity Catalog permissions; FastMCP composition; customer quote and benchmark loop", "access": "Editor read the original vendor customer account. Architecture and benefits remain vendor/customer-reported.", "excerpt": "now takes days"}]}, {"id": "W18AM-PETROREC", "record": null, "related_ids": [], "company": "PetroReconcavo", "title": "PetroReconcavo connects its field network and back-office systems", "date": "2025-07-14", "publisher": "PetroReconcavo", "source_type": "Operator sustainability report and publication notice", "source_title": "2024 Sustainability Report · digital transformation", "url": "https://www.petroreconcavo.com.br/Download.aspx?Arquivo=OXSl%2FB7SEmrx%2BHoNzBeFlQ%3D%3D#page=78", "locator": "PDF / printed page 78: SAP implementation, Projeto Waves and data lake; issuer notice confirms publication date", "status": "ERP implemented and six LTE towers activated in 2024 reporting; GenAI scope unresolved", "summary": "PetroReconcavo reports implementing SAP and activating six LTE towers in Bahia and Rio Grande do Norte.", "mechanism": "The ERP work connected production data and material tracking. The field network supports monitoring and prepares for IoT and automation. A separate data-lake passage describes GenAI analytics, reports and assistants without naming a tool or defining live use.", "measure": "The report lists more than 2,200 training hours for SAP. Faster response and stronger controls are qualitative claims; no measured savings, cost or ROI is disclosed.", "operating_period": "Portuguese report published 14 July 2025, confirmed by issuer notice. It covers 2024; exact go-live dates are unknown, so an in-window deployment is not established.", "limit": "No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology.", "question": "Which field and finance tasks improved after the systems went live, and what did people still have to do manually?", "department": "IT, data & enterprise transformation", "source_id": "WATCH18AM-PETROREC-DATE", "excerpt": "published, on this date, its 2024 Sustainability Report", "sources": [{"id": "WATCH18AM-PETROREC", "title": "Relatório de Sustentabilidade 2024", "url": "https://www.petroreconcavo.com.br/Download.aspx?Arquivo=OXSl%2FB7SEmrx%2BHoNzBeFlQ%3D%3D#page=78", "date": "2025-07-14", "publisher": "PetroReconcavo", "locator": "PDF / printed page 78: SAP, Projeto Waves and data lake", "access": "Editor retrieved the original 100-page Portuguese PDF after encoding the plus sign in its download parameter; page 78 read and rendered.", "excerpt": "A nova rede possibilita monitoramento em tempo real"}, {"id": "WATCH18AM-PETROREC-DATE", "title": "2024 Sustainability Report · Notice to the Market", "url": "https://ri.petroreconcavo.com.br/Download.aspx?Arquivo=6hzxLJkhGOrPzcEDtUKx3A%3D%3D&IdCanal=ckeQVNzVtM%2FZZ4T5tqRLLw%3D%3D", "date": "2025-07-14", "publisher": "PetroReconcavo", "locator": "PDF pages 1–2; publication announcement and English-edition timing", "access": "Original issuer notice read; Portuguese publication 14 July, English edition separately planned for 25 July 2025.", "excerpt": "published, on this date, its 2024 Sustainability Report"}]}]}, "watch_2026_10_18_pm": {"as_of": "2026-10-18", "run": "2026-10-18-pm", "label": "18 October 2026 / Evening research", "headline": "The useful question is what changed in the work.", "intro": "A planning pilot, a maintenance estimate and two field workflows show why recommendations, automatic adjustments and realized benefits need separate checks.", "count_note": "Three existing profiles clarified; one anonymous-customer supplier implementation added as context. Profiles remain 295, company-adoption 209, operating/scaled 144 and counted-result profiles 60. The library also indexes an already-covered JPT field case; no deployment is counted twice.", "method": "Three Luna tracks used eight search clauses each, followed by editor source checks and full-corpus deduplication. Sources predate the last update. No new named failed deployment verified; this is a bounded public-source search.", "featured_count": 4, "archive_label": "More company context and earlier research", "findings": [{"id": "W18PM-OMV", "record": "AI-133", "related_ids": ["AI-133"], "company": "OMV", "title": "OMV tests field-development decisions with analytical agents", "department": "Reservoir engineering & field development", "status": "CCS planning pilot; early internal results", "summary": "OMV’s Norwegian CCS pilot uses decision theory and geostatistics with Stanford and TerraAI.", "mechanism": "Seismic, geological and well data feed scenario planning for drilling, injection and monitoring. Engineers challenge assumptions and validate recommendations. OMV explicitly distinguishes this analytical system from generative AI.", "measure": "OMV reports over 25% higher project value, around 50% less quantified subsurface risk and simulations up to 1,000 times faster. These are early internal planning claims, not realized production or net ROI.", "operating_period": "Norwegian CCS planning pilot; article published 14 April 2026. Test dates, asset name and project count unknown.", "limit": "No metric definitions, comparator details, validation sample or complete costs. The wider six-year Stanford sponsorship is not this pilot’s cost. Pilot remains separate from NAIC document search (AI-132).", "question": "Do better plans still outperform after engineering review and actual field execution?", "sources": [{"id": "WATCH18PM-OMV", "title": "How AI is powering better decisions in the race to net zero", "url": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "date": "2026-04-14", "publisher": "OMV", "locator": "Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker", "excerpt": "not generative AI, but analytical, agent-based AI", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "date": "2026-04-14", "publisher": "OMV", "source_type": "Operator technical account", "source_id": "WATCH18PM-OMV", "source_title": "How AI is powering better decisions in the race to net zero", "url": "https://www.omv.com/en/magazine/blogs/2026/how-ai-is-powering-better-decisions-in-the-race-to-net-zero", "locator": "Sections: 1,000 times faster; Creating value while reducing risk; A discussion partner, not a decision maker", "excerpt": "not generative AI, but analytical, agent-based AI"}, {"id": "W18PM-WOODSIDE", "record": "AI-042", "related_ids": ["AI-042"], "company": "Woodside", "title": "Woodside’s maintenance estimate has a traceable source", "department": "Maintenance & reliability", "status": "Operating at North West Shelf; potential maintenance-hours reduction", "summary": "Woodside’s AI page supports the previously unresolved 15% maintenance-hours estimate at Angel.", "mechanism": "Maint Intel compares maintenance history with reliability targets to recommend intervals. A separate January article identifies AWS support and the North West Shelf launch.", "measure": "The AI page describes potential 15% lower annual maintenance hours over five years, not hours already avoided. The January article reports model processing cut from five days to under two hours; that is not a field-labor saving.", "operating_period": "Tested at Angel and launched at North West Shelf. January article: 5 January 2026. AI page last updated 27 May 2026; original publication date and estimate period unknown.", "limit": "The earlier audit checked the January article alone. No independent validation, achieved maintenance-hours result or full costs. Global rollout was planned for 2026; completion remains unverified.", "question": "Were the recommended intervals adopted, and did reliability hold while maintenance hours fell?", "sources": [{"id": "S24", "title": "Artificial Intelligence at Woodside", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "date": "Updated 2026-05-27; publication day unknown", "publisher": "Woodside", "locator": "Maintenance section; last-updated stamp at foot of page", "excerpt": "a potential 15% reduction of maintenance hours", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}, {"id": "RS006", "title": "Right fix at right time – AI delivers for Woodside", "url": "https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside", "date": "2026-01-05", "publisher": "Woodside", "locator": "Angel pilot / North West Shelf launch, AWS collaboration and model processing-time paragraphs", "excerpt": "five days to under two hours", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "date": "Updated 2026-05-27; publication day unknown", "publisher": "Woodside", "source_type": "Operator publication", "source_id": "S24", "source_title": "Artificial Intelligence at Woodside", "url": "https://www.woodside.com/what-we-do/innovation/artificial-intelligence", "locator": "Maintenance section; last-updated stamp at foot of page", "excerpt": "a potential 15% reduction of maintenance hours"}, {"id": "W18PM-NOV", "record": null, "related_ids": [], "company": "NOV", "title": "NOV connects two agents to recommend maintenance", "department": "Maintenance & reliability", "status": "Supplier reports implementation; customer unnamed", "summary": "NOV describes its agent interfacing with a Norwegian operator’s agent at a sulphate-removal unit.", "mechanism": "Real-time equipment diagnostics on the Max Platform support maintenance recommendations. The release does not establish automatic execution or identify an LLM.", "measure": "No measured accuracy, downtime reduction, savings, cost or ROI disclosed.", "operating_period": "Release: 28 July 2026. Implementation start and duration unknown.", "limit": "Supplier-only account; operator, unit and user population unnamed. Approval, override and safety rules undisclosed. Separate service-rig monitoring in the same release is not evidence of AI.", "question": "Who checks the recommendation, and which system can authorize maintenance work?", "sources": [{"id": "WATCH18PM-NOV", "title": "NOV Reports Second Quarter 2026 Earnings", "url": "https://investors.nov.com/news-releases/news-release-details/nov-reports-second-quarter-2026-earnings", "date": "2026-07-28", "publisher": "NOV", "locator": "Significant Achievements: Norwegian sulphate-removal unit; separate remote service-rig monitoring paragraph", "excerpt": "recommend maintenance actions using the Max Platform", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "date": "2026-07-28", "publisher": "NOV", "source_type": "Supplier earnings release", "source_id": "WATCH18PM-NOV", "source_title": "NOV Reports Second Quarter 2026 Earnings", "url": "https://investors.nov.com/news-releases/news-release-details/nov-reports-second-quarter-2026-earnings", "locator": "Significant Achievements: Norwegian sulphate-removal unit; separate remote service-rig monitoring paragraph", "excerpt": "recommend maintenance actions using the Max Platform"}, {"id": "W18PM-PRESIDIO", "record": "AI-231", "related_ids": ["AI-231"], "company": "Presidio Production", "title": "Presidio separates field recommendations from controller adjustments", "department": "Production & field operations", "status": "Existing operating profile; Q2 workflow detail clarified", "summary": "Presidio’s Q2 call distinguishes DOUG recommendations, weekend visit routing and automatic plunger-cycle adjustments.", "mechanism": "DOUG alerts field engineers; pumpers receive a prioritized map. Controllers on 24 wells continuously adjust cycle timing. Management says field judgment remains with staff; controller safeguards are unspecified.", "measure": "Reported gas rates on the controller population: 2,680 to 2,946 Mcf/day, roughly 10%. Weekend production reportedly rose 2.5%. These overlap the existing 2.3% program-uplift claim and must not be added to it.", "operating_period": "24 controller wells within Presidio’s operated base. Call: 12 August 2026; controller pre/post periods and weekend baseline unknown. Platform scope is a separate population.", "limit": "Already-reviewed source, not a new deployment. No matched control, decline adjustment or full costs. Existing $4.5m annualized program revenue is not net cash. Platform coverage (~2,000 call / ~2,300 later goal slide) is not controller coverage.", "question": "What happened on comparable wells, and what net benefit remained after controller and field-work costs?", "sources": [{"id": "OCT5-S024", "title": "Presidio Production Company Q2 2026 Earnings Call Transcript", "url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "date": "2026-08-12", "publisher": "Presidio Production", "locator": "Chris Hammack remarks, printed page 3: weekend coverage, 24 plunger boxes, human judgment; Will Ulrich page 2 for aggregate uplift/revenue", "excerpt": "None of these replace field judgment", "access": "Original publisher text read by editor 18 October 2026; claims remain source-reported."}], "date": "2026-08-12", "publisher": "Presidio Production", "source_type": "SEC-filed operator earnings-call transcript", "source_id": "OCT5-S024", "source_title": "Presidio Production Company Q2 2026 Earnings Call Transcript", "url": "https://www.sec.gov/Archives/edgar/data/2083125/000121390026089284/ea030206301ex99-2.htm", "locator": "Chris Hammack remarks, printed page 3: weekend coverage, 24 plunger boxes, human judgment; Will Ulrich page 2 for aggregate uplift/revenue", "excerpt": "None of these replace field judgment"}]}, "watch_2026_10_19_am": {"as_of": "2026-10-19", "run": "2026-10-19-am", "label": "19 October 2026 / Morning research", "headline": "Useful AI work still needs a way to check the answer.", "intro": "A petroleum safety workshop and a trading-data example show where shared definitions, traceability and human review matter.", "count_note": "Two supplementary notes and one research-library entry added. The main register stays at 295 profiles, including 209 company-adoption profiles; operating/scaled and counted-result totals are unchanged.", "method": "Three research tracks, followed by editor original-source checks and full-corpus deduplication. No post-cutoff source or new named failed deployment verified. This is a bounded public-source search, not an exhaustive census.", "featured_count": 2, "archive_label": "More company context and earlier research", "findings": [{"id": "W19AM-HAVTIL", "record": null, "related_ids": [], "company": "DNV / Havtil", "title": "Safety learning needs common definitions and traceable answers", "department": "HSE & emissions", "status": "Workshop and tool demonstration; operating deployment unestablished", "summary": "Havtil’s account of a 40-person petroleum workshop describes shared incident-learning work and concerns about unreliable or untraceable AI answers.", "mechanism": "The slides show PDF extraction, a shared map of incident concepts and a searchable database. The approach supports review; it does not replace root-cause investigation for high-severity incidents.", "measure": "No accuracy score, incident reduction, cost or ROI. Participant concerns are general experience, not a measured evaluation of the demonstrated tool.", "operating_period": "Workshop: 28 October 2025. Article: 7 January 2026. The tool was developed in 2024; exact development and slide-posting dates unknown.", "limit": "Forty representatives are not forty companies or a survey sample. No named failure or agreed follow-up project is identified. Listed framework user-test participants do not establish AI deployments. This is distinct from the 2024 responsible-AI overview (INS004).", "question": "Can reviewers trace each extracted cause to the original passage and correct a classification?", "sources": [{"id": "WATCH19AM-HAVTIL", "title": "Can artificial intelligence help improve learning from incidents?", "url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2026/can-artificial-intelligence-help-improve-learning-from-incidents/", "date": "2026-01-07", "publisher": "Norwegian Ocean Industry Authority (Havtil)", "locator": "Workshop date/attendance; How AI can help; Quality and trust; Obstacles; The way forward", "excerpt": "No specific further activities were identified", "access": "Original regulator account read by editor 19 October 2026. Workshop experience is not a representative survey or a tool benchmark."}, {"id": "WATCH19AM-HAVTIL-SLIDES", "title": "Bruk av kunstig intelligens til læring etter hendelser", "url": "https://www.havtil.no/contentassets/2152e5ff08734a219a61315d4a50464c/havtil-arbeidsmote-20251028-ki-for-laring-av-hendelser.pdf#page=15", "date": "2025-10-28 workshop; posting date unknown", "publisher": "Havtil / DNV", "locator": "36-page deck: physical/printed p. 15 workflow; p. 20 contributing-factor framework and high-severity root-cause-analysis boundary", "excerpt": "Prat med data", "access": "Original PDF text read; pages 15 and 20 rendered and inspected. Web screenshot failed; local PyMuPDF rendering succeeded. Event date is not a verified posting date."}], "date": "2026-01-07", "publisher": "Norwegian Ocean Industry Authority (Havtil)", "source_type": "Regulator workshop account and technical slides", "source_id": "WATCH19AM-HAVTIL", "source_title": "Can artificial intelligence help improve learning from incidents?", "url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2026/can-artificial-intelligence-help-improve-learning-from-incidents/", "locator": "Workshop date/attendance; How AI can help; Quality and trust; Obstacles; The way forward", "excerpt": "No specific further activities were identified"}, {"id": "W19AM-TOTAL-AWS", "record": null, "related_ids": [], "company": "TotalEnergies", "title": "TotalEnergies uses generative AI to check trading data", "department": "Trading & marketing", "status": "Vendor reports implementation; date and rollout scope unknown", "summary": "AWS describes a TotalEnergies Trading & Shipping system that matches structured and unstructured transaction data.", "mechanism": "The workflow supports transaction traceability and input-error detection. AWS lists Bedrock and ECS; the model, connectors and exception-handling process are unspecified.", "measure": "A named customer manager reports more reliable controls and fewer financial losses, without amounts, error rates, a baseline or an observation period.", "operating_period": "AWS page undated. Implementation date, trading-desk scope, user count and transaction volume unknown; an in-window launch is not established.", "limit": "Vendor-published customer testimony, not independent validation. No costs or net ROI. Matching records does not establish authority to place trades, release payments or change books. No specific model provider can be inferred from the Bedrock service description.", "question": "Which mismatches need a person’s approval, and how much error-review work does the system remove?", "sources": [{"id": "WATCH19AM-TOTAL-AWS", "title": "TotalEnergies puts Gen AI to work on its trading operations", "url": "https://aws.amazon.com/solutions/case-studies/totalenergies/", "date": "Undated", "publisher": "AWS / TotalEnergies Trading & Shipping", "locator": "Overview; Olivier Collon customer quotation; AWS Services Used", "excerpt": "matching structured and unstructured data within transaction flows", "access": "Original AWS customer account read by editor 19 October 2026. No visible publication date; a secondary 2024-10-28 clue could not be confirmed from an original dated source."}], "date": "Undated", "publisher": "AWS / TotalEnergies Trading & Shipping", "source_type": "Vendor-published customer account", "source_id": "WATCH19AM-TOTAL-AWS", "source_title": "TotalEnergies puts Gen AI to work on its trading operations", "url": "https://aws.amazon.com/solutions/case-studies/totalenergies/", "locator": "Overview; Olivier Collon customer quotation; AWS Services Used", "excerpt": "matching structured and unstructured data within transaction flows"}]}, "watch_2026_10_19_extra": {"as_of": "2026-10-19", "run": "2026-10-19-extra", "label": "19 October 2026 / Additional research", "headline": "The hard part starts after the pilot.", "intro": "Murphy’s maintenance project shows what can go wrong when alerts meet real work. Two offshore examples show how much useful technology sits outside the AI label.", "count_note": "Three supplementary notes and one technical-library entry added. The main register remains at 295 profiles, including 209 company-adoption profiles; operating/scaled and counted-result totals are unchanged.", "method": "Three bounded Luna research tracks, followed by editor checks of accessible original sources and full-corpus deduplication. These are missed earlier disclosures, not new announcements since this morning. This is not an exhaustive search.", "featured_count": 3, "archive_label": "More company context and earlier research", "findings": [{"id": "W19EX-MURPHY", "record": null, "related_ids": [], "company": "Murphy Oil", "title": "A good pilot did not make maintenance AI easy to deploy", "department": "Maintenance & reliability", "status": "Retrospective operating project; current status unknown", "summary": "Murphy’s two-platform, 24-month project deployed 46 rotating-equipment models. Missing data delayed the first model until six months after kickoff.", "mechanism": "Incomplete maintenance records and generic anomaly alerts created investigation work. Engineers and operations staff reviewed alerts; retraining reportedly reduced false positives. Verified alerts fed maintenance notifications, later through an API.", "measure": "No quantified downtime savings, project cost or net ROI. A known-failure pilot’s 100–121-day warning was a replay, not demonstrated live lead time.", "operating_period": "Article: 20 January 2025. Alert records begin January/March 2023, before this report’s rolling window; exact project start/end dates unknown.", "limit": "The reviewed 2023 alerts are older context; current operating status is unknown. The authors cite insufficient supplier equipment expertise and reported mitigation—not abandonment. Platform names, supplier and precision/recall are undisclosed; no autonomous maintenance established.", "question": "How many alerts lead to useful work, and what does reviewing the rest cost?", "sources": [{"id": "WATCH19EX-MURPHY", "title": "Framework for AI- and ML-Based Predictive Maintenance for Offshore Rotating Equipment", "url": "https://jpt.spe.org/framework-for-ai-and-ml-based-predictive-maintenance-for-offshore-rotating-equipment", "date": "2025-01-20", "publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "locator": "Methodology; Observations; Discussion: Project Planning and Management, Data Prerequisites, Domain Knowledge, Model Development, Performance, and User Experience; Figs. 4–6", "excerpt": "the first predictive model was deployed 6 months after the project kickoff", "access": "Original technical account read by editor 19 October 2026. Text and qualitative figure discussion reviewed; numeric chart values not transcribed. Not an independent outcome audit."}], "date": "2025-01-20", "publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "source_type": "Technical project account", "source_id": "WATCH19EX-MURPHY", "source_title": "Framework for AI- and ML-Based Predictive Maintenance for Offshore Rotating Equipment", "url": "https://jpt.spe.org/framework-for-ai-and-ml-based-predictive-maintenance-for-offshore-rotating-equipment", "locator": "Methodology; Observations; Discussion: Project Planning and Management, Data Prerequisites, Domain Knowledge, Model Development, Performance, and User Experience; Figs. 4–6", "excerpt": "the first predictive model was deployed 6 months after the project kickoff"}, {"id": "W19EX-NOV-AURA", "record": null, "related_ids": [], "company": "NOV", "title": "Remote crane control still has a person at the controls", "department": "Operations & production", "status": "Supplier test and one delivery; customer shore operation unestablished", "summary": "NOV reports controlling a full-size crane from about 20 km away and delivering one Aura-equipped remote-operated crane to an unnamed customer.", "mechanism": "Aura brings camera feeds, load information, equipment diagnostics and environmental data into the human operator’s screen. Customers are expected to start onboard before moving control ashore.", "measure": "No measured cost, safety or crew reduction. Two further cranes were being built for delivery by early 2027; these are planned deliveries.", "operating_period": "Article: 20 August 2026. Supplier test: Kristiansand to Søgne, Norway; exact test and first-delivery dates unknown.", "limit": "No AI/ML identified. Delivery does not establish shore-controlled customer operations or autonomous lifting. Customer, vessel, operational hours, interlocks and fallback arrangements are undisclosed.", "question": "What must be proven onboard before a customer can safely move the operator ashore?", "sources": [{"id": "WATCH19EX-NOV-AURA", "title": "Aura moves remote crane operations from concept to reality", "url": "https://www.nov.com/news/aura-moves-remote-crane-operations-from-concept-to-reality", "date": "2026-08-20", "publisher": "NOV", "locator": "Remote-access paragraph; Kristiansand/Søgne test; first delivery; two planned deliveries; phased customer adoption", "excerpt": "One fully remote-operated crane has already been delivered", "access": "Original supplier account read by editor 19 October 2026. Supplier demonstration and delivery are distinguished from customer operating use."}], "date": "2026-08-20", "publisher": "NOV", "source_type": "Supplier original announcement", "source_id": "WATCH19EX-NOV-AURA", "source_title": "Aura moves remote crane operations from concept to reality", "url": "https://www.nov.com/news/aura-moves-remote-crane-operations-from-concept-to-reality", "locator": "Remote-access paragraph; Kristiansand/Søgne test; first delivery; two planned deliveries; phased customer adoption", "excerpt": "One fully remote-operated crane has already been delivered"}, {"id": "W19EX-OCEANEERING", "record": null, "related_ids": [], "company": "Oceaneering", "title": "One survey vessel brings several offshore jobs together", "department": "Geoscience & subsurface", "status": "Supplier reports campaign started; completion unverified", "summary": "Oceaneering says an Ocean Intervention II survey campaign began in April 2026 for an unnamed international operator offshore Trinidad and the U.S. Gulf.", "mechanism": "The scope combines geophysical and geotechnical work with surface-vessel-supported autonomous underwater surveys. U.S. Gulf pipeline inspections include non-contact cathodic-protection measurement.", "measure": "The supplier says combining work on one vessel and crew can reduce mobilizations, exposure and emissions. No measured savings, baseline or net ROI disclosed.", "operating_period": "Announcement: 30 June 2026. About six months of work expected, including roughly four in Trinidad and two in the U.S. Gulf; not verified completed durations.", "limit": "Autonomous survey is not proof of ML or LLM use. The vessel still has a crew. Operator identity, human supervision, intervention rules, achieved campaign length and comparative performance are undisclosed.", "question": "Does combining survey tasks cut total vessel days once weather, rework and supervision are included?", "sources": [{"id": "WATCH19EX-OCEANEERING", "title": "Oceaneering Secures Multi-Discipline Cross-Regional Offshore Survey Contracts for Ocean Intervention II", "url": "https://www.oceaneering.com/oceaneering-secures-multi-discipline-cross-regional-offshore-survey-contracts-for-ocean-intervention-ii/", "date": "2026-06-30", "publisher": "Oceaneering", "locator": "June 30 release: opening contract/program paragraphs; Trinidad/U.S. Gulf scope; simultaneous operations; supplier quotation; 2025 vessel upgrades", "excerpt": "The program commenced in April 2026", "access": "Original supplier release read by editor 19 October 2026. Begun campaign is distinguished from planned scopes, durations and unverified completion."}], "date": "2026-06-30", "publisher": "Oceaneering", "source_type": "Supplier original announcement", "source_id": "WATCH19EX-OCEANEERING", "source_title": "Oceaneering Secures Multi-Discipline Cross-Regional Offshore Survey Contracts for Ocean Intervention II", "url": "https://www.oceaneering.com/oceaneering-secures-multi-discipline-cross-regional-offshore-survey-contracts-for-ocean-intervention-ii/", "locator": "June 30 release: opening contract/program paragraphs; Trinidad/U.S. Gulf scope; simultaneous operations; supplier quotation; 2025 vessel upgrades", "excerpt": "The program commenced in April 2026"}]}, "watch_2026_10_19_pm": {"as_of": "2026-10-19", "run": "2026-10-19-pm", "label": "19 October 2026 / Evening research", "headline": "What has to be true before you rely on an AI answer?", "intro": "Two useful readings address different questions: whether a model works beyond its training data, and who can approve a change that affects the money.", "count_note": "Two supplementary readings added to the research library. The main register remains at 295 profiles, including 209 company-adoption profiles; deployment and counted-result totals are unchanged.", "method": "Three bounded Luna tracks, then editor checks of accessible original sources and full-corpus deduplication. No post-cutoff announcement or new named abandoned deployment verified. These are earlier missed readings, not an exhaustive market search.", "featured_count": 2, "archive_label": "More company context and earlier research", "findings": [{"id": "W19PM-METERING", "record": null, "related_ids": [], "company": "SPE / JPT", "title": "An AI finding is not permission to change an invoice", "department": "Finance, investment & strategy", "status": "Expert guidance; no reported deployment", "summary": "SPE’s flow-measurement editorial describes AI support for investigating discrepancies in the quantities used for commercial settlement.", "mechanism": "Rules can flag frozen signals; ML can prioritize anomalies; language tools can organize procedures and evidence. Settlement changes still need validated calculations, preserved records and approval.", "measure": "No deployment count, recovered revenue or ROI. Measurement uncertainty multiplied by volume and price is not a demonstrated loss.", "operating_period": "Published 12 October 2026. Illustrative workflows; no named asset, operating period or installed system.", "limit": "Author affiliations include operators and suppliers, but this is not an ExxonMobil or service-company deployment announcement. Shared inputs can make checks agree while retaining the same bias. No evidence that AI changed billed quantities.", "question": "Can an investigator reproduce the correction and see who approved it?", "sources": [{"id": "WATCH19PM-METERING", "title": "From Measurement to Commercial Confidence: An Operator’s View of Custody Transfer—and a Responsible Role for AI", "url": "https://jpt.spe.org/from-measurement-to-commercial-confidence-an-operators-view-of-custody-transfer-and-a-responsible-role-for-ai", "date": "2026-10-12", "publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "locator": "Meter Accuracy Is Not System Credibility; Custody Transfer is an Organizational System; Automation and AI Can Assist, but Accountability Cannot Be Automated; author biographies", "excerpt": "an unvalidated output cannot independently change a billed quantity", "access": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes."}], "date": "2026-10-12", "publisher": "Society of Petroleum Engineers / Journal of Petroleum Technology", "source_type": "Technical-section editorial", "source_id": "WATCH19PM-METERING", "source_title": "From Measurement to Commercial Confidence: An Operator’s View of Custody Transfer—and a Responsible Role for AI", "url": "https://jpt.spe.org/from-measurement-to-commercial-confidence-an-operators-view-of-custody-transfer-and-a-responsible-role-for-ai", "locator": "Meter Accuracy Is Not System Credibility; Custody Transfer is an Organizational System; Automation and AI Can Assist, but Accountability Cannot Be Automated; author biographies", "excerpt": "an unvalidated output cannot independently change a billed quantity"}, {"id": "W19PM-STIMULATION", "record": null, "related_ids": [], "company": "Discover Artificial Intelligence", "title": "A model that fits one field may not travel well", "department": "Drilling & completions", "status": "Literature review; not a deployment census", "summary": "A stimulation-AI review highlights weak external testing, small datasets and inconsistent reporting in the published research it examines.", "mechanism": "The authors discuss validation on independent fields, time-series leakage, physical constraints and integration with field systems. Their readiness assessments are judgments about applications, not audited operator rankings.", "measure": "The quantitative subset contains 23 studies, not 23 deployments. No independently established project return; broad cost and benefit estimates are not admitted as benchmarks.", "operating_period": "Published 15 October 2026. Literature focus: 2020–2025, with earlier foundational studies; much of the underlying work predates the rolling window.", "limit": "Review conclusions depend on heterogeneous studies and incomplete reporting. We have not re-audited every underlying experiment. No named abandonment or industry failure rate; performance averages cannot rank companies or predict another field’s results.", "question": "Was the final model tested on a different field, with tuning kept separate from the test?", "sources": [{"id": "WATCH19PM-STIMULATION", "title": "A systematic and critical review of artificial intelligence applications for enhancing well stimulation operations and sustainable reservoir management", "url": "https://link.springer.com/article/10.1007/s44163-026-02455-x", "date": "2026-10-15", "publisher": "Discover Artificial Intelligence / Springer Nature", "locator": "Abstract; §§2.2–2.3 review selection/quality; §5.8 heterogeneity; §§6.1.1–6.1.5 validation, integration, readiness and economics; §7.4 denominators", "excerpt": "Field-scale and cross-field validation are limited", "access": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes."}], "date": "2026-10-15", "publisher": "Discover Artificial Intelligence / Springer Nature", "source_type": "Academic literature review", "source_id": "WATCH19PM-STIMULATION", "source_title": "A systematic and critical review of artificial intelligence applications for enhancing well stimulation operations and sustainable reservoir management", "url": "https://link.springer.com/article/10.1007/s44163-026-02455-x", "locator": "Abstract; §§2.2–2.3 review selection/quality; §5.8 heterogeneity; §§6.1.1–6.1.5 validation, integration, readiness and economics; §7.4 denominators", "excerpt": "Field-scale and cross-field validation are limited"}]}, "watch_2026_10_20_am": {"as_of": "2026-10-20", "run": "2026-10-20-am", "label": "20 October 2026 / Morning research", "headline": "What Oil India wants to connect.", "intro": "We could now read the company’s detailed DRIVE 2.0 plan. It spans wells, engineering, maintenance and corporate work; the page does not establish what is live.", "count_note": "One planning-context note added, with a department-by-department breakdown. The main register remains at 295 profiles, including 209 company-adoption profiles; deployment and counted-result totals are unchanged.", "method": "Three Luna tracks plus editor source checks and full-corpus deduplication. A previously inaccessible issuer page is now verified through its public canonical path. No new launch date or abandoned deployment established. Coverage remains bounded.", "featured_count": 1, "archive_label": "More company context and earlier research", "findings": [{"id": "W20AM-OILINDIA", "record": null, "related_ids": [], "company": "Oil India", "title": "Oil India’s plan connects field operations with corporate systems", "department": "IT, data & enterprise transformation", "status": "Programme being undertaken; project completion unverified", "summary": "Oil India describes ten DRIVE 2.0 workstreams, including a control centre, a digital oilfield covering 380 producing wells, subsurface agents, maintenance and corporate workflows.", "mechanism": "The proposed systems would connect operating data with engineering and business records. Rig-planning approvals and safety notifications to designated people are described; AI decision authority and safeguards are not.", "measure": "The 380 wells are planned scope, not a verified connected population. No measured uplift, savings, implementation cost or net ROI.", "operating_period": "Publication and DRIVE 2.0 start dates unknown. Header: “11/09/2026 09:35 AM”; footer: “20/10/2026 11:28 AM”. These update strings do not date deployment. The 2019 start belongs to earlier DRIVE.", "limit": "One undated issuer plan, now accessible after an earlier blocked route. No completion dates, live user population or model providers established. Ten workstreams are not ten operating AI deployments.", "question": "Which parts have been commissioned, and who can approve or override their recommendations?", "sources": [{"id": "WATCH20AM-OILINDIA", "title": "Digitalization in OIL — DRIVE 2.0 programme", "url": "https://www.oil-india.com/digitalfootprint", "date": "Undated", "publisher": "Oil India Limited", "locator": "Digitalization in OIL; DRIVE 2.0 §§1–10; §2 380-well scope; §4 approvals; §7 notifications; header and footer update strings", "excerpt": "10 digital initiatives are being taken up under Project DRIVE 2.0", "access": "Original issuer page read by editor 20 October 2026. Public /digitalfootprint path accessible; earlier /index.php/digitalfootprint attempt was blocked. Header and footer update strings are not verified publication or launch dates."}], "date": "Undated", "publisher": "Oil India Limited", "source_type": "Issuer programme overview; undated", "source_id": "WATCH20AM-OILINDIA", "source_title": "Digitalization in OIL — DRIVE 2.0 programme", "url": "https://www.oil-india.com/digitalfootprint", "locator": "Digitalization in OIL; DRIVE 2.0 §§1–10; §2 380-well scope; §4 approvals; §7 notifications; header and footer update strings", "excerpt": "10 digital initiatives are being taken up under Project DRIVE 2.0", "technology_note": "W20AM-OILINDIA", "components": [{"title": "Operations data", "department": "IT, data & enterprise transformation", "scope": "A shared control centre and OT data lake; corporate, field, pipeline and Rajasthan locations.", "locator": "DRIVE 2.0 §1", "stage": "Proposed scope; completion unverified"}, {"title": "Connected wells", "department": "Production & field operations", "scope": "Proposed sensors on 380 producing wells, remote monitoring/control, lift optimization, integrity models and virtual flow measurement.", "locator": "DRIVE 2.0 §2", "stage": "Proposed scope; completion unverified"}, {"title": "Subsurface assistants", "department": "Exploration & geoscience", "scope": "Proposed GenAI/agent tools linking subsurface and operating datasets for forecasts and drilling/workover recommendations.", "locator": "DRIVE 2.0 §3", "stage": "Proposed scope; completion unverified"}, {"title": "Drilling and rig logistics", "department": "Drilling & completions", "scope": "Rig scheduling, readiness/approval workflows and bit-selection support using operating and ERP data.", "locator": "DRIVE 2.0 §4(a)–(b)", "stage": "Proposed scope; completion unverified"}, {"title": "Surface-facility twins", "department": "Refining, LNG & process operations", "scope": "Planned models at OCS Makum, FGGS Madhuban and GCS Hebeda, with engineering records and historian/SAP links.", "locator": "DRIVE 2.0 §5", "stage": "Proposed scope; completion unverified"}, {"title": "Equipment maintenance", "department": "Maintenance & reliability", "scope": "An asset registry and lifecycle/maintenance workflows using SAP S/4HANA; proposed sensor-based failure warnings.", "locator": "DRIVE 2.0 §6", "stage": "Proposed scope; completion unverified"}, {"title": "Camera safety alerts", "department": "HSE & emissions", "scope": "Existing camera feeds and edge processing would flag PPE, access and other hazards for designated people.", "locator": "DRIVE 2.0 §7", "stage": "Proposed scope; completion unverified"}, {"title": "Project control", "department": "Finance, investment & strategy", "scope": "Forecasts of project dates, cost overruns and bottlenecks, with scenario views and audit trails.", "locator": "DRIVE 2.0 §8", "stage": "Proposed scope; completion unverified"}, {"title": "Purchasing", "department": "Supply chain & logistics", "scope": "Requisition-to-order automation, bid support and procurement analysis linked to SAP S/4HANA and GeM.", "locator": "DRIVE 2.0 §9", "stage": "Proposed scope; completion unverified"}, {"title": "Workforce tools", "department": "Workforce & learning", "scope": "A proposed HR system joining onboarding, workforce analytics/chatbot, access, time, benefits, training and contractors.", "locator": "DRIVE 2.0 §10", "stage": "Proposed scope; completion unverified"}]}]}, "watch_2026_10_20_update": {"as_of": "2026-10-20", "run": "2026-10-20-update", "label": "20 October 2026 / Evening source review", "headline": "What is running, what is planned, and what customers actually want.", "intro": "Atlas’s driverless trucks offer an operating example. bp’s appraisal contract is still a plan. SPE’s experience explains why trusted information can matter more than another AI product.", "count_note": "Three source-linked additions to the research notes; the LOGIX history is expanded. The main register remains at 295 profiles, including 209 company-adoption profiles. These notes add no main-register deployment or counted-result profiles.", "method": "Editor review of complete original issuer pages and SPE’s own account, with a clearly labeled syndicated issuer release where the distributor page timed out. Existing initiatives were checked for overlap. This targeted check is not exhaustive.", "featured_count": 3, "archive_label": "Earlier company context and research", "findings": [{"id": "W20UP-KODIAK", "record": null, "related_ids": [], "related_companies": ["Kodiak AI"], "company": "Atlas Energy Solutions", "title": "A driverless truck stayed available for 64 days", "department": "Supply chain & logistics", "status": "Operating industrial deployment; supplier-reported availability", "summary": "Kodiak reports a 64-day uptime streak for one driverless truck in Atlas’s Permian Basin fleet, plus five trucks with streaks above 45 days.", "mechanism": "The truck and autonomy software stayed powered on and available for service. This is physical driving automation, not a language-model assistant. No one was in the cab; remote support and intervention boundaries are not quantified.", "measure": "As of June 30: 35 Atlas-owned and operated driverless trucks and over 40,000 cumulative paid driverless hours. These are supplier-reported fleet figures, not the uptime of every truck. No fleet availability rate, savings or net ROI.", "operating_period": "Announced October 20, 2026. The 64-day streak ran June 30–September 1. The fleet and paid-hours figures are as of June 30; 100 trucks by mid-2027 and a year-end highway launch remain plans.", "limit": "Available does not mean continuously driving or hauling. No mileage, dispatch utilization, collision/exposure comparison or cost denominator disclosed. Kodiak AI is separate from Kodiak Gas Services (AI-193).", "question": "How much paid hauling did those available hours produce, and what did remote support, maintenance and autonomy cost per load?", "sources": [{"id": "WATCH20UP-KODIAK", "title": "Kodiak Driver Sets New Uptime Record By Running 64 Consecutive Days on an Unsupervised Driverless Truck", "url": "https://kodiak.ai/news/kodiak-driver-uptime-record", "date": "2026-10-20", "publisher": "Kodiak AI", "locator": "Opening uptime definition and dates; Atlas fleet as-of paragraph; future fleet and highway plans", "excerpt": "running and available for service", "access": "Full original supplier release retrieved (HTTP 200) and read. Supplier-reported availability and fleet figures; no independent reliability, safety or financial audit."}], "date": "2026-10-20", "publisher": "Kodiak AI", "source_type": "Supplier announcement", "source_id": "WATCH20UP-KODIAK", "source_title": "Kodiak Driver Sets New Uptime Record By Running 64 Consecutive Days on an Unsupervised Driverless Truck", "url": "https://kodiak.ai/news/kodiak-driver-uptime-record", "locator": "Opening uptime definition and dates; Atlas fleet as-of paragraph; future fleet and highway plans", "excerpt": "running and available for service", "technology_note": "W20UP-KODIAK", "technology_label": "Read the operating detail"}, {"id": "W20UP-SPE", "record": null, "related_ids": [], "related_companies": ["Halliburton"], "company": "SPE / OnePetro", "title": "Engineers may want the knowledge more than another AI product", "department": "IT, data & enterprise transformation", "status": "Classification agreement announced; separate licensing lesson", "summary": "SPE and i2k announced extended automated classification of OnePetro content. Separately, SPE says customers often wanted its technical content inside tools they already used.", "mechanism": "i2k describes metadata tagging using more than 20 domain knowledge bases. SPE’s original JPT account says EnRG-LLM remains licensable and identifies Halliburton Landmark as its first content-license customer in an undated follow-up paragraph.", "measure": "No measured retrieval accuracy, time saving, token saving, license revenue, usage population or net ROI. General claims about i2k’s geotechnical agents do not establish that SPE uses every advertised workflow.", "operating_period": "Classification release: October 20, 2026. JPT article: September 10 publication header, with an undated following-conference licensing update. Neither dates completed classification rollout.", "limit": "The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model.", "question": "Our reading: make trusted content useful inside existing tools before assuming customers want another assistant. Ask about content rights, retrieval quality and cost per completed task.", "sources": [{"id": "WATCH20UP-SPE", "title": "SPE extends i2k Connect information-classification collaboration", "url": "https://wrex.marketminute.com/article/bizwire-2026-10-20-society-of-petroleum-engineers-spe-partners-with-i2k-connect-for-information-classification", "date": "2026-10-20", "publisher": "i2k Connect / Business Wire; syndicated by MarketMinute", "locator": "Opening announcement; classification paragraph; SPE CEO statement", "excerpt": "The classification is fully automated", "access": "Complete issuer-authored Business Wire release read on an accessible syndication page. Distributor original timed out; identical release also retrieved from AOL. No independent outcome validation. Distributor URL: https://www.businesswire.com/news/home/20261020770911/en/", "original_url": "https://www.businesswire.com/news/home/20261020770911/en/"}, {"id": "WATCH20UP-SPE-JPT", "title": "Industry Executives Offer Practical Advice on How To Use Modern AI Models Wisely", "url": "https://jpt.spe.org/industry-executives-offer-practical-advice-on-how-to-use-modern-ai-models-wisely", "date": "2026-09-10", "publisher": "Society of Petroleum Engineers / JPT", "locator": "SPE’s AI Journey; licensing-demand discussion; following-conference Landmark paragraph; cost and metadata discussion", "excerpt": "EnRG-LLM remains available for licensing", "access": "Full original publisher article retrieved and read 20 October. September 10 publication header; later following-conference paragraph is undated. Direct current text takes precedence over stale search extracts."}], "date": "2026-10-20", "publisher": "i2k Connect / Business Wire; syndicated by MarketMinute", "source_type": "Issuer announcement and original publisher account", "source_id": "WATCH20UP-SPE", "source_title": "SPE extends i2k Connect information-classification collaboration", "url": "https://wrex.marketminute.com/article/bizwire-2026-10-20-society-of-petroleum-engineers-spe-partners-with-i2k-connect-for-information-classification", "locator": "Opening announcement; classification paragraph; SPE CEO statement", "excerpt": "The classification is fully automated", "technology_note": "W20UP-SPE", "technology_label": "Read the operating detail"}, {"id": "W20UP-BP", "record": null, "related_ids": ["AI-016"], "related_companies": ["Halliburton"], "company": "bp", "title": "Bumerangue’s appraisal contract includes drilling automation", "department": "Drilling & completions", "status": "Contract awarded; LOGIX deployment planned", "summary": "Halliburton’s August contract for bp’s Bumerangue appraisal campaign includes LOGIX automation and remote operations offshore Brazil.", "mechanism": "The integrated drilling and evaluation scope connects planning, execution and evaluation. The release says LOGIX will be deployed; it does not identify an LLM, individual control permissions or human override arrangements.", "measure": "No number of wells, completed automated wells, operating comparator, quantified benefit, contract value or net ROI disclosed.", "operating_period": "Original announcement August 25, 2026; checked October 20. This is older evidence newly added to this review, not a launch announced today.", "limit": "A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform.", "question": "Which drilling tasks will run automatically, who can intervene, and how will bp compare outcomes with a credible baseline?", "sources": [{"id": "WATCH20UP-BP", "title": "bp awards Halliburton integrated contract for Bumerangue field appraisal in Brazil", "url": "https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil", "date": "2026-08-25", "publisher": "Halliburton", "locator": "Opening and contract-scope paragraphs; LOGIX sentence", "excerpt": "LOGIX™ automation and remote operations will be deployed", "access": "Full original supplier release retrieved (HTTP 200) and read. Future-tense scope; no completion or measured result established."}], "date": "2026-08-25", "publisher": "Halliburton", "source_type": "Supplier announcement", "source_id": "WATCH20UP-BP", "source_title": "bp awards Halliburton integrated contract for Bumerangue field appraisal in Brazil", "url": "https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil", "locator": "Opening and contract-scope paragraphs; LOGIX sentence", "excerpt": "LOGIX™ automation and remote operations will be deployed", "technology_note": "W20UP-BP", "technology_label": "Read the operating detail"}]}, "watch_2026_10_20_pm": {"as_of": "2026-10-20", "run": "2026-10-20-pm", "label": "20 October 2026 / Scheduled research", "headline": "What changes when a recommendation can reach the equipment?", "intro": "A missed Permian field account explains how engineers moved to supervising exceptions. Ecopetrol adds dated operating scale to its energy platform, without proving an AI saving.", "count_note": "One named operator field profile added and one existing platform history expanded. The main register now has 296 profiles, including 210 company-adoption profiles. The additional reported-result profile is not independently verified net ROI.", "method": "Three bounded Luna tracks, six search clauses each, followed by editor checks of accessible originals and corpus deduplication. Publication dates, operating periods and future scope remain separate. Coverage is not exhaustive.", "featured_count": 2, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W20PM-USEDC", "company": "U.S. Energy Development Corporation", "title": "Twelve wells moved from recommendations to automatic adjustments", "department": "Production & field operations", "status": "Operating field deployment; operator/vendor-reported results", "summary": "U.S. Energy and ROAM-AI describe automatic ESP adjustments on 12 Permian wells. Engineers define the operating limits and handle exceptions.", "mechanism": "Physics-aware models use live data to adjust setpoints within approved frequency, pressure, temperature and pump-curve limits. The authors say thousands of adjustments were executed automatically. No LLM is identified; exact override and shutdown procedures are undisclosed.", "measure": "The authors report 12.2% higher production against a 90-day pre-activation baseline, and a return exceeding four times initial cost within six months. Full costs and return definitions are absent; this is not verified net ROI. Six months was too early for a definitive pump-failure interval conclusion.", "operating_period": "Started December 2025. July 2026 article; exact publication day unknown. Six-month results, not a year or the operator’s full fleet. This older case was missed in earlier research.", "limit": "Operator/vendor coauthored. No controlled comparator or disclosed adjustment for decline, workovers or other changes. Do not import the article’s separate unnamed 250-plus-well operator results into this 12-well case.", "question": "Ask for the all-in cost and benefit bridge, the baseline calculation and the log of exceptions, overrides and rejected adjustments.", "sources": [{"id": "WATCH20PM-USEDC", "title": "The Bottleneck Isn’t the Model. It’s the Execution.", "url": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "date": "2026-07", "publisher": "The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors", "locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "excerpt": "We activated the system across 12 Permian wells beginning in December 2025.", "access": "Full original publisher article retrieved (HTTP200) and read by editor. Operator/vendor coauthored, not independently validated. Month precision only; vendor’s July20 promotion is not the article’s exact publication date."}], "related_ids": ["AI-296"], "record": "AI-296", "date": "2026-07", "publisher": "The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors", "source_type": "Operator/vendor-coauthored field account", "source_id": "WATCH20PM-USEDC", "source_title": "The Bottleneck Isn’t the Model. It’s the Execution.", "url": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "excerpt": "We activated the system across 12 Permian wells beginning in December 2025.", "technology_note": "W20PM-USEDC", "technology_label": "Read the operating detail"}, {"id": "W20PM-CADE", "company": "Ecopetrol", "title": "Rubiales adds a dated operating footprint to CADE", "department": "IT, data & enterprise transformation", "status": "CADE operating coverage; specific AI-model commissioning unverified", "summary": "Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid.", "mechanism": "CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority.", "measure": "The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI.", "operating_period": "Issuer announcement May 4, 2026; Rubiales coverage reported from March 2026. Older evidence newly checked, not a new October launch.", "limit": "No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population.", "question": "Which decisions are actually automated, and what changes in delivered electricity cost after all platform and operating costs?", "sources": [{"id": "WATCH20PM-CADE", "title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "url": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "date": "2026-05-04", "publisher": "Ecopetrol", "locator": "May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs", "excerpt": "En el campo Rubiales, desde marzo de 2026, se gestionan 173 MW de energía", "access": "Full original operator page retrieved (HTTP200) and read by editor. Coverage and plans remain company-reported; no independent cost or performance validation."}], "related_ids": ["AI-148"], "record": "AI-148", "date": "2026-05-04", "publisher": "Ecopetrol", "source_type": "Operator announcement", "source_id": "WATCH20PM-CADE", "source_title": "Grupo Ecopetrol inaugura centro para optimizar su gestión energética", "url": "https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica", "locator": "May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs", "excerpt": "En el campo Rubiales, desde marzo de 2026, se gestionan 173 MW de energía", "technology_note": "W20PM-CADE", "technology_label": "Read the operating detail"}]}, "watch_2026_10_21_am": {"as_of": "2026-10-21", "run": "2026-10-21-am", "label": "21 October 2026 / Scheduled research", "headline": "What is actually inside the company’s AI engine?", "intro": "PTTEP describes its internal engine, an adjacent power-site case shows how operators confirm diagnoses, and Quantum separates a solver improvement from its assistant.", "count_note": "One announced engine profile added; two source reviews add operating detail. There are now 297 profiles, including 211 company-adoption profiles. Operating and reported-result counts are unchanged.", "method": "Three Luna tracks searched six clauses each, followed by editor review of accessible originals and corpus checks. These are missed historical or undated sources; no qualifying new post-cutoff announcement was retained. Coverage is not exhaustive.", "featured_count": 3, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W21AM-PTTEP", "company": "PTTEP", "record": "AI-297", "related_ids": ["AI-297", "AI-067", "AI-245"], "title": "PTTEP describes an in-house engine for specialist agents", "department": "IT, data & enterprise transformation", "status": "Engine launch described; eight-agent rollout unverified", "summary": "PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data.", "mechanism": "The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live.", "measure": "No X.brain user count, adoption rate or measured return is disclosed. The same article separately reports LogSight interpretation falling from two days to one hour and DORA drilling savings described only as “millions of US dollars annually”. Those are product-specific claims, not demonstrated X.brain results.", "operating_period": "Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement.", "limit": "No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons.", "question": "Ask which agents people actually use, what data they can reach and which decisions still need a named human approver.", "sources": [{"id": "WATCH21AM-PTTEP", "title": "PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy", "url": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "date": "2025-11-03", "publisher": "PTTEP", "locator": "Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples", "excerpt": "PTTEP has launched its in-house Enterprise AI Engine, “X.brain,”", "access": "Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation."}], "date": "2025-11-03", "source_type": "Operator technology feature", "technology_note": "W21AM-PTTEP", "technology_label": "Read the operating detail", "source_id": "WATCH21AM-PTTEP", "source_title": "PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy", "url": "https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy", "publisher": "PTTEP", "locator": "Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples", "excerpt": "PTTEP has launched its in-house Enterprise AI Engine, “X.brain,”", "components_label": "Read the two operating examples separately", "components_note": "The same feature describes these tools. It does not establish that either runs on X.brain, or that their reported benefits came from the engine.", "components": [{"title": "DORA: anticipate stuck pipe", "department": "Drilling & completions", "scope": "Drilling Operation Real-Time Analyses combines machine learning with physics-based simulations to predict stuck-pipe risk and support parameter adjustments. The issuer claims annual savings of millions of US dollars, without an exact figure, well population, comparator or full costs. Automatic control and human approval rules are not disclosed.", "locator": "PTTEP feature: DORA example", "stage": "Operating use implied by issuer; rollout and evaluation population unknown"}, {"title": "LogSight: interpret well logs", "department": "Reservoir engineering & field development", "scope": "AI supports petrophysical log interpretation. PTTEP reports a reduction from two days to one hour, without sample size, quality comparison, calendar period, human sign-off rule or full costs.", "locator": "PTTEP feature: LogSight example", "stage": "Reported issuer comparison; deployment population unknown"}]}, {"id": "W21AM-MNVR", "company": "Energy Solutions / MNVR AI", "record": null, "related_ids": [], "title": "Five diagnosed equipment faults, with people confirming the fixes", "department": "Maintenance & reliability", "status": "Vendor/partner-reported operating case; adjacent gas-to-power site", "summary": "An undated MNVR and Energy Solutions case describes ODIE supporting seven gas generators at the Muskrat microgrid in North Dakota’s Bakken Basin.", "mechanism": "Across six weeks, three cases came from agent alerts and two from operator questions. The authors say all five diagnoses were confirmed on inspection. Confirmed diagnoses and operator corrections are retained for future use. The product page keeps decisions with operators; automatic equipment control is not established.", "measure": "The authors jointly estimate $37,015 in avoided costs across the five selected faults. These are counterfactual estimates, not measured cash savings or net ROI. A $25,500 relay case includes the hypothesized cost of replacing the wrong controller; do not add that controller’s price again.", "operating_period": "Publication date unknown; checked 21 October 2026. Six-week calendar window unknown. PDF filename is not a verified date. Adjacent gas-to-power operations, not an upstream adoption count.", "limit": "Selected equipment faults are not failed AI deployments. No total alert population, missed-fault count, false-positive rate, control group or full system cost. Some repairs were scheduled or could be planned; not all five were completed. Roughly 5,200 signals means data signals, not assets.", "question": "Can the team show every alert, inspection and resulting work order—including alerts that turned out to be wrong?", "sources": [{"id": "WATCH21AM-MNVR-CASE", "title": "Five Failures Across Five Systems", "url": "https://www.mnvrai.ai/s/MNVR-Customer-Story-Energy-Solutions-5-Examples-09222026.pdf", "date": "Unknown", "publisher": "MNVR AI / Energy Solutions", "locator": "PDF pp. 1–3: seven generators; five fault cases; estimated costs; six-week period and retained diagnoses", "excerpt": "Over six weeks, ODIE identified five separate developing failures", "access": "Original three-page vendor/partner PDF retrieved HTTP200, full text read and pages 1–2 visually inspected. No publication date; filename is not date evidence."}, {"id": "WATCH21AM-MNVR-PRODUCT", "title": "ODIE — Operations Diagnostic Intelligence Engine", "url": "https://www.mnvrai.ai/odie", "date": "Unknown", "publisher": "MNVR AI", "locator": "Human in the loop / operators make decisions section", "excerpt": "ODIE informs decisions. Operators make them.", "access": "Original product page retrieved HTTP200 and read by editor. Undated product description; not a field performance validation."}], "date": "Unknown", "source_type": "Vendor/maintenance-partner customer case", "technology_note": "W21AM-MNVR", "technology_label": "Read the operating detail", "source_id": "WATCH21AM-MNVR-CASE", "source_title": "Five Failures Across Five Systems", "url": "https://www.mnvrai.ai/s/MNVR-Customer-Story-Energy-Solutions-5-Examples-09222026.pdf", "publisher": "MNVR AI / Energy Solutions", "locator": "PDF pp. 1–3: seven generators; five fault cases; estimated costs; six-week period and retained diagnoses", "excerpt": "Over six weeks, ODIE identified five separate developing failures", "components_label": "Read the five equipment faults", "components_note": "These selected equipment faults illustrate diagnosis and maintenance follow-through. They are not failures of the AI system, and the estimated avoided costs are not measured cash savings.", "components": [{"title": "Control relay", "department": "Maintenance & reliability", "scope": "A gas-pressure trip left a generator motoring and drawing grid power for over nine hours. The authors say ODIE identified a relay rather than the initially suspected $18,000 controller. One hour of labor and a $100 relay are described; $25,500 estimated avoidance already includes the hypothesized wrong-controller path.", "locator": "Customer PDF p. 1, case 1", "stage": "Relay correction described"}, {"title": "Ignition transformer", "department": "Maintenance & reliability", "scope": "An agent flagged degradation; operators confirmed it and replaced the transformer during scheduled maintenance. Estimated cost avoided: $4,641, without a measured counterfactual.", "locator": "Customer PDF p. 1, case 2", "stage": "Replacement described"}, {"title": "Valve recession / detonation", "department": "Maintenance & reliability", "scope": "An alert prompted operator inspection. A head replacement could then be planned. Estimated avoidance: $4,000; the account does not confirm the planned replacement was completed.", "locator": "Customer PDF p. 2, case 3", "stage": "Inspection confirmed; replacement could be planned"}, {"title": "Silent NOx sensor", "department": "HSE & emissions", "scope": "A sensor fault reportedly absent from traditional monitoring was identified. Replacement was proposed for the next convenient opportunity. Estimated avoidance: $1,000; completion is unverified.", "locator": "Customer PDF p. 2, case 4", "stage": "Diagnosis described; replacement pending/unverified"}, {"title": "Fouling gas filter", "department": "Maintenance & reliability", "scope": "The team checked and confirmed a restriction and changed the filter during scheduled maintenance. Estimated avoidance: $1,874, based on the authors’ hypothetical failure progression.", "locator": "Customer PDF p. 2, case 5", "stage": "Scheduled replacement described"}]}, {"id": "W21AM-QUANTUM", "company": "Quantum Capital Group", "record": "AI-054", "related_ids": ["AI-054"], "title": "Quantum’s planning gain includes a separate algorithm improvement", "department": "Reservoir engineering & field development", "status": "Existing operating planning workflow; reported updated-version comparison", "summary": "Microsoft says algorithm changes in an updated version of Quantum’s field-planning tool reduced solve time by 50–65%.", "mechanism": "The integer-linear-programming optimizer selects feasible pads and wells. A Copilot Studio agent explains outputs. The numerical improvement belongs to the optimization algorithm; the source does not attribute it to the language model.", "measure": "50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison.", "operating_period": "Microsoft story dated 30 April 2026; later operator identification remains QB Energy from Quantum’s 2026 report. This is an enrichment of AI-054, not a new deployment.", "limit": "Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model.", "question": "Separate the solver, data preparation and assistant when testing which change actually improves the planning decision.", "sources": [{"id": "OCT5-S032", "title": "Field development planning at Quantum Capital Group", "url": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "date": "2026-04-30", "publisher": "Microsoft", "locator": "Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph", "excerpt": "algorithmic improvements have reduced solve times by 50 to 65%", "access": "Full original Microsoft customer story read again on 21 October. Updated-version comparison is vendor-reported; no independent validation."}], "date": "2026-04-30", "source_type": "Microsoft customer story", "technology_note": "W21AM-QUANTUM", "technology_label": "Read the operating detail", "source_id": "OCT5-S032", "source_title": "Field development planning at Quantum Capital Group", "url": "https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service", "publisher": "Microsoft", "locator": "Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph", "excerpt": "algorithmic improvements have reduced solve times by 50 to 65%"}]}, "watch_2026_10_21_extra": {"as_of": "2026-10-21", "run": "2026-10-21-extra", "label": "21 October 2026 / Additional research", "headline": "An internal assistant becomes easier to assess when the company describes daily use.", "intro": "An issuer-hosted Devon call closes an evidence gap. SLB adds a named automation example, while PETRONAS shows why field models should not automatically be called AI.", "count_note": "The 297-profile register is unchanged in size. ChatDVN moves from pending review to company adoption: 212 company-use profiles, 146 described as operating or scaled, and 61 with a counted reported result. No new result or net ROI is credited.", "method": "Three Luna researchers used six search clauses each. The editor checked originals, downloaded and inspected the two relevant PDF pages, and deduplicated the findings. These are older disclosures; this is not an exhaustive review or a set of new October launches.", "featured_count": 3, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W21EX-DEVON", "company": "Devon Energy", "record": "AI-086", "related_ids": ["AI-086"], "title": "Devon confirms ChatDVN is part of daily work", "department": "IT, data & enterprise transformation", "status": "Operating internal workflow; issuer-described daily use", "summary": "Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data.", "mechanism": "The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006.", "measure": "No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings.", "operating_period": "Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch.", "limit": "Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status.", "question": "Which jobs do employees finish through ChatDVN, and does the full review process improve time, accuracy and follow-through?", "sources": [{"id": "FU001", "title": "Devon Q1 2026 earnings call — internal ChatDVN workflow", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "date": "2026-05-06", "publisher": "Devon Energy / FactSet CallStreet", "locator": "PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI", "excerpt": "a fully-firewalled internal tool called ChatDVN", "access": "Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation."}], "date": "2026-05-06", "source_type": "Issuer-hosted earnings transcript", "asset": "Internal company-data workflow; user and asset populations undisclosed.", "ai_role": "Internal AI assistant and company-data access; coding and operations-data examples described separately. No external model provider disclosed.", "partner": "Not disclosed", "source_id": "FU001", "source_title": "Devon Q1 2026 earnings call — internal ChatDVN workflow", "url": "https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf", "publisher": "Devon Energy / FactSet CallStreet", "locator": "PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI", "excerpt": "a fully-firewalled internal tool called ChatDVN", "technology_note": "W21EX-DEVON", "technology_label": "Read the operating detail"}, {"id": "W21EX-SLB", "company": "SLB", "record": null, "related_ids": [], "related_companies": ["Shell"], "title": "Automated cementing has moved beyond its first job", "department": "Drilling & completions", "status": "Supplier reports commercial operating deployments", "summary": "SLB says its cementing technology had passed 100 remote and automated jobs by July 2026, following first-quarter commercialization.", "mechanism": "The release also names an automated dual-slurry operation on Shell’s Velox-1 in Egypt. The 100-plus figure combines remote and automated jobs; it is not 100 fully autonomous wells or a Shell-only fleet count.", "measure": "Job count shows reported operating reach. The release gives no controlled cycle-time, cement-quality, safety or cash-return comparison for this cohort.", "operating_period": "Issuer release 24 July 2026; first-quarter 2026 commercialization described. Individual job dates and the remote/automated split are undisclosed.", "limit": "Automation and remote operation do not establish machine learning or LLM use. Human supervision, control envelopes, fallback and override rules are not specified. The same release’s AI Production Studio is a pre-field demonstration environment, not proof that those capabilities ran on Shell’s well.", "question": "Which steps execute automatically, what still needs the crew, and how does job quality compare under similar well conditions?", "sources": [{"id": "WATCH21EX-SLB", "title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "date": "2026-07-24", "publisher": "SLB", "locator": "Core / Technology Highlights; automated cementing paragraph and Shell Velox-1 follow-up", "excerpt": "surpassed 100 remote and automated cementing jobs", "access": "Original issuer release opened and relevant technology-highlight paragraphs read by editor. Remote and automated categories are not broken out; no AI method specified."}], "date": "2026-07-24", "source_type": "Issuer earnings release", "asset": "Reported global 100-plus combined-job cohort; named dual-slurry case at Shell Velox-1, Egypt.", "ai_role": "Cementing automation and remote operation; AI/ML/LLM attribution unestablished.", "partner": "Shell named for Velox-1, not the entire cohort", "source_id": "WATCH21EX-SLB", "source_title": "SLB Announces Second-Quarter 2026 Results", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results", "publisher": "SLB", "locator": "Core / Technology Highlights; automated cementing paragraph and Shell Velox-1 follow-up", "excerpt": "surpassed 100 remote and automated cementing jobs", "technology_note": "W21EX-SLB", "technology_label": "Read the operating detail"}, {"id": "W21EX-IAM", "company": "PETRONAS", "record": null, "related_ids": [], "title": "Eight fields use physics-based asset modeling", "department": "Production & field operations", "status": "Issuer reports deployed workflow during 2025", "summary": "PETRONAS reports Integrated Asset Management across eight fields, combining physics-based models with near-real-time data for surveillance and production optimization.", "mechanism": "The report says the workflow helps teams identify production opportunities faster. Its own passage does not attribute AI to IAM. Nearby PROTEAN and ERMAI AI descriptions concern different systems.", "measure": "The issuer says IAM improved production efficiency and value realization in 2025, without giving an amount, baseline, comparison group, attribution method or full costs. Eight is a field count, not eight models or platforms.", "operating_period": "Operating period 2025. Source register dates the 2025 integrated report 27 February 2026; exact posting day was not re-audited. Printed p.55 / PDF p.29 newly inspected.", "limit": "No field names, user population, human approval or automatic control permissions are disclosed. The chapter’s broad AI heading cannot establish IAM’s method. Do not transfer PROTEAN’s equipment footprint or ERMAI’s time comparison to IAM.", "question": "What production decisions changed, and what incremental result can be separated from other field improvements?", "sources": [{"id": "RS017", "title": "PETRONAS Integrated Report 2025 — IAM", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "date": "2026-02-27", "publisher": "PETRONAS", "locator": "PDF p.29 / printed p.55; Turning Data into Production Upside", "excerpt": "IAM has been deployed across eight fields", "access": "Original 10,849,024-byte PDF retrieved HTTP200; editor extracted and visually inspected spread54–55 (PDF29). Publication day retained from existing source register; exact posting day not re-audited in this pass."}], "date": "2026-02-27", "source_type": "Operator integrated report", "asset": "Eight unnamed fields; country/basin, well and equipment populations undisclosed.", "ai_role": "Physics-based models and near-real-time data; this passage does not establish AI use.", "partner": "Not disclosed", "source_id": "RS017", "source_title": "PETRONAS Integrated Report 2025 — IAM", "url": "https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf", "publisher": "PETRONAS", "locator": "PDF p.29 / printed p.55; Turning Data into Production Upside", "excerpt": "IAM has been deployed across eight fields", "technology_note": "W21EX-IAM", "technology_label": "Read the operating detail"}]}, "watch_2026_10_21_pm": {"as_of": "2026-10-21", "run": "2026-10-21-pm", "label": "21 October 2026 / Evening research", "headline": "Shell’s pump model needed an operator’s knowledge to move forward.", "intro": "An undated Shell case shows where analysis stalled. Quantum adds detail on its data foundation, and a new IEA report helps separate gas-power proposals from committed demand.", "count_note": "Three source-reviewed notes; one existing sponsor dossier expanded and one industry report added. The register remains at 297 profiles, including 212 company-use profiles, 146 operating or scaled and 61 with a counted reported result. No new adoption, result or net ROI is credited.", "method": "Three bounded Luna tracks used six search clauses each; editor original-source checks and corpus deduplication followed. No qualifying post-cutoff disclosure was retained. Dates, source populations and missing evidence remain explicit; coverage is not exhaustive.", "featured_count": 3, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W21PM-SHELL", "company": "Shell", "record": null, "related_ids": [], "title": "Shell’s pump analysis stalled until an operator joined the team", "department": "Maintenance & reliability", "status": "Retrospective precursor analysis; operational deployment unknown", "summary": "Shell describes an initially unsuccessful search for pump-disruption precursors at Perdido. A field operator helped the data team refine its code and find a chemical signature.", "mechanism": "The team first drew a blank, later plateaued and considered stopping. The operator supplied equipment knowledge. Shell then reports a signature before 70% of disruptions, with one example about two days beforehand.", "measure": "The 70% is a reported precursor association in the analyzed events, not demonstrated live prediction accuracy. Event count, false alarms, held-out validation, intervention results, uptime benefit and full costs are undisclosed.", "operating_period": "Undated webpage and video; operating and evaluation dates unknown. Already mentioned in the coverage log, now structured as a readable case. It is not a new 2026 deployment.", "limit": "The transcript does not establish a live alert or control workflow, prevented disruptions or realized savings. The two-day example is not an average warning time. Initial stalled analysis is not proof of an abandoned or failed field deployment.", "question": "Could the team reproduce the result on later events, with all false alerts recorded, and show what operators changed because of it?", "sources": [{"id": "WATCH21PM-SHELL", "title": "The Data Detectives — Solving the 10-year-old problem at Perdido", "url": "https://www.shell.com/what-we-do/digitalisation/artificial-intelligence.html", "date": "Unknown", "publisher": "Shell", "locator": "Artificial intelligence examples / Perdido video transcript; initial blank, plateau, operator input and refined precursor signature", "excerpt": "the refined algorithm identifies the chemical fingerprint ahead of 70 percent of pump disruptions", "access": "Original operator webpage and relevant video transcript read by editor. Page/video publication and operating dates are not established; historical or undated context only."}], "date": "Unknown", "source_type": "Operator video transcript", "technology_note": "W21PM-SHELL", "technology_label": "Read the operating detail", "asset": "Perdido subsea pumps, Gulf of Mexico; pump/event population undisclosed.", "ai_role": "AI-assisted analysis of stored pump data, refined with operator expertise; no LLM or live control established.", "partner": "Not disclosed", "source_id": "WATCH21PM-SHELL", "source_title": "The Data Detectives — Solving the 10-year-old problem at Perdido", "url": "https://www.shell.com/what-we-do/digitalisation/artificial-intelligence.html", "publisher": "Shell", "locator": "Artificial intelligence examples / Perdido video transcript; initial blank, plateau, operator input and refined precursor signature", "excerpt": "the refined algorithm identifies the chemical fingerprint ahead of 70 percent of pump disruptions"}, {"id": "W21PM-DBX", "company": "Quantum Capital Group", "record": "AI-052", "related_ids": ["AI-052"], "title": "Quantum adds a relational layer to its deal-evaluation data", "department": "Finance, investment & strategy", "status": "Vendor describes an adopted sponsor-level data layer", "summary": "Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data.", "mechanism": "A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker.", "measure": "Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns.", "operating_period": "Undated source; adoption and evaluation periods unknown. Extends existing sponsor platform AI-052, separately from QB Energy’s field-planning workflow AI-054.", "limit": "No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution.", "question": "Can the team trace a deal assumption to its source and show whether shared data reduces reconciliation and review work?", "sources": [{"id": "WATCH21PM-DBX", "title": "Transforming energy deal evaluation with Lakebase", "url": "https://www.databricks.com/customers/quantum-capital-group", "date": "Unknown", "publisher": "Databricks", "locator": "Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs", "excerpt": "The firm now runs over 11,000 unique data queries per month", "access": "Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation."}], "date": "Unknown", "source_type": "Vendor customer story", "technology_note": "W21PM-DBX", "technology_label": "Read the operating detail", "asset": "Quantum sponsor-level energy deal evaluation; asset and portfolio-company populations undisclosed.", "ai_role": "Relational and governed data foundation; no AI model or autonomous investment authority established by this source.", "partner": "Databricks", "source_id": "WATCH21PM-DBX", "source_title": "Transforming energy deal evaluation with Lakebase", "url": "https://www.databricks.com/customers/quantum-capital-group", "publisher": "Databricks", "locator": "Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs", "excerpt": "The firm now runs over 11,000 unique data queries per month"}, {"id": "W21PM-IEA", "company": "International Energy Agency", "record": null, "related_ids": [], "report_id": "WATCH21PM-IEA", "title": "A gas-power project queue is not committed gas demand", "department": "Finance, investment & strategy", "status": "Industry policy and project-pipeline research", "summary": "The IEA’s 14 October report explains why data-centre power announcements need to be checked against project maturity, equipment delivery and actual grid access.", "mechanism": "It distinguishes islanded, mixed and temporary bridge-power arrangements. Connection queues can include duplicate or speculative requests, while operating loads often ramp up in stages. A later grid connection can change how onsite generation is used.", "measure": "The report describes roughly 210 GW of planned onsite gas generation globally, around 190 GW at proposal stage and 17 GW under construction. These are rounded project-capacity figures, not additional additive totals, operating load or contracted gas volumes. No residual operating capacity is inferred.", "operating_period": "Published 14 October 2026. Gas-project figure uses IEA analysis of Global Energy Monitor data from August 2026. A distinct October edition, not a new announcement today or an amount to add to the April IEA report.", "limit": "Project announcements do not establish completion, utilization, fuel burn or firm offtake. The report’s abstract covers 34 policy jurisdictions and about 90% of global data-centre capacity; those are policy-study scope measures, not an oil-and-gas AI-adoption sample.", "question": "Before relying on a project’s gas demand, ask for its customer commitment, generation equipment, permits, grid-access terms and phased load schedule.", "sources": [{"id": "WATCH21PM-IEA", "title": "Powering Artificial Intelligence: Strategies for integrating data centres into electricity systems", "url": "https://www.iea.org/reports/powering-artificial-intelligence", "date": "2026-10-14", "publisher": "International Energy Agency", "locator": "Dated landing page; executive summary; PDF pp.6–10, 36–47 and 58–62; gas-generation pipeline on p.60", "excerpt": "Connection queues contain duplicative or speculative projects", "access": "Original IEA landing page, executive summary and relevant report passages read. Date verified on landing page. Policy and project-pipeline analysis, not operator AI adoption or contracted gas demand."}], "date": "2026-10-14", "source_type": "Institutional policy and project-pipeline report", "source_id": "WATCH21PM-IEA", "source_title": "Powering Artificial Intelligence: Strategies for integrating data centres into electricity systems", "url": "https://www.iea.org/reports/powering-artificial-intelligence", "publisher": "International Energy Agency", "locator": "Dated landing page; executive summary; PDF pp.6–10, 36–47 and 58–62; gas-generation pipeline on p.60", "excerpt": "Connection queues contain duplicative or speculative projects"}]}, "watch_2026_10_22_am": {"as_of": "2026-10-22", "run": "2026-10-22-am", "label": "22 October 2026 / Morning research", "headline": "An equipment alert is only useful when someone acts on it.", "intro": "TC Energy’s presentation shows the people and process behind its reliability work. It also separates an existing benefit claim from a new platform’s forecast and future AI features.", "count_note": "One supplementary company case added. The register remains at 297 profiles: 212 company-use profiles, 146 operating or scaled and 61 with a counted reported result. No new adoption, result or net ROI is credited.", "method": "We checked original company and partner material against the existing report. This adds older context, not a new announcement today. Publication dates, operating dates and unverified claims remain separate. Coverage is not exhaustive.", "featured_count": 1, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W22AM-FAST", "company": "TC Energy", "record": null, "related_ids": [], "title": "TC Energy puts people between equipment alerts and field action", "department": "Maintenance & reliability", "status": "Existing advisory process; FAST release completion unverified", "summary": "TC Energy describes an equipment-surveillance process built around specialists and field teams. FAST is its consolidation plan with Accenture and AVEVA, separate from Pipeline IQ and compressor-dispatch optimization.", "mechanism": "Alerts pass through triage, specialist investigation, corrective action and documentation. The proposed platform joins existing tools, AVEVA and AWS. SageMaker model development is marked development-only; language assistants and deeper SAP coordination sit under future work.", "measure": "The presentation credits existing surveillance with over $100m/year in avoided failure costs. FAST’s additional $250m over ten years is a forecast. Neither is measured LLM value or all-in net ROI.", "operating_period": "Undated AVEVA World 2026 presentation, internally dated 29 May. Page 16 schedules Canada for Q3 2026 and US/Mexico for Q4; page 23 describes a first release in under seven months. The 16 October opinion does not resolve completion.", "limit": "The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features.", "question": "What reached production, who acts on an alert, and what remains after false alarms, specialist time and deployment costs?", "sources": [{"id": "WATCH22AM-FAST", "title": "Practical AI in Action: TC Energy’s Path to a Unified, Scalable Asset Reliability Platform", "url": "https://cdn.mediavalet.com/eunl/content/W8TklEW5w0msg_eHg0KrTA/gz5XgM6QhE6XBVDXhH9-pA/Original/Practical%20AI%20in%20Action%3A%20TC%20Energy%27s%20Path%20to%20a%20Unified%2C%20Scalable%20Asset%20Reliability%20Platform.pdf", "date": "Unknown", "publisher": "TC Energy and Accenture / AVEVA World 2026", "locator": "PDF pp.14–16: advisory workflow, upcoming FAST and release roadmap; p.17: architecture; p.20: forecast; p.22: future AI work; p.23: release wording", "excerpt": "FAST is TC Energy’s upcoming Facility Asset Surveillance Technology", "access": "Original 27-page presentation downloaded and read by editor; pp.14,16,22 visually inspected. Deck internally dated 29 May 2026; exact publication date unknown."}, {"id": "WATCH22AM-AVEVA", "title": "How the oil and gas industry is putting data first", "url": "https://www.aveva.com/en/our-industrial-life/type/opinion/how-the-oil-and-gas-industry-is-putting-data-first/", "date": "2026-10-16", "publisher": "AVEVA / Cindy Crow", "locator": "Posted date; Improved reliability through AI monitoring section", "excerpt": "the company saved $100 million of value in failure avoidance per year", "access": "Original vendor opinion read by editor. Named-customer program claims are not independent outcome validation or release confirmation."}], "date": "Unknown", "source_type": "Joint customer conference presentation; vendor opinion", "technology_note": "W22AM-FAST", "technology_label": "Read the operating detail", "asset": "North American gas-pipeline facilities; exact FAST rollout population undisclosed.", "ai_role": "Future AI triage, language assistance and workflow coordination are proposals; operating model-by-model AI scope is undisclosed.", "partner": "Accenture; AVEVA; AWS", "source_id": "WATCH22AM-FAST", "source_title": "Practical AI in Action: TC Energy’s Path to a Unified, Scalable Asset Reliability Platform", "url": "https://cdn.mediavalet.com/eunl/content/W8TklEW5w0msg_eHg0KrTA/gz5XgM6QhE6XBVDXhH9-pA/Original/Practical%20AI%20in%20Action%3A%20TC%20Energy%27s%20Path%20to%20a%20Unified%2C%20Scalable%20Asset%20Reliability%20Platform.pdf", "publisher": "TC Energy and Accenture / AVEVA World 2026", "locator": "PDF pp.14–16: advisory workflow, upcoming FAST and release roadmap; p.17: architecture; p.20: forecast; p.22: future AI work; p.23: release wording", "excerpt": "FAST is TC Energy’s upcoming Facility Asset Surveillance Technology"}]}, "watch_2026_10_22_extra": {"as_of": "2026-10-22", "run": "2026-10-22-extra", "label": "22 October 2026 / Additional research", "headline": "Who checks whether an AI project is worth the money?", "intro": "Diamondback’s hiring plan makes that responsibility explicit. An Italgas voice-agent case shows why beta claims still need careful checks before they become a business case.", "count_note": "Two supplementary notes; no new core initiative or reported-result count. The register remains 297 profiles: 212 company-use, 146 operating or scaled and 61 with a counted reported result.", "method": "We checked original material and the existing register. Hiring plans remain separate from implemented controls, and the gas-distribution example remains an adjacent case. Neither source establishes a new disclosure since the last update. Coverage is not exhaustive.", "featured_count": 2, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W22EX-FANG", "company": "Diamondback Energy", "record": null, "related_ids": [], "title": "Diamondback wants someone to challenge the economics of its AI projects", "department": "Finance, investment & strategy", "status": "Hiring signal; role and framework implementation unverified", "summary": "An official AI Lead – Value and Economics posting calls for someone to test AI business cases and distinguish actual impact from correlation.", "mechanism": "The role would set baselines and comparison methods, challenge benefit assumptions, compare forecasts with realized value, and recommend whether projects continue, scale, change course or stop. It works with Finance, product owners and the AI portfolio lead.", "measure": "No project results, filled-role evidence, implemented measurement framework, cost or return is disclosed.", "operating_period": "The posting was available when checked on 22 October 2026. Its publication date and whether the role has been filled are unverified.", "limit": "A job description establishes intended responsibilities. It does not show that these controls exist today or that any project was stopped. Existing Diamondback operating initiatives retain their own evidence.", "question": "Who can challenge an AI business case, and what evidence would make the company stop funding it?", "sources": [{"id": "WATCH22EX-FANG", "title": "AI Lead – Value and Economics · R100885", "url": "https://diamondbackenergy.wd12.myworkdayjobs.com/DBE/job/Midland-TX/AI-Lead---Value-and-Economics_R100885", "date": "Unknown", "publisher": "Diamondback Energy / official Workday careers", "locator": "Job Responsibilities: valuation framework, baselines/comparison groups, benefits attribution, continue/scale/pivot/stop recommendations. Official public jobPostingInfo payload retrieved.", "excerpt": "recommend whether investments should continue, scale, pivot, or stop", "access": "Editor retrieved and read the full official public Workday job payload. Human-facing page returns no extracted text; search copy corroborates responsibilities. Publication date not verified; metadata startDate is not a verified hire or operating date.", "retrieval_url": "https://diamondbackenergy.wd12.myworkdayjobs.com/wday/cxs/diamondbackenergy/DBE/job/Midland-TX/AI-Lead---Value-and-Economics_R100885", "metadata_start_date": "2026-08-25", "metadata_date_limit": "Workday startDate field; meaning not documented, not treated as publication, employment or deployment date."}], "date": "Unknown", "source_type": "Official hiring description", "technology_note": "W22EX-FANG", "technology_label": "Read the hiring signal and its limits", "asset": "Enterprise AI portfolio; Midland and Dallas role locations, no named operating asset.", "ai_role": "Intended economics and attribution governance; no new model or operating workflow established.", "partner": "No outside partner named", "core_company": true, "category": "AI investment governance", "baseline": "Intended measurement standards, not evidence of a measured intervention.", "relationship": "Context for Diamondback readiness; separate from AI-045 and AI-251. No new adoption or counted result.", "source_id": "WATCH22EX-FANG", "source_title": "AI Lead – Value and Economics · R100885", "url": "https://diamondbackenergy.wd12.myworkdayjobs.com/DBE/job/Midland-TX/AI-Lead---Value-and-Economics_R100885", "publisher": "Diamondback Energy / official Workday careers", "locator": "Job Responsibilities: valuation framework, baselines/comparison groups, benefits attribution, continue/scale/pivot/stop recommendations. Official public jobPostingInfo payload retrieved.", "excerpt": "recommend whether investments should continue, scale, pivot, or stop"}, {"id": "W22EX-ITALGAS", "company": "Italgas", "record": null, "related_ids": [], "title": "Italgas’s voice agent can help with appointments; the reported gains need more checking", "department": "Commercial, power & infrastructure", "status": "Adjacent gas-distribution case; limited production beta described", "summary": "A vendor-hosted talk describes Italgas testing a voice agent on a subset of customer calls, with people handling complex requests.", "mechanism": "A real-time language model uses governed knowledge and authenticated CRM tools for appointment work. Databricks supports retrieval and tracing; Azure hosts the voice model. Broader channels are planned.", "measure": "Speakers claim 50% fewer dropped calls and 40% more automated resolutions in beta. Test size, period, metric definitions, quality, costs and net return are absent.", "operating_period": "Undated talk. Full call coverage was planned for September, with no verified year or completion. It is not confirmed as a fresh announcement.", "limit": "The transcript’s earlier baseline figures do not reconcile. The summary omits the beta boundary. Do not reconstruct absolute outcomes, infer full rollout or treat these claims as measured upstream results.", "question": "Does the bot complete the right task, escalate safely, and reduce total effort after corrections?", "sources": [{"id": "WATCH22EX-ITALGAS", "title": "Building Production Voice Agents with Databricks and AI Gateway", "url": "https://answers.databricks.com/building-production-voice-agents-databricks-ai-gateway-w8XJClcgjx4", "date": "Unknown", "publisher": "Databricks / named Italgas and Reply speakers", "locator": "Full transcript: 02:04 and09:38 baseline conflict;10:10–16:57 architecture and CRM;19:28–20:15 limited beta, planned September rollout and claimed changes;21:05 human operators.", "excerpt": "we are in a kind of beta testing moment", "access": "Editor read vendor-hosted named-speaker transcript, not just its summary/FAQs. Original video link fetch failed. Transcript baseline numbers conflict; exact talk/publication date and full rollout remain unverified."}], "date": "Unknown", "source_type": "Vendor-hosted customer conference transcript", "technology_note": "W22EX-ITALGAS", "technology_label": "Read the adjacent gas-distribution case", "asset": "Gas-distribution contact centre; beta call population undisclosed.", "ai_role": "Customer-service language agent with CRM tools; no physical gas-network control described.", "partner": "Databricks; Reply; Microsoft Azure", "core_company": false, "category": "Adjacent gas-distribution customer service", "baseline": "Conflicting transcript baseline; no absolute result reconstruction.", "relationship": "Adjacent gas-distribution analog, not an upstream E&P initiative or addition to core adoption/results.", "source_id": "WATCH22EX-ITALGAS", "source_title": "Building Production Voice Agents with Databricks and AI Gateway", "url": "https://answers.databricks.com/building-production-voice-agents-databricks-ai-gateway-w8XJClcgjx4", "publisher": "Databricks / named Italgas and Reply speakers", "locator": "Full transcript: 02:04 and09:38 baseline conflict;10:10–16:57 architecture and CRM;19:28–20:15 limited beta, planned September rollout and claimed changes;21:05 human operators.", "excerpt": "we are in a kind of beta testing moment"}]}, "watch_2026_10_22_failures": {"as_of": "2026-10-22", "run": "2026-10-22-failures", "label": "22 October 2026 / Failures and scaling review", "headline": "One pressure-prediction model performed poorly in testing.", "intro": "One missed SPE study gives us a concrete negative result. The named-company searches did not establish another abandoned AI project.", "count_note": "One technical note and one library paper added. No new core deployment, result or net ROI count. The register remains 297 profiles: 212 company-use, 146 operating or scaled and 61 with a counted reported result.", "method": "We checked named-company claims, supplier programs and technical studies against the existing report. Equipment failures, ordinary ERP cutovers, anonymous claims and missing ROI were not treated as proof that an AI deployment failed. Coverage remains incomplete.", "featured_count": 1, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W22FAIL-SPE", "company": "SPE pressure-prediction study", "record": null, "related_ids": [], "title": "A neural network struggled to predict pressure in this test", "department": "Reservoir engineering & field development", "status": "Model-test weakness; no operating deployment established", "summary": "An SPE study compared a random forest and a neural network on 206 historical pressure samples. The neural network performed poorly on the reported test split.", "mechanism": "The authors used 10 inputs and an 80/20 training/test split. They describe a small, noisy dataset; the abstract does not establish testing on entirely unseen wells.", "measure": "Neural-network mean absolute error: 209.55 psig; random forest: 93.73 psig. Test R² was 0.05 versus 0.81. R² is not a percent-correct accuracy score. No operating loss or deployment cost is reported.", "operating_period": "Published 10 August 2026. The ProBHP data are attributed to sources from 1975 and 1986; collection dates, well count, geography and test chronology are unspecified.", "limit": "Only the publisher-indexed abstract was reviewed; direct access failed. No raw data, full validation protocol, deployed system, stop decision or later outcome verified. This does not establish a failure rate for industry AI projects.", "question": "Would performance hold on entirely new wells, with the full error distribution and an operating decision attached?", "sources": [{"id": "WATCH22FAIL-SPE", "title": "Evaluation of Influential Parameters for Prediction of Flowing Bottom Hole Pressure in Dry Gas Wells Using Machine Learning", "url": "https://onepetro.org/SPENAIC/proceedings-abstract/26NAIC/26NAIC/SPE-234889-MS/802231", "date": "2026-08-10", "publisher": "Society of Petroleum Engineers / OnePetro", "locator": "SPE-234889-MS; publisher-indexed abstract: dataset/split and Random Forest/ANN comparison paragraphs. Published 10 August 2026; conference 3–5 August 2026.", "excerpt": "ANN model performed poorly", "access": "Editor independently read the search-indexed original publisher abstract and metadata. Direct OnePetro/DOI opens failed; local page fetch returned 403. Full paper, raw data and complete methods not reviewed. No claim of field deployment or independent replication."}], "date": "2026-08-10", "source_type": "Publisher-indexed technical abstract", "technology_note": "W22FAIL-SPE", "technology_label": "Read the model-test limits", "asset": "Historical dry-gas pressure dataset; 206 samples, not 206 identified wells.", "ai_role": "Random forest versus artificial neural network for pressure prediction; no LLM or field control.", "partner": "Authors: Fred Akpososo, Victor Aimikhe, David Kalu and Precious Okolaa", "core_company": false, "category": "Technical model-validation limits", "baseline": "Reported test split of a 206-sample dataset; exact test count and independence by well unverified.", "relationship": "Technical benchmark only, separate from Santos AI-139 and all operating pressure-model claims. No core adoption or counted result.", "report_id": "WATCH22FAIL-SPE", "source_id": "WATCH22FAIL-SPE", "source_title": "Evaluation of Influential Parameters for Prediction of Flowing Bottom Hole Pressure in Dry Gas Wells Using Machine Learning", "url": "https://onepetro.org/SPENAIC/proceedings-abstract/26NAIC/26NAIC/SPE-234889-MS/802231", "publisher": "Society of Petroleum Engineers / OnePetro", "locator": "SPE-234889-MS; publisher-indexed abstract: dataset/split and Random Forest/ANN comparison paragraphs. Published 10 August 2026; conference 3–5 August 2026.", "excerpt": "ANN model performed poorly"}]}, "watch_2026_10_22_roi": {"reviewed_on": "2026-10-22", "source_id": "WATCH22ROI-EY", "disposition": "Existing sidecar survey newly indexed; no new deployment or failure case."}, "watch_2026_10_22_pm": {"as_of": "2026-10-22", "run": "2026-10-22-pm", "label": "22 October 2026 / Evening research", "headline": "The useful detail is in the work people actually do.", "intro": "Vista describes real-time completion decisions. Saipem has an installed maintenance pilot. Worley names four internal tools. Whitecap and Technip Energies show why a platform, an MVP and a deployed workflow need different labels.", "count_note": "Three profiles added; one existing program updated and one context-only note added. Vista’s technical measures receive source-reported result credit; its ambiguous dollar claim does not. No new verified net cash ROI.", "method": "Three bounded Luna assignments searched operators, services and private-company evidence. The editor checked retained claims against original issuer material or clearly labeled indexed excerpts, corrected slide locations and separated plans from use. This is a targeted review, not exhaustive coverage.", "featured_count": 5, "archive_label": "Earlier source reviews and company context", "findings": [{"core_company": true, "technology_label": "Read the source and limits", "source_type": "Company investor presentation", "id": "W22PM-VISTA", "company": "Vista Energy", "record": "AI-298", "related_ids": ["AI-298"], "title": "Vista uses AI monitoring to adjust completion designs", "department": "Drilling & completions", "status": "Operating since March 2025, per issuer", "summary": "Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta.", "mechanism": "The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025.", "measure": "Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established.", "operating_period": "Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed.", "limit": "No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted.", "question": "What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review?", "sources": [{"id": "WATCH22PM-VISTA", "title": "Vista Investor Day 2025", "url": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "date": "2025-11-12", "publisher": "Vista Energy", "locator": "Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.", "excerpt": "Implemented since March 2025", "access": "Editor downloaded and read the original issuer PDF, inspected slide 19 visually and cross-checked the SEC exhibit. Company-reported, not independently validated."}, {"id": "WATCH22PM-VISTA-SEC", "title": "Vista Investor Day 2025 — SEC exhibit", "url": "https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm", "date": "2025-11-12", "publisher": "Vista Energy / SEC EDGAR", "locator": "Exhibit heading: November 12, 2025; same slide 19 presentation.", "excerpt": "Vista Investor Day", "access": "Original SEC exhibit read. A second access path to the same issuer presentation, not independent corroboration of outcomes."}], "date": "2025-11-12", "asset": "Vaca Muerta, Argentina; specific block, pad and well count undisclosed.", "ai_role": "Proprietary AI monitoring for real-time completion-design decisions; no LLM or autonomous pumping authority established.", "partner": "No implementation partner named on the reviewed slide.", "category": "Operating completion decisions", "baseline": "2025 technical measures; matched baseline and denominator undisclosed.", "relationship": "One new named operating profile. The PDF and SEC exhibit are the same publication, not two deployments.", "stage_norm": "operating", "segment_norm": "Independent E&P", "outcome_type": "realized", "outcome_counted": true, "problem": "Recognize frac hits while a completion is still underway and adjust the design before more stages are lost.", "human": "Engineering decision support is described. The slide does not define human approval, override or execution rights.", "secondary": ["Reservoir engineering & field development"], "technology_note": "W22PM-VISTA", "source_id": "WATCH22PM-VISTA", "source_title": "Vista Investor Day 2025", "url": "https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf", "publisher": "Vista Energy", "locator": "Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading.", "excerpt": "Implemented since March 2025"}, {"core_company": true, "technology_label": "Read the source and limits", "source_type": "Company announcement", "id": "W22PM-SAIPEM", "company": "Saipem", "record": "AI-299", "related_ids": ["AI-299", "AI-160", "AI-241", "AI-242"], "title": "Saipem installs a predictive-maintenance pilot on one drillship", "department": "Maintenance & reliability", "status": "Installed pilot on Saipem 12000; fleet extension planned", "summary": "Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it.", "mechanism": "With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests.", "measure": "No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment.", "operating_period": "Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective.", "limit": "The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed.", "question": "Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs?", "sources": [{"id": "WATCH22PM-SAIPEM", "title": "Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000", "url": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "date": "2026-02-23", "publisher": "Saipem", "locator": "Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.", "excerpt": "This pilot project", "access": "Editor read the full original issuer announcement. Installation is described; no measured reliability result or fleet rollout independently verified."}], "date": "2026-02-23", "asset": "Saipem 12000 ultra-deepwater drillship. Saipem 7000 is separate future context.", "ai_role": "Equipment-data anomaly detection and failure prediction; no LLM or autonomous maintenance execution established.", "partner": "ADC Energy for Saipem 12000; BIP for the separate Saipem 7000 project.", "category": "Installed predictive-maintenance pilot", "baseline": "No pre-pilot reliability comparator or evaluation period disclosed.", "relationship": "Distinct from Saipem scheduling, video safety and drone initiatives. The separate 7000 plan is not another operating profile.", "stage_norm": "pilot", "segment_norm": "Oilfield services", "outcome_type": "none", "outcome_counted": false, "problem": "Find emerging equipment problems early enough to plan maintenance on a drillship.", "human": "Maintenance planning is described. Approval, execution, override and fallback boundaries are undisclosed.", "secondary": ["Production & field operations"], "technology_note": "W22PM-SAIPEM", "source_id": "WATCH22PM-SAIPEM", "source_title": "Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000", "url": "https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000", "publisher": "Saipem", "locator": "Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs.", "excerpt": "This pilot project"}, {"core_company": true, "technology_label": "Read the source and limits", "source_type": "Company investor presentation", "id": "W22PM-WORLEY", "company": "Worley", "record": "AI-300", "related_ids": ["AI-300"], "title": "Worley puts four internal tools into engineering and bid work", "department": "IT, data & enterprise transformation", "status": "Four recently deployed internal solutions, per issuer", "summary": "Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting.", "mechanism": "AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest.", "measure": "Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed.", "operating_period": "FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown.", "limit": "These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established.", "question": "Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved?", "sources": [{"id": "WATCH22PM-WORLEY", "title": "FY2025 full-year results presentation", "url": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "date": "2025-08-27", "publisher": "Worley", "locator": "Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.", "excerpt": "Recently deployed solutions through our Advanced Development Lab", "access": "Editor downloaded and read the original issuer presentation and visually inspected slide 31. Named internal tools, not measured client or field outcomes."}], "date": "2025-08-27", "asset": "Internal engineering and commercial work; no client asset, basin or user population disclosed.", "ai_role": "Internal AI-assisted document review, search, comparison and drafting. Exact model architecture and execution permissions undisclosed.", "partner": "No tool-specific implementation partner established by the reviewed slide.", "category": "Internal engineering and commercial tools", "baseline": "No tool-level before/after or quality-controlled comparator disclosed.", "relationship": "One internal program profile covering four named tools, not four field deployments.", "stage_norm": "operating", "segment_norm": "Oilfield services", "outcome_type": "scale_only", "outcome_counted": false, "problem": "Reduce repeated work finding engineering knowledge, checking documents and preparing technical or commercial responses.", "human": "Tools support review and drafting. Named approval roles and execution rights are undisclosed.", "secondary": ["Supply chain & logistics"], "technology_note": "W22PM-WORLEY", "source_id": "WATCH22PM-WORLEY", "source_title": "FY2025 full-year results presentation", "url": "https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf", "publisher": "Worley", "locator": "Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover.", "excerpt": "Recently deployed solutions through our Advanced Development Lab"}, {"core_company": false, "technology_label": "Read the source and limits", "source_type": "Vendor customer story", "id": "W22PM-WHITECAP", "company": "Whitecap Resources", "record": null, "related_ids": [], "title": "Whitecap describes the data foundation it is building around AI", "department": "IT, data & enterprise transformation", "status": "Operating data platform; predictive and agentic uses remain prospective", "summary": "A Shakudo customer story describes Whitecap’s governed data platform and business-led AI prototypes, with IT bringing selected work into production.", "mechanism": "The account describes relational data and operating telemetry, on-premise language models and selective frontier-model access. Human review remains part of the approach.", "measure": "The page’s less-than-one-hour analytics headline conflicts with narrative describing an endpoint still being built toward. It is not counted as achieved savings or ROI. Company production and team size are not an AI user population.", "operating_period": "Vendor index dated 23 April 2026; story itself undated. Platform use described, but project start and evaluation dates are unknown.", "limit": "Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved.", "question": "Which prototypes have become routine work, and what does the production team measure before approving wider use?", "sources": [{"id": "WATCH22PM-WHITECAP", "title": "Whitecap Resources customer story", "url": "https://www.shakudo.io/customers/whitecap-resources", "date": "2026-04-23", "publisher": "Shakudo", "locator": "Key Results; govern-first platform, AI operating model, human review and endpoint being built toward. Page is undated; date comes from vendor customer index.", "excerpt": "The endpoint being built toward", "access": "Editor read the original vendor customer story. Named customer executive is quoted; operator-side corroboration and outcome methods not verified."}, {"id": "WATCH22PM-WHITECAP-DATE", "title": "Shakudo customer stories — Whitecap entry", "url": "https://www.shakudo.io/customers", "date": "2026-04-23", "publisher": "Shakudo", "locator": "Whitecap customer-story listing dated April 23, 2026.", "excerpt": "Whitecap Resources", "access": "Original vendor customer index read; verifies listed publication date, not a deployment or benefit."}], "date": "2026-04-23", "asset": "Whitecap/Veren operating data; no asset-specific AI evaluation population disclosed.", "ai_role": "Governed data and internal AI experimentation; no deployed predictive field model or autonomous control established.", "partner": "Shakudo", "category": "Data foundations and production governance", "baseline": "The analytics-cycle headline is not a verified before/after measurement.", "relationship": "Context-only technology note. Does not add a core AI adoption, counted outcome or net ROI.", "technology_note": "W22PM-WHITECAP", "source_id": "WATCH22PM-WHITECAP", "source_title": "Whitecap Resources customer story", "url": "https://www.shakudo.io/customers/whitecap-resources", "publisher": "Shakudo", "locator": "Key Results; govern-first platform, AI operating model, human review and endpoint being built toward. Page is undated; date comes from vendor customer index.", "excerpt": "The endpoint being built toward"}, {"core_company": true, "technology_label": "Read the source and limits", "source_type": "Indexed original annual-report excerpt", "id": "W22PM-TEN", "company": "Technip Energies", "record": "AI-199", "related_ids": ["AI-199"], "title": "Technip Energies reports progress, without proving 55 AI deployments", "department": "IT, data & enterprise transformation", "status": "Mixed digital and AI program; end-2025 output counts", "summary": "Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025.", "mechanism": "The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established.", "measure": "The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI.", "operating_period": "2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date.", "limit": "Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved.", "question": "Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs?", "sources": [{"id": "WATCH22PM-TEN", "title": "Technip Energies 2025 Annual Report", "url": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "date": "2026 (exact publication day unverified)", "publisher": "Technip Energies", "locator": "Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.", "excerpt": "more than 55 digital and AI initiatives", "access": "Editor read the search-indexed original annual-report excerpt. Direct PDF returned 403; full report unreviewed. URL timestamp is not used as an exact publication date."}], "date": "2026 (exact publication day unverified)", "asset": "Company transformation program; project and asset populations undisclosed.", "ai_role": "Mixed digital and AI initiatives and staff upskilling; no inference of 55 AI production deployments.", "partner": "No partner verified in the reviewed excerpt.", "category": "Program output and scaling limits", "baseline": "End-2025 program output counts, not a benefit baseline.", "relationship": "Enriches existing AI-199. No second program, additional deployment count or numerical-result credit.", "stage_norm": "operating", "segment_norm": "Oilfield services", "outcome_type": "scale_only", "outcome_counted": false, "problem": "Move useful digital and AI ideas into working tools and train staff to use them.", "human": "Program and upskilling outputs disclosed; project-level approval and execution boundaries unknown.", "secondary": ["Workforce & learning"], "technology_note": "W22PM-TEN", "source_id": "WATCH22PM-TEN", "source_title": "Technip Energies 2025 Annual Report", "url": "https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf", "publisher": "Technip Energies", "locator": "Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts.", "excerpt": "more than 55 digital and AI initiatives"}]}, "watch_2026_10_23_am": {"as_of": "2026-10-23", "run": "2026-10-23-am", "label": "23 October 2026 / Morning research", "headline": "From the well plan to the rig.", "intro": "New SLB disclosures connect planning with rig execution, describe chemical dosing controls and name on-premise software work. The operator, the scope and the measurement still matter.", "count_note": "Two profiles added: Azule’s operating example and BSP’s software award. Eni and ADNOC are updates to existing records. The unnamed chemistry account stays outside named adoption and result totals. No verified full-cost net cash ROI added.", "method": "Three bounded Luna searches found no additional qualifying case; the editor then checked the newly published SLB Q3 release and an original supplier case study. All claims remain attributed to their sources. This is not an exhaustive earnings-call census.", "featured_count": 5, "archive_label": "Earlier source reviews and company context", "findings": [{"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-AZULE", "company": "Azule Energy", "record": "AI-301", "related_ids": ["AI-301"], "title": "Azule connects the well plan to drilling automation", "department": "Drilling & completions", "status": "Operating single-well example; wider rollout unverified", "summary": "SLB describes Azule using one digital workflow from well design through rig execution offshore Angola.", "mechanism": "DrillPlan passes the plan into DrillOps automation. The supplier case identifies AI-driven operations; it does not isolate AI from engineering automation.", "measure": "Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return.", "operating_period": "Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown.", "limit": "No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate.", "question": "Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted?", "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / Azule Energy paragraph; same case and key measures.", "excerpt": "cut planning time from 192 to 31 hours", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf"}, {"id": "WATCH23AM-AZULE", "title": "Azule Energy pioneers digital well planning and rig automation in Africa", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa", "date": "Unknown", "publisher": "SLB", "locator": "Location; Results: Achieving operational excellence; Products Used: DrillOps AI-driven workflows.", "excerpt": "digitally ingested the well plan", "access": "Editor read the original supplier case study. No exact publication date displayed; the 2026 URL path and search crawl dates are not publication dates. Same supplier as the Q3 release; not independent corroboration."}], "asset": "One offshore Angola well; no named field or basin disclosed.", "ai_role": "AI-enabled DrillOps execution alongside digital well planning; underlying algorithms and model type undisclosed.", "category": "Plan-to-execution operating workflow", "baseline": "Planning before/after and an unspecified field ROP benchmark; no matched controlled comparison.", "relationship": "One named operator profile. Case study and quarterly release are both SLB publications; no independent result validation.", "stage_norm": "operating", "segment_norm": "Independent E&P", "outcome_type": "realized", "outcome_counted": true, "problem": "Carry an engineering plan into rig execution with fewer manual transfers.", "human": "Autonomous section execution is reported. Approval, override, stop and fallback boundaries are not disclosed.", "core_company": true, "secondary": [], "technology_note": "W23AM-AZULE", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 16: Digital / Azule Energy paragraph; same case and key measures.", "excerpt": "cut planning time from 192 to 31 hours"}, {"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-CHEMISTRY", "company": "SLB / unnamed Permian operator", "record": null, "related_ids": [], "title": "A chemistry-control result is useful, even though the operator is unnamed", "department": "Production & field operations", "status": "Supplier-reported deployment; operator identity withheld", "summary": "SLB reports AI injection control and anomaly detection on nearly 1,400 producing Permian wells, with remote recalibration and weekly reviews.", "mechanism": "Analyst takeaway: measure dosage accuracy alongside cost, and ask who reviews exceptions.", "measure": "Reported on-target injection: 63% to 92%; overtreatment: 17% to 3%; $4.4 million in net costs avoided in the first nine months. The supplier does not define the net-cost calculation or disclose a full-cost ROI.", "operating_period": "Published 23 October 2026. First nine months of the program; exact start/end dates not stated.", "limit": "Unnamed operator: no company attribution, core adoption or result-count addition. No controls, cost reconciliation, model error rate or approval/override policy. Do not infer a specific Permian producer or treat the figure as audited net cash ROI.", "question": "What costs were included, and did lower overtreatment persist without compromising corrosion or scale protection?", "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 14: Technology Highlights / United States production chemistry paragraph.", "excerpt": "AI-driven injection control", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf"}], "asset": "Nearly 1,400 producing Permian wells; operator and individual assets unnamed.", "ai_role": "AI injection control plus ML anomaly detection; human review cadence disclosed, execution permissions and fallback unspecified.", "category": "Anonymous chemistry-control result", "baseline": "Reported pre/post dose metrics; no control group or matched operating conditions.", "relationship": "Context only; anonymous supplier account excluded from named operator adoption and outcome totals.", "core_company": false, "secondary": [], "technology_note": "W23AM-CHEMISTRY", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 14: Technology Highlights / United States production chemistry paragraph.", "excerpt": "AI-driven injection control"}, {"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-ENI", "company": "Eni", "record": "AI-027", "related_ids": ["AI-027"], "title": "Eni adds a concrete on-premise subsurface workflow", "department": "IT, data & enterprise transformation", "status": "Supplier reports expanded operating workflows", "summary": "SLB says Eni expanded petrophysics and geomechanics AI workflows using an on-premise Lumi platform.", "mechanism": "Analyst takeaway: the location of the data platform matters when deciding what information a new assistant or model can access.", "measure": "No task-level benefit, asset population or full cost quantified in this update. Eni’s existing portfolio metrics and targets remain separately sourced; none is allocated to Lumi.", "operating_period": "Published 23 October 2026; exact deployment and expansion dates undisclosed.", "limit": "Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications.", "question": "Which petrophysics or geomechanics decisions changed, and what data and review boundaries apply?", "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / SLB and Eni paragraph.", "excerpt": "AI-driven workflows", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf"}], "asset": "Eni subsurface work; no field, basin or user population identified.", "ai_role": "AI-enabled petrophysics and geomechanics on on-premise Lumi; specific model architecture undisclosed.", "category": "Existing program: named operating detail", "baseline": "No matched quality, time or operating comparison disclosed.", "relationship": "Adds source-bounded detail to AI-027; no new core profile or counted result.", "secondary": ["Reservoir engineering & field development"], "core_company": true, "technology_note": "W23AM-ENI", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 16: Digital / SLB and Eni paragraph.", "excerpt": "AI-driven workflows"}, {"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-BSP", "company": "Brunei Shell Petroleum", "record": "AI-302", "related_ids": ["AI-302"], "title": "BSP awards an on-premise software contract", "department": "IT, data & enterprise transformation", "status": "Contract awarded; go-live and usage unverified", "summary": "BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work.", "mechanism": "Analyst takeaway: a broad contract establishes buying intent and implementation scope; it does not show which AI tasks employees complete.", "measure": "No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment.", "operating_period": "Published 23 October 2026. Award date, implementation timetable and active-use dates unknown.", "limit": "Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout.", "question": "Which workflows will go live first, and what acceptance test separates installed software from useful work?", "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / Brunei BSP software contract paragraph.", "excerpt": "on-premises software contract", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf"}], "asset": "BSP, Brunei; subsurface and production scope, no specific asset or user count.", "ai_role": "Contract includes AI capabilities; no particular live model or LLM workflow established.", "category": "Named software award", "baseline": "No live-use or outcome baseline established.", "relationship": "One announced customer profile. Distinct from Shell’s global AI-infrastructure collaboration and from BSP’s separate Retina equipment use.", "stage_norm": "announced", "segment_norm": "Integrated & supermajor", "outcome_type": "none", "outcome_counted": false, "problem": "Connect subsurface and production software with a consistent implementation and support arrangement.", "human": "No project-level AI approval or execution boundaries disclosed.", "core_company": true, "secondary": [], "technology_note": "W23AM-BSP", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 16: Digital / Brunei BSP software contract paragraph.", "excerpt": "on-premises software contract"}, {"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-ADNOC", "company": "ADNOC", "record": "AI-226", "related_ids": ["AI-226"], "title": "SLB repeats ADNOC’s rig-center scope, without a new rollout count", "department": "Drilling & completions", "status": "Supplier repeats existing 120-plus-rig scope", "summary": "SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs.", "mechanism": "Analyst takeaway: this is another source for an existing footprint, not evidence that a new set of rigs was added.", "measure": "No additional rollout count or newly measured benefit. Earlier ADNOC claims retain their own source, period and limitations.", "operating_period": "Published 23 October 2026; earlier operator disclosure was 4 August 2026. No new commissioning date.", "limit": "Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet.", "question": "What is monitored centrally, what action can the platform take and who approves a change on a rig?", "sources": [{"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 16: Digital / ADNOC Real-Time Operations Center paragraph.", "excerpt": "Real-Time Operations Center", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf"}], "asset": "Existing ADNOC rig-center fleet; no new asset-specific population.", "ai_role": "AI-enabled drilling oversight; per-rig autonomous execution is not established by this paragraph.", "category": "Existing deployment: source corroboration", "baseline": "No new comparable outcome or rollout baseline.", "relationship": "Existing AI-226 only. No new deployment, result or ROI credit.", "core_company": true, "secondary": [], "technology_note": "W23AM-ADNOC", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 16: Digital / ADNOC Real-Time Operations Center paragraph.", "excerpt": "Real-Time Operations Center"}]}, "watch_2026_10_23_pm": {"as_of": "2026-10-23", "run": "2026-10-23-pm", "label": "23 October 2026 / Evening research", "headline": "AI can do the first pass. Who signs off?", "intro": "A Petrobras supplier pilot gives a clear answer: technical staff still check the documents and make the decision.", "count_note": "One pilot profile added. Operating deployments, counted results and verified net cash ROI are unchanged. This is an undated source with a September 2026 schedule, not a new announcement today.", "method": "Three bounded Luna tracks checked SLB call material, anonymous field cases and operator implementation limits. The editor verified the retained Petrobras notice against the original. No new named failure or disappointing full-cost return was substantiated; this is not an exhaustive search.", "featured_count": 1, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W23PM-PETROBRAS", "company": "Petrobras", "record": "AI-303", "related_ids": ["AI-303"], "title": "AI checks supplier documents. Petrobras staff make the decision.", "department": "Supply chain & logistics", "status": "Wave 2 pilot; rollout unconfirmed", "summary": "Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them.", "mechanism": "The evaluation team records its decision and feeds it back to the AI. Petronect handles registration; the notice does not identify the AI model or developer.", "measure": "No accuracy, workload, completed-case count, cost, saving or ROI reported.", "operating_period": "Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed.", "limit": "Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established.", "question": "Does the first AI pass reduce total review work while preserving the quality of supplier decisions?", "sources": [{"id": "WATCH23PM-PETROBRAS", "title": "Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice", "url": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "date": "Unknown", "publisher": "Petrobras", "locator": "Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.", "excerpt": "A supervisão e a decisão final permanecem sempre com nossos avaliadores técnicos.", "access": "Editor read the original Petrobras Supplier Channel notice in Portuguese. Exact publication date unknown; the notice describes Wave 2 and a form-opening schedule from September 2026. Current completed-case volume and pilot results are not supplied."}], "date": "Unknown", "source_type": "Company supplier-channel pilot notice", "partner": "Petronect registration portal; AI technology partner unknown", "record_type": "company_adoption", "asset": "Petrobras corporate supplier registration; no producing asset or basin applies. Participating supplier volume is undisclosed.", "ai_role": "Preliminary document analysis; model type and developer undisclosed.", "category": "Supplier qualification", "baseline": "No before/after comparison or measurement period disclosed.", "relationship": "One named pilot. Possible suspension is a future contingency, not a recorded setback.", "stage_norm": "pilot", "segment_norm": "National oil company", "outcome_type": "none", "outcome_counted": false, "problem": "Check supplier evidence during registration.", "human": "Technical evaluators perform the final check and retain the decision.", "core_company": true, "secondary": [], "technology_note": "W23PM-PETROBRAS", "technology_label": "Read the pilot and its limits", "source_id": "WATCH23PM-PETROBRAS", "source_title": "Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice", "url": "https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores", "publisher": "Petrobras", "locator": "Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date.", "excerpt": "A supervisão e a decisão final permanecem sempre com nossos avaliadores técnicos."}]}, "watch_2026_10_24_am": {"as_of": "2026-10-24", "run": "2026-10-24-am", "label": "24 October 2026 / Morning research", "headline": "An early warning is only the start.", "intro": "Halliburton describes turning equipment alerts into planned maintenance. ONGC shows the older enterprise systems behind today’s AI discussion. The two examples answer different questions.", "count_note": "One supplier-product profile added, plus historical ONGC context. Named operator adoption, operating adoption, counted results and net cash ROI are unchanged.", "method": "We checked regional operator disclosures, private portfolios and platform customer stories, then read the original sources for these additions. This targeted review found no substantiated new named failure or disappointing full-cost return. It does not cover every company or document.", "featured_count": 2, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "W24AM-HAL-RELIABILITY", "company": "Halliburton", "record": "AI-304", "related_ids": ["AI-304"], "sources": [{"id": "WATCH24AM-HAL-RELIABILITY", "title": "How predictive intelligence reduces downtime and improves equipment reliability", "url": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "date": "2026-10-22", "publisher": "Halliburton", "locator": "A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.", "excerpt": "helped teams avoid an estimated $125,000 to $150,000 in downtime costs", "access": "Editor read the complete original Halliburton article and displayed publication date. Supplier-selected examples; operator names, operating dates and measurement methods are not provided."}, {"id": "WATCH24AM-HAL-PRODUCT", "title": "LOGIX Unit Vitality — cement equipment health monitoring", "url": "https://www.halliburton.com/en/products/logix-unit-vitality", "date": "Unknown", "publisher": "Halliburton", "locator": "Automated cement equipment health assessment; AI-driven models; Real-time monitoring closes the loop.", "excerpt": "more than 400 real-time parameters", "access": "Editor read the original product page. Exact publication date unknown. Same supplier as the article, not independent validation; product capabilities are not proof each feature was used in the four reported deployments."}], "date": "2026-10-22", "source_id": "WATCH24AM-HAL-RELIABILITY", "source_title": "How predictive intelligence reduces downtime and improves equipment reliability", "url": "https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability", "publisher": "Halliburton", "locator": "A simple question led to a new approach; How predictive reliability works; What early risk identification delivers.", "excerpt": "helped teams avoid an estimated $125,000 to $150,000 in downtime costs", "source_type": "Supplier product account", "partner": "Halliburton", "core_company": true, "secondary": [], "technology_note": "W24AM-HAL-RELIABILITY", "technology_label": "Read the work and its limits", "title": "A maintenance alert only helps if a crew can act on it.", "department": "Maintenance & reliability", "status": "Supplier reports four deployments; customers unidentified", "summary": "Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs.", "mechanism": "Sensors feed more than 400 equipment parameters into cloud models. Staff check warnings, arrange parts and schedule corrective work. The product page separately describes transferring approved job designs from VIDA and iCem into controllers; this is not proof of autonomous maintenance.", "measure": "The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined.", "operating_period": "Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed.", "limit": "No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016.", "question": "Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs?", "asset": "Cementing equipment; four deployments reported, with no identified basin, asset or customer.", "ai_role": "Statistical models and machine learning for equipment condition; no LLM identified.", "human": "Crews investigate alerts, obtain parts and schedule repairs; authorization, override and fallback rules are not specified.", "category": "Equipment reliability and planned maintenance", "baseline": "No comparator, productivity definition or cost model disclosed.", "relationship": "Supplier product with four unidentified deployment examples; not a new named operator deployment.", "stage_norm": "operating", "outcome_type": "scale_only", "outcome_counted": false, "problem": "Find developing equipment problems early enough for a useful maintenance intervention.", "record_type": "vendor_product"}, {"id": "W24AM-ONGC-ICE", "company": "ONGC", "record": null, "related_ids": [], "sources": [{"id": "WATCH24AM-ONGC-ICE", "title": "ONGC Technology — Project ICE", "url": "https://www.ongcindia.com/web/eng/about-ongc/technology", "date": "Unknown", "publisher": "ONGC", "locator": "Project ICE, five paragraphs. Site-wide last-updated stamp does not establish when this section or its roadmap was published.", "excerpt": "24,000 users, and 23 SAP modules", "access": "Editor read the original ONGC Project ICE section. Legacy SAP R/3 rollout is dated 2004; section publication date and S/4HANA transition date are unknown. Website-wide last-updated label is not a section publication date."}], "date": "Unknown", "source_id": "WATCH24AM-ONGC-ICE", "source_title": "ONGC Technology — Project ICE", "url": "https://www.ongcindia.com/web/eng/about-ongc/technology", "publisher": "ONGC", "locator": "Project ICE, five paragraphs. Site-wide last-updated stamp does not establish when this section or its roadmap was published.", "excerpt": "24,000 users, and 23 SAP modules", "source_type": "Company technology overview", "partner": "SAP platform; implementation partner undisclosed", "core_company": true, "secondary": [], "technology_note": "W24AM-ONGC-ICE", "technology_label": "Read the work and its limits", "title": "ONGC’s shared enterprise system predates the current AI wave.", "department": "IT, data & enterprise transformation", "status": "Historical operating foundation; current roadmap undated", "summary": "ONGC says Project ICE connects finance, materials, HR, production revenue and other functions on SAP S/4HANA.", "mechanism": "The operator reports more than 500 locations, 24,000 users and 23 modules. It describes a three-site disaster recovery setup and a SAP RISE journey over the next two to three years; that roadmap has no section-specific date.", "measure": "These are reported platform-scale figures, not active AI users, measured uptime, productivity, cost savings or ROI.", "operating_period": "Undated page checked 24 October 2026. The program began with SAP R/3 in 2004. S/4HANA transition date and the roadmap’s starting date are unknown.", "limit": "Historical non-AI context only. AI-enabled innovation is a roadmap aim; no specific live AI workflow or decision authority established. A shared platform does not by itself prove clean data, adoption quality or AI readiness.", "question": "Can teams retrieve the same trusted records across functions, and which transactions still require manual reconciliation?", "asset": "Corporate enterprise systems across more than 500 locations; no basin or asset mapping.", "ai_role": "Enterprise software and data foundation; deployed AI not established by this section.", "category": "Historical enterprise software foundation", "baseline": "No independently measured platform quality, availability or economic comparison.", "relationship": "Adds company context, not an AI adoption, result or new project profile.", "record_type": "context"}]}, "slb_data_ai_thesis": {"as_of": "2026-10-24", "run": "slb-data-ai-2026-10-24", "label": "24 October 2026 / SLB deep dive", "headline": "What is SLB’s data worth in the AI era?", "intro": "The investment case becomes more useful when you separate data access, model capability and the money customers actually pay.", "count_note": "Existing SLB platform dossier expanded. No new initiative, adoption or customer-result count. Digital financial results are broader than AI.", "method": "Editor checked original SLB investor scripts, full-year and quarterly financial tables, and public online-service terms. Investor opinion is identified separately. An exact Eisman passage remains unconfirmed. This is a targeted assessment, not a valuation or a full contract audit.", "featured_count": 1, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "SLB-DATA-AI-THESIS", "company": "SLB", "record": "AI-009", "related_ids": ["AI-009", "AI-008", "AI-047", "AI-105", "AI-301"], "sources": [{"id": "SLBTH-Q3", "title": "SLB third-quarter 2026 results — Digital economics and category definitions", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "date": "2026-10-23", "publisher": "SLB", "locator": "Digital results; Supplementary Information questions 10–13; Digital customer highlights.", "excerpt": "$693 million decreased 1% sequentially", "type": "Supplier earnings release", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-FY25", "title": "SLB full-year 2025 results — Digital financial baseline", "url": "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "date": "2026-01-23", "publisher": "SLB", "locator": "Supplementary Information question 12, Digital FY2025 table.", "excerpt": "Digital revenue increased 9% on a full-year basis", "type": "Supplier earnings release", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-MENON", "title": "Shashi Menon: The Race to Scale Digital Operations and AI", "url": "https://investorcenter.slb.com/static-files/cf773c3d-348f-479f-aeb0-81f5418844ca", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF pp. 3–5: data access and control; p. 11: seismic training; pp. 14–15: monetization.", "excerpt": "we trained it with public and SLB multiclient seismic data sets", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-BIGUET", "title": "Stephane Biguet: Proven Growth. Accretive Returns.", "url": "https://investorcenter.slb.com/static-files/0835f6dc-d2fb-4bc0-b85e-afecadcf36fe", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF p. 4: customer base; pp. 7–10: commercial models; p. 11: 2030 ambitions.", "excerpt": "more than 1,500 customers, including 90 of the world’s top 100 oil and gas producers", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}, {"id": "SLBTH-TERMS", "title": "SLB Online Services Terms and Conditions", "url": "https://www.software.slb.com/slb-online-services-terms-and-conditions", "date": "Unknown", "publisher": "SLB", "locator": "Definitions: Customer Data and Interaction Data; sections 8(a)–(b), 9(a), 10(a).", "excerpt": "SLB claims no ownership rights in Customer Data.", "type": "Published online-service terms", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement."}, {"id": "SLBTH-GENAI", "title": "SLB Online Services Service Specific Terms", "url": "https://www.software.slb.com/slb-online-services-service-specific-terms", "date": "Unknown", "publisher": "SLB", "locator": "Generative AI Services: definitions; Connected Gen AI A(i), A(iv); Embedded Gen AI B(iii).", "excerpt": "SLB shall not use Customer Prompts or Output to train models or further develop the Online Services without Customer’s consent.", "type": "Published online-service terms", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation. Effective/publication date unknown. Public standard terms do not establish the terms of every negotiated customer agreement."}, {"id": "SLBTH-MULLIN", "title": "Robert Mullin interview: 3.5% of the S&P, 20% of the Cash", "url": "https://georgenoble.substack.com/p/35-of-the-s-and-p-20-of-the-cash", "date": "2026-10-20", "publisher": "The Noble Update", "locator": "Host introduction and linked interview; exact SLB interview timestamp not established.", "excerpt": "The oil service giant hiding a real AI business that could be worth more than the whole company", "type": "Investor opinion / interview", "access": "Original host introduction reviewed; confirms the investment framing and interview identity. Full interview passage and exact timestamp not independently checked. An Eisman episode containing the same argument remains unconfirmed."}, {"id": "SLBTH-RANNEN", "title": "Trygve Randen: Our Digital Advantage", "url": "https://investorcenter.slb.com/static-files/7aa731c6-2f38-47e4-9aa2-473a8d04ec6e", "date": "2026-06-17", "publisher": "SLB", "locator": "PDF pp. 2–3: governed platforms; pp. 5–6: hosting and trust; p. 7: operational footprint.", "excerpt": "more than ninety percent of global production is simulated or modelled using at least one of our digital solutions", "type": "Investor presentation and prepared script", "access": "Editor read the accessible original publisher text on 24 October 2026. Statements remain publisher-reported; no independent performance validation."}], "date": "2026-10-23", "source_id": "SLBTH-MENON", "source_title": "Shashi Menon: The Race to Scale Digital Operations and AI", "url": "https://investorcenter.slb.com/static-files/cf773c3d-348f-479f-aeb0-81f5418844ca", "publisher": "SLB", "locator": "PDF pp. 3–5: data access and control; p. 11: seismic training; pp. 14–15: monetization.", "excerpt": "we trained it with public and SLB multiclient seismic data sets", "source_type": "Original investor materials, financial disclosures and public terms", "title": "What holds up in SLB’s data and AI investment thesis?", "status": "Existing commercial platform; investor valuation argument remains unproven", "department": "IT, data & enterprise transformation", "summary": "SLB has a credible route to earning more from AI: it can put specialized models into software and operating workflows that customers already use. The public evidence supports a platform, engineering and governed-access advantage. It does not support treating every customer’s downhole data as an unrestricted SLB-owned training pool.", "mechanism": "Menon describes a seismic model trained on public and SLB multiclient datasets, with customers able to fine-tune for their own assets. Lumi connects governed data; domain models handle technical formats; Tela provides interaction and workflow orchestration. SLB describes subscription upgrades, metered usage, tailored workflows and marketplace revenue sharing. These are commercial mechanisms, not disclosed AI-only earnings.", "measure": "FY2025 Digital revenue was $2.660bn and adjusted EBITDA $928m (34.9% margin). Q3 2026 Digital revenue was $693m, adjusted EBITDA $228m and ARR $1.04bn. These are division-wide figures, not AI revenue or customer ROI. The CFO’s 2030 $1.8–2.0bn adjusted EBITDA and approximately $2bn ARR are ambitions, not results.", "operating_period": "Original investor-day scripts: 17 June 2026. FY2025 baseline released 23 January 2026; Q3 results released 23 October 2026. Terms dates unknown. Investor interview framing published 20 October 2026.", "limit": "SLB disclaims ownership of Customer Data in its published online-service terms, while allowing specified service/product-improvement uses. Embedded Gen AI prompts and outputs require customer consent for training. Those clauses do not settle rights to every historical field dataset, derivative or negotiated contract. Digital includes software, operations, data licensing and professional services, with no AI-only profit bridge. Digital Operations overlaps Core segment reporting and is eliminated in consolidation. No demonstrated undervaluation or complete customer net cash ROI follows from these disclosures.", "question": "Ask for AI-specific paid usage and retention, incremental profit after compute and support, training rights by data category, held-out-basin model performance, customer-specific control boundaries, and a valuation using post-elimination earnings. For the full evidence map and interactive revenue chart, read the SLB briefing linked from the homepage.", "problem": "Evaluate whether SLB can convert specialized data, technical software and customer reach into durable incremental AI earnings.", "ai_role": "Technical domain models plus general LLMs and agents; not a single text model trained without restriction on all downhole data.", "human": "Presentation describes engineering recommendations under human control and more autonomous workflows under physical constraints. It does not establish one common approval boundary for every customer deployment.", "asset": "Global Digital business; named operating cases retain their own asset and evidence boundaries.", "partner": "SLB, customer operators and technology partners; no new partner adoption count.", "baseline": "Segment accounts are not AI-only economics. No pooled value or return estimate is admitted.", "category": "Data rights and commercial AI strategy", "relationship": "Expanded existing Lumi dossier and analytical context; no new initiative, operator adoption or reported-result count.", "core_company": true, "technology_note": "SLB-DATA-AI-THESIS", "technology_label": "Read the source review"}]}, "watch_2026_10_24_pm": {"as_of": "2026-10-24", "run": "2026-10-24-pm", "label": "24 October 2026 / Evening research", "headline": "Weatherford says PDO moved ESP prediction beyond a pilot.", "intro": "Weatherford’s April 2026 results describe named operating use in Oman. A separate October release covers non-AI energy recovery.", "count_note": "One supplier-reported operator deployment added, with no measured-result credit. A separate anonymous energy-recovery example is context only.", "method": "Three bounded Luna tracks used six search clauses each across services, private operators/platforms and adverse outcomes. The editor checked retained claims against original Weatherford releases. No new named failed AI rollout or disappointing full-cost return was substantiated. Not an exhaustive company or document census.", "featured_count": 2, "archive_label": "Earlier source reviews and company context", "findings": [{"source_type": "Supplier earnings release", "department": "Production & field operations", "core_company": true, "id": "W24PM-PDO", "company": "Petroleum Development Oman", "record": "AI-305", "related_ids": ["AI-305", "AI-113"], "sources": [{"id": "WATCH24PM-WFRD-Q1", "title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "date": "2026-04-21", "publisher": "Weatherford", "locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "excerpt": "moving the technology from pilot to operational use", "access": "Editor read original issuer HTML and its displayed 21 April 2026 publication date on 24 October. Supplier statement; no independent operator confirmation. Exact operating start date unknown."}], "date": "2026-04-21", "title": "PDO has moved ESP prediction from pilot to operating use", "status": "Operating deployment, supplier-reported; scale unknown", "problem": "Help production teams anticipate electric submersible pump failures and estimate remaining run life.", "summary": "In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework.", "mechanism": "The on-premise installation provides failure forecasts, run-life estimates and real-time pump-performance insight while meeting the customer’s cybersecurity requirements. The release does not say whether forecasts trigger automatic changes or how people review alerts and approve maintenance.", "measure": "No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed.", "operating_period": "Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment.", "limit": "A named operating deployment is supported by Weatherford’s own statement. There is no independent PDO confirmation in this review, asset/basin or pump-population denominator, stated maintenance approval boundary or measured economic result. Regional expansion is a supplier aspiration, not established regional scale. Do not attribute Weatherford’s consolidated results to this use case.", "question": "Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs.", "ai_role": "PetroVisor machine-learning framework supports pump failure/run-life prediction. No LLM or autonomous pump control established.", "human": "Human review, override, alert triage and maintenance approval rules are undisclosed; do not infer autonomous control.", "asset": "Oman; exact field, basin, pump population and operating period undisclosed.", "partner": "Weatherford; ForeSite Well Management System; PetroVisor machine-learning framework.", "baseline": "No counterfactual, baseline or result denominator disclosed.", "category": "Predictive maintenance", "relationship": "Distinct named customer deployment; related Weatherford AI-113 is a supplier product portfolio, not another PDO deployment.", "source_id": "WATCH24PM-WFRD-Q1", "source_title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "publisher": "Weatherford", "locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "excerpt": "moving the technology from pilot to operational use"}, {"source_type": "Supplier earnings release", "department": "Production & field operations", "core_company": true, "id": "W24PM-WFRD-ENERGY", "company": "Weatherford", "record": null, "related_ids": [], "sources": [{"id": "WATCH24PM-WFRD-Q3", "title": "Weatherford Announces Third Quarter 2026 Results — rod-lift energy recovery", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18871", "date": "2026-10-20", "publisher": "Weatherford", "locator": "Technology Highlights → Production and Intervention; first U.S. ForeSite Power Regenerative System retrofit paragraph.", "excerpt": "delivering over 50% in energy savings", "access": "Editor read original issuer HTML and displayed 20 October 2026 publication date on 24 October. One unidentified U.S. well; supplier result, not independent validation."}], "date": "2026-10-20", "title": "A rod-lift retrofit recovers energy on one U.S. well", "status": "First commercial installation, supplier-reported", "problem": "Recover energy during the pumping cycle without replacing an existing rod-lift drive.", "summary": "Weatherford reports the first commercial installation of its ForeSite Power Regenerative System retrofit on an existing U.S. rod-lift well. It captures and reuses energy generated during each pumping cycle.", "mechanism": "The supplier says the retrofit avoids replacement of the installed drive. This is energy recovery equipment; the release does not identify AI or a learning model in this installation.", "measure": "Weatherford reports over 50% energy savings on this installation. Metering method, comparison baseline, test duration and operating conditions are unspecified. No capex, total operating cost or payback is reported.", "operating_period": "Disclosed 20 October 2026 in Q3 results. Exact installation/measurement dates unknown. One unnamed well, not a fleet result.", "limit": "Anonymous single-installation supplier claim. Do not treat the percentage as company-wide energy savings, production uplift, net cash ROI or an AI result. The word ForeSite links a brand family, not proof this equipment used the PDO predictive model.", "question": "Ask for metered before/after energy under comparable duty, uptime and load; retrofit cost and maintenance burden; and the well conditions under which the result transfers.", "ai_role": "No AI identified; practical non-AI energy recovery.", "human": "Installation and operating approval boundaries undisclosed; no autonomous operational change inferred.", "asset": "One unidentified U.S. rod-lift well; basin and operator unknown.", "partner": "Weatherford", "baseline": "Comparison baseline, observation period and duty conditions unknown.", "category": "Energy efficiency", "relationship": "Supplementary non-AI technology context, outside AI adoption and result counts.", "source_id": "WATCH24PM-WFRD-Q3", "source_title": "Weatherford Announces Third Quarter 2026 Results — rod-lift energy recovery", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18871", "publisher": "Weatherford", "locator": "Technology Highlights → Production and Intervention; first U.S. ForeSite Power Regenerative System retrofit paragraph.", "excerpt": "delivering over 50% in energy savings"}]}, "deep_failures_2026_10_24": {"as_of": "2026-10-24", "run": "2026-10-24-deep-failures", "label": "24 October 2026 / What makes AI hard to put into use", "headline": "A good model still needs a workable operating process.", "intro": "False alarms and incomplete oversight are concrete obstacles. This review adds two documented lessons; it did not establish another named operator cancellation or poor financial return.", "count_note": "Two supplementary notes and three original publications added. Core coverage stays at 305 profiles: 219 company-use, 150 operating or scaled, and 63 with a counted reported result. None establishes complete net cash ROI.", "method": "Three Luna researchers used eight search clauses each. The editor checked retained original passages, deduplicated existing surveys and kept inaccessible or unsupported claims out. Public evidence and this search remain incomplete.", "featured_count": 2, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "D24-MARS", "company": "UNEP IMEO / MARS", "record": null, "related_ids": [], "title": "False alarms changed what the team could put into use", "department": "HSE & emissions", "status": "Operational system after redesign; model-test limitations", "summary": "MARS researchers found that testing on small image chips hid the false-alarm burden in full satellite scenes.", "mechanism": "On the same 200-granule EMIT test, a single-model variant averaged 3,565 false alarms, 213 detected events and 98 misses across five runs. Its five-model ensemble produced 1,501 false alarms, 207 detections and 104 misses. This is a trade-off, not error-free detection.", "measure": "These are test counts, not a live alert rate or measured labor savings. The paper reports 11 months of operational use; analysts verify candidate plumes before notifications. No full-cost ROI is disclosed.", "operating_period": "Preprint v2: 24 March 2026; first submitted 11 November 2025. Test uses 105 granules with known plumes and 95 without, containing 311 labeled events. Operational start dates differ by sensor.", "limit": "A successful redesign, not a cancelled operator project. Results span emissions sectors; no new named oil-and-gas operator adoption is established. Raw data and code were not rerun. A separate 17 October conference abstract reports a similar benchmark-to-operation gap; its full chapter was unavailable.", "question": "Before scaling, how many false alarms and missed events will a team handle on the full live workload?", "sources": [{"id": "DEEP24-MARS", "title": "Operational machine learning for remote spectroscopic detection of CH4 point sources", "url": "https://arxiv.org/html/2511.07719v2", "date": "2026-03-24", "publisher": "Růžička and coauthors / arXiv", "locator": "Version 2, 24 March 2026; sections 3.1, 5.1, 6.2 and Table 2; section 6.4 and Table 5.", "excerpt": "the models remain useful in guiding the reviewers", "access": "Editor read original full-text methods, Table 2 and operational workflow. Author preprint, listed as in review on arXiv; raw data/code not replicated. First version 11 November 2025."}, {"id": "DEEP24-SPRINGER", "title": "Reducing False Positives in Operational Remote Sensing: Lessons from Methane Plume Detection", "url": "https://doi.org/10.1007/978-3-032-41086-3_46", "date": "2026-10-17", "publisher": "Springer Nature / Discovery Science 2026", "locator": "Publisher abstract and First Online metadata; chapter pp. 721–736. Full subscription chapter not reviewed.", "excerpt": "generate substantially more operational false positives", "access": "Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass."}], "date": "2026-03-24", "source_type": "Original technical preprint and publisher abstract", "technology_note": "D24-MARS", "technology_label": "Read the test and operating limits", "asset": "Global satellite methane screening; EMIT test granules, not a well population.", "ai_role": "Computer vision ensembles for candidate plume detection; human verification retained.", "partner": "UNEP International Methane Emissions Observatory and research coauthors", "core_company": false, "category": "Model validation and operational workload", "baseline": "Single-model mean across five runs versus ensemble on the same 200-granule, 311-event EMIT test. Counted objects; not percentages or rates.", "relationship": "Supplementary cross-sector methane-monitoring evidence; no core adoption, result or ROI count."}, {"id": "D24-HAVTIL", "company": "Havtil / Norwegian petroleum operators", "record": null, "related_ids": [], "title": "Some operators cannot yet account for all the AI they use", "department": "IT, data & enterprise transformation", "status": "Regulator survey finding; no named failure or incident", "summary": "Norway’s petroleum regulator says operators vary widely in how they oversee AI. Many lack a complete inventory, including AI inside supplier products.", "mechanism": "Some companies reported no AI with safety implications while describing AI in well planning, maintenance or operational advice. Havtil questions whether indirect effects on decisions are being recognized.", "measure": "No respondent count, project denominator, incident rate, cost or return is disclosed on the reviewed page.", "operating_period": "Published 23 October 2026. Survey fieldwork dates and responding companies are not identified on the page; May 2025 guidance is earlier context.", "limit": "This is a regulator’s interpretation of company responses, not proof of a failed deployment, an enforcement finding against every operator or measured poor ROI. It does not rank companies.", "question": "Can the operating team name every AI system it relies on, including supplier features, and who owns its review?", "sources": [{"id": "DEEP24-HAVTIL", "title": "Artificial intelligence (AI) in the petroleum industry is in need of better oversight and management", "url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/reports-from-projects/2026/artificial-intelligence-ai-in-the-petroleum-industry-is-in-need-of-better-oversight-and-management/", "date": "2026-10-23", "publisher": "Norwegian Ocean Industry Authority (Havtil)", "locator": "Varying levels of maturity; An overview is essential for effective risk management; Prudent development and use.", "excerpt": "Many companies lack a comprehensive overview", "access": "Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page."}], "date": "2026-10-23", "source_type": "Original regulator survey summary", "technology_note": "D24-HAVTIL", "technology_label": "Read the oversight findings", "asset": "Norwegian petroleum sector; operator and asset identities not disclosed.", "ai_role": "Inventory and oversight of AI in planning, maintenance and operational decision support.", "partner": "Norwegian Ocean Industry Authority", "core_company": false, "category": "AI inventory and operational oversight", "baseline": "Survey population and response count unknown; no adoption or failure rate calculated.", "relationship": "Sector governance evidence; no named operator adoption, failure count or ROI credit."}]}, "sol_five_rounds_2026_10_24": {"as_of": "2026-10-24", "run": "sol-five-rounds-oct24", "label": "24 October 2026 / Five SOL research rounds", "headline": "A working model can still add little value.", "intro": "An undated West Brae trial is unusually candid about asset fit and value. Two technical studies show where record extraction and engineering-document search went wrong—and how far those tests let us judge.", "count_note": "Three supplementary notes and four publications added. Core coverage remains 305 profiles: 219 company-use, 150 operating or scaled and 63 with a counted reported result. No complete net cash ROI established.", "method": "Five sequential rounds of three gpt-6-sol researchers: 15 assignments and 90 search clauses. The editor checked retained original sources, challenged contrary evidence and kept research tests, historical trials and live deployments distinct. Coverage remains incomplete.", "featured_count": 3, "archive_label": "Earlier source reviews and company context", "findings": [{"id": "SOL5-BRAE", "company": "Marathon Oil / West Brae (historical trial)", "record": null, "related_ids": [], "title": "West Brae’s model worked, but the value case was weak", "department": "Production & field operations", "status": "Undated historical live-asset trial; site-specific limited value", "summary": "NZTC says Intelligent Plant’s ML tool recommended valve settings at Marathon’s West Brae, but appeared to add little value on that already optimized field.", "mechanism": "The eight-well field yielded five usable verification points in four weeks and 12 in a two-year retrospective. Stable operation and larger unrelated production changes made small gains hard to establish.", "measure": "The listed 1% potential and £26.2 million industry-value scenario at 10% uptake are not realized trial benefits or net ROI. Trial costs are undisclosed.", "operating_period": "Undated historical account. Exact trial, publication and end dates unknown; not a new deployment in the rolling 24-month window.", "limit": "No cancellation, negative net cash return or later scale is established. Intelligent Plant has an undated anonymous account with similar trial details that claims benefits; we cannot confirm identical case identity or reconcile it with NZTC’s named assessment.", "question": "Is there enough improvement left in the chosen asset to measure and pay for this tool?", "sources": [{"id": "SOL5-WESTBRAE", "title": "Well Intelligence – Optimised Production", "url": "https://www.netzerotc.com/case-studies/well-intelligence-optimised-production/", "date": "Unknown", "publisher": "Net Zero Technology Centre", "locator": "Project overview; Industry value; Lessons learned. Undated historical account naming Marathon and West Brae.", "excerpt": "the application does not appear to add significant value to Marathon", "access": "Editor read original project summary and lessons. Exact trial/publication dates unknown; website copyright is not a date. Project assessment, not independently audited cash returns."}, {"id": "SOL5-WELLINT-SUPPLIER", "title": "Case Study: Enhancing Production with Well Intelligence AI Model", "url": "https://www.intelligentplant.com/case-study-well-intelligence", "date": "Unknown", "publisher": "Intelligent Plant", "locator": "Challenge and Testing; Historical Data Review; Results and Impact. Anonymous client, no publication date.", "excerpt": "This finding alone justified the deployment of Well Intelligence.", "access": "Editor read original supplier case. Anonymous account shares four-week/~8,000-bbl/day descriptors with NZTC account; same-trial identity not confirmed. No measured financial result or operator corroboration."}], "date": "Unknown", "source_type": "Original project assessment and supplier case", "technology_note": "SOL5-BRAE", "technology_label": "Read the trial and the different accounts", "asset": "West Brae; NZTC describes eight production wells and roughly 8,000 bbl/day. These are historical scope figures.", "ai_role": "Machine learning uses historical production data to recommend valve movements.", "partner": "Intelligent Plant; NZTC; Marathon as named in historical source", "core_company": false, "category": "Asset fit and value verification", "baseline": "One field; five live usable test points and 12 historical points, not 17 independent projects. No confirmed realized uplift.", "relationship": "Supplementary undated historical context; Marathon Oil is not Marathon Petroleum. No core adoption/result/ROI count."}, {"id": "SOL5-RECORDS", "company": "Los Alamos / NETL well-record study", "record": null, "related_ids": [], "title": "A well-record model picked the wrong depth field", "department": "IT, data & enterprise transformation", "status": "Research extraction test; no operating deployment established", "summary": "A 2024 study shows why plausible extraction needs field-level checks. With a simple prompt, Llama 2 70B returned depths within 10 ft for 72 of 150 clean Colorado reports. Revised prompts reached 150/150 on the same corpus.", "mechanism": "On 10 Pennsylvania scanned records, its best prompt located 7/10 wells within 10 m and extracted 9/10 depths within 10 ft. Errors involved source/OCR signs, coordinate conversion and measured depth versus true vertical depth.", "measure": "The paper also tested Llama 3.1 405B on those 10 scans: 9/10 locations and 10/10 depths. These are threshold-based test results, not a current-model ceiling, exact transcription score or live error rate.", "operating_period": "Published 30 December 2024. Historical documents and 2024-era models; no prospective deployment date.", "limit": "Unequal 150- and 10-document cohorts; no independent untouched prompt-selection holdout established. Do not attribute all errors to the LLM, or infer failed operator rollout, measured savings or net ROI.", "question": "Does the test check the right depth, coordinates, units and source quality before values enter an operating record?", "sources": [{"id": "SOL5-WELLRECORDS", "title": "Information extraction from historical well records using a large language model", "url": "https://www.nature.com/articles/s41598-024-81846-5", "date": "2024-12-30", "publisher": "Ma, Santos, Lackey, Viswanathan and O’Malley / Scientific Reports", "locator": "Methods: LLMs performance evaluation; Results: Prompt engineering, Tables 2–3, chain-of-thought and comparison with other LLMs. DOI 10.1038/s41598-024-81846-5.", "excerpt": "none of the prompts resulted in completely correct extraction for the 10 Pennsylvania well records", "access": "Editor read original open-access publisher full text, metric definitions and results. Paper independently cross-checked by two SOL tracks; data/code not rerun. Peer-reviewed publication date 30 December 2024; tests use 2024-era models, not current capability."}], "date": "2024-12-30", "source_type": "Original peer-reviewed technical study", "technology_note": "SOL5-RECORDS", "technology_label": "Read the extraction test and sample limits", "asset": "150 Colorado completion reports and 10 Pennsylvania records; research corpus, not 160 operating wells.", "ai_role": "OCR/text conversion plus Llama-family extraction of location and true vertical depth; prompt variations compared.", "partner": "Authors affiliated with Los Alamos National Laboratory and National Energy Technology Laboratory", "core_company": false, "category": "Document extraction and field validation", "baseline": "Accuracy accepts location within 10 m and depth within 10 ft. Same-corpus prompt comparison, not independently validated production reliability.", "relationship": "Research test only. No named corporate adoption, field failure or ROI count; richer prompts and newer models improved results."}, {"id": "SOL5-PID", "company": "Nigeria LNG-affiliated document-retrieval study", "record": null, "related_ids": [], "title": "Text search missed information in engineering diagrams", "department": "IT, data & enterprise transformation", "status": "Controlled retrieval experiment; no failed assistant deployment established", "summary": "Authors affiliated with Nigeria LNG tested four text-retrieval methods on 60 proprietary oil-and-gas documents: 20 narrative, 20 tabular and 20 piping and instrumentation diagrams. All four struggled with the diagrams.", "mechanism": "The 60 LLM-assisted queries were human-reviewed; people judged pooled search results. On diagrams, the best mean reciprocal rank was 0.106. That ranks the first relevant passage; it is not 10.6% answer accuracy.", "measure": "Across the full corpus, preserving structure led on MRR/NDCG, while recursive splitting led some MAP/F1 measures. No method won every measure. No cash return or final-answer correctness was measured.", "operating_period": "Author manuscript posted 25 March 2026; proceedings marked 2026, exact issue day unknown.", "limit": "Small proprietary corpus; diagram-query count and annotator agreement undisclosed. No live-user, hallucination or adoption rate. Employer affiliation does not establish that these tested configurations were deployed at Nigeria LNG.", "question": "Can the system retrieve the right information from drawings and tables, as well as prose, before anyone relies on its answers?", "sources": [{"id": "SOL5-RAG", "title": "Evaluating Chunking Strategies For Retrieval-Augmented Generation in Oil and Gas Enterprise Documents", "url": "https://arxiv.org/pdf/2603.24556", "date": "2026-03-25", "publisher": "Taiwo and Yusoff / CS&IT proceedings author manuscript", "locator": "Original 11-page PDF: methods PDF pp. 4–7; Tables 1–3 PDF pp. 8–9, printed pp. 64–65. arXiv first submitted 25 March 2026; exact proceedings publication day unknown.", "excerpt": "all four methods demonstrated limited effectiveness on P&IDs", "access": "Editor read original author PDF and visually inspected Tables 1–3. arXiv says manuscript matches CCSEIT 2026 proceedings; raw proprietary corpus and relevance labels unavailable. Author affiliation does not establish deployment or corpus ownership."}], "date": "2026-03-25", "source_type": "Original technical proceedings author manuscript", "technology_note": "SOL5-PID", "technology_label": "Read what the retrieval test measured", "asset": "60 documents and 60 queries;20 P&ID files, not an operator or asset population.", "ai_role": "Azure document extraction; OpenAI text-embedding-3-large; Azure CosmosDB vector retrieval. Answer generation was not evaluated.", "partner": "Samuel Taiwo and Mohd Amaluddin Yusoff; Nigeria LNG Digital and Innovation Department affiliations", "core_company": false, "category": "Retrieval quality and visual documents", "baseline": "Four separately indexed methods, pooled top 10 results, human relevance grades 0/1/2. Corpus and judgments not independently replicated.", "relationship": "Supplementary retrieval-component limitation, not evidence a live Nigeria LNG assistant failed. No core adoption/result/ROI count."}]}, "watch_2026_10_25_am": {"as_of": "2026-10-25", "run": "2026-10-25-am", "label": "25 October 2026 / Morning research", "headline": "PTTEP reports faster purchase-order work. The cash return is still unknown.", "intro": "PTTEP reports a faster purchase-order workflow. BKV, Harbour and ConocoPhillips show the operating systems and exceptions behind everyday work.", "count_note": "One operator-reported AI workflow added; three non-AI technology/process notes kept separate. The sources do not establish complete net cash ROI.", "method": "Three Luna researchers used six search clauses each. The editor ran four regional searches and checked the retained findings against original sources, including the relevant PTTEP annual-report page. These are earlier disclosures we had missed. The search did not substantiate another named failed AI rollout or a complete calculation of disappointing cash returns. Coverage is not exhaustive.", "featured_count": 4, "archive_label": "Earlier source reviews", "findings": [{"id": "W25AM-PTTEP-XBOT", "company": "PTTEP", "title": "PTTEP reports faster purchase-order processing with X.Bot", "record": "AI-306", "related_ids": ["AI-306"], "sources": [{"id": "WATCH25AM-PTTEP-XBOT", "title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "date": "Unknown", "publisher": "PTTEP", "locator": "Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate.", "excerpt": "average processing time by five hours per purchase order", "access": "Editor downloaded original 557-page operator PDF, read and visually inspected the complete relevant page on 25 October 2026. Report covers 2025; exact publication date and six-month evaluation dates unknown. Operator claims, not independent outcomes.", "format": "pdf"}, {"id": "WATCH25AM-PTTEP-OPEX", "title": "PTTEP organizes OpEx Day 2025, advancing operational excellence", "url": "https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence", "date": "2025-07-09", "publisher": "PTTEP", "locator": "Press release, procurement AI sentence in operational-excellence examples.", "excerpt": "application of AI to streamline procurement processes", "access": "Editor read original issuer release and displayed 9 July 2025 date. It supports procurement AI use, not X.Bot identity or its measured results.", "type": "Operator press release"}], "source_type": "Operator annual report", "department": "Supply chain & logistics", "status": "Operating procurement workflow; scale undisclosed", "summary": "PTTEP says Procurement X.Bot combines AI and automation in purchase-order processing.", "mechanism": "The annual report describes improved processing, decision support and compliance. It does not identify model, vendor, user population or approval permissions.", "measure": "Operator reports five fewer processing hours per purchase order and 50% less pending-order backlog within six months. These are system-level process measures, not AI-only cash savings.", "operating_period": "Report covers 2025; exact publication day, go-live and six-month comparison dates unknown. A 9 July 2025 release independently describes procurement AI use without naming X.Bot.", "limit": "No PO count, baseline hours, starting backlog, comparison design, implementation costs or cash return. AI and conventional automation are not isolated. Adjacent 172-supplier event, green-spend and vessel-fuel figures do not measure X.Bot.", "ai_role": "AI plus automation; no LLM, autonomous purchasing or X.brain integration established.", "human": "Who approves or releases purchase orders, reviews the output or overrides it is not disclosed.", "asset": "Procurement process; geography, business-unit and supplier coverage unknown.", "partner": "Procurement X.Bot; supplier/model not named.", "baseline": "Average per-PO time and six-month backlog change; baseline values and cohort unknown.", "question": "Which steps changed, who approves orders, and do the time gains persist after all operating costs?", "problem": "Purchase-order processing and a backlog of pending orders; underlying task bottlenecks are not described.", "category": "Procurement AI and automation", "relationship": "Distinct from AI-245 internal audit and AI-297 enterprise engine. No evidence that X.Bot is one of its eight planned agents.", "core_company": true, "source_id": "WATCH25AM-PTTEP-XBOT", "source_title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "publisher": "PTTEP", "date": "Unknown", "locator": "Section 9.4.2, printed p.284 / physical PDF p.273 of 557. Procurement X.Bot paragraph; adjacent vessel and green-procurement figures are separate."}, {"id": "W25AM-BKV-RUNWAY", "company": "BKV Corporation", "title": "BKV rebuilt its field-data system around how operators work", "record": null, "related_ids": [], "sources": [{"id": "WATCH25AM-BKV-RUNWAY", "title": "BKV 2025 Sustainability Report — Project Runway", "url": "https://www.bkv.com/2025-sustainability-report-governance", "date": "Unknown", "publisher": "BKV Corporation", "locator": "Governance: Project Runway; Informed by field operations; Collaboration beyond implementation.", "excerpt": "the consolidated Barnett platform went live on November 19, 2025", "access": "Editor read original operator page on 25 October 2026. Publication day not displayed; report landing page CEO letter dated September 2026. Metrics are issuer-reported."}, {"id": "WATCH25AM-BKV-REPORT", "title": "BKV 2025 Sustainability Report — letter date", "url": "https://www.bkv.com/2025-sustainability-report", "date": "Unknown", "publisher": "BKV Corporation", "locator": "Letter from CEO footer: September 2026.", "excerpt": "Letter from the CEO | September 2026", "access": "Editor read original report landing page; letter month is not exact report publication date.", "type": "Operator report landing page"}], "source_type": "Operator annual report", "department": "IT, data & enterprise transformation", "status": "Operating SCADA foundation; not AI", "summary": "BKV consolidated four inherited SCADA/CygNet environments. Its Barnett platform went live 19 November 2025.", "mechanism": "Field staff reviewed terminology and screens. Peer-led training and field ride-alongs supported the change.", "measure": "BKV reports 2.5 million data points standardized, over 250,000 obsolete points removed, polling 60 to 15 minutes, and no production downtime. Acquisition integration reportedly fell from 6–8 months to 6–8 weeks.", "operating_period": "Go-live 19 November 2025. Report letter is dated September 2026; exact report publication day unknown.", "limit": "Issuer-reported measures; no costs, controlled evaluation, cash benefit or AI use established. About 1,100 Bedrock wells were positioned for onboarding, not confirmed live at the go-live date.", "ai_role": "Conventional operating-data infrastructure; not AI adoption.", "human": "Field operators use the system; autonomous control and permission boundaries undisclosed.", "asset": "Barnett Shale; broader portfolio and acquisition populations differ.", "partner": "SCADA/CygNet; external implementation partner unnamed.", "baseline": "Four legacy environments; timings are reported process comparisons.", "question": "Does the common data model survive the next acquisition, and how much ongoing support does it need?", "problem": "Does the common data model survive the next acquisition, and how much ongoing support does it need?", "category": "Operating-system transformation", "relationship": "Supplementary non-AI operating foundation; not a new AI register profile.", "core_company": false, "source_id": "WATCH25AM-BKV-RUNWAY", "source_title": "BKV 2025 Sustainability Report — Project Runway", "url": "https://www.bkv.com/2025-sustainability-report-governance", "publisher": "BKV Corporation", "date": "Unknown", "locator": "Governance: Project Runway; Informed by field operations; Collaboration beyond implementation."}, {"id": "W25AM-HARBOUR-COUPA", "company": "Harbour Energy", "title": "Harbour documents the exceptions in its supplier invoice workflow", "record": null, "related_ids": [], "sources": [{"id": "WATCH25AM-HARBOUR-COUPA", "title": "Harbour Energy Coupa Supplier Portal FAQ, version 4", "url": "https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf", "date": "2025-06", "publisher": "Harbour Energy", "locator": "PDF p.1 version 4, June 2025 and PO process; pp.2–4 permissions, invoices, receipts and disputes; p.5 refers to ex-Wintershall Dea vendor guides.", "excerpt": "All new purchase orders of Harbour Energy are generated through the Coupa platform.", "access": "Editor read original 5-page operator PDF. Exact revision day, go-live and legal-entity scope unknown. The linked vendor page separately retains SAP Business Network.", "format": "pdf"}, {"id": "WATCH25AM-HARBOUR-VENDORS", "title": "Harbour vendor relations — distinct supplier channels", "url": "https://www.harbourenergy.com/vendor-relations/", "date": "Unknown", "publisher": "Harbour Energy", "locator": "Separate sections: vendors using SAP Business Network; ex-Wintershall Dea vendors.", "excerpt": "Applicable to ex-Wintershall Dea vendors", "access": "Editor read original operator page on 25 October 2026. Undated current channel guidance; SAP start November 2022 is older context, not new AI.", "type": "Operator supplier webpage"}], "source_type": "Operator supplier guide", "department": "Supply chain & logistics", "status": "Documented supplier transaction channel; not AI", "summary": "Harbour’s June 2025 Coupa guide covers purchase orders, confirmations, invoices and status tracking.", "mechanism": "It documents access roles, drafts, receipt matching, disputes and credit/reissue paths. The guide points to ex-Wintershall Dea materials; a separate channel uses SAP Business Network.", "measure": "No measured time, error, cash saving or ROI result. Free supplier access does not establish the operator’s total cost.", "operating_period": "Guide version 4, June 2025; exact day, system start and entity rollout unknown.", "limit": "Do not extend the guide’s all-new-PO wording to every Harbour entity or infer that Coupa replaced SAP. Exception instructions do not prove high failure rates.", "ai_role": "Supplier process digitization; no AI established.", "human": "Suppliers submit; Harbour approves service receipts and can resolve or reject invoices.", "asset": "Affected supplier-channel scope; no asset, user or invoice denominator.", "partner": "Coupa; separate SAP Business Network channel.", "baseline": "Published process instructions, not before/after evaluation.", "question": "Where do invoices stall, which exceptions require people, and which supplier channel applies?", "problem": "Where do invoices stall, which exceptions require people, and which supplier channel applies?", "category": "Procure-to-pay foundation", "relationship": "Supplementary administrative workflow, not AI adoption or failed deployment.", "core_company": false, "source_id": "WATCH25AM-HARBOUR-COUPA", "source_title": "Harbour Energy Coupa Supplier Portal FAQ, version 4", "url": "https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf", "publisher": "Harbour Energy", "date": "2025-06", "locator": "PDF p.1 version 4, June 2025 and PO process; pp.2–4 permissions, invoices, receipts and disputes; p.5 refers to ex-Wintershall Dea vendor guides."}, {"id": "W25AM-COP-GEP", "company": "ConocoPhillips", "title": "ConocoPhillips shows how suppliers enter orders and invoices", "record": null, "related_ids": [], "sources": [{"id": "WATCH25AM-COP-GEP", "title": "ConocoPhillips GEP SMART Introduction", "url": "https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf", "date": "2024-12", "publisher": "ConocoPhillips", "locator": "PDF p.1 updated December 2024; p.3 transition target; p.5 intended benefits; p.6 supplier invoice flow.", "excerpt": "all business units should have transitioned to GEP SMART", "access": "Editor read original 7-page issuer PDF. January 2025 is a target, not verified unit-by-unit completion. Exact revision day unknown.", "format": "pdf"}, {"id": "WATCH25AM-COP-VENDORS", "title": "ConocoPhillips technical upgrades to supplier systems", "url": "https://vendors.conocophillips.com/supplier-guide/technical-upgrades-to-supplier-systems/", "date": "Unknown", "publisher": "ConocoPhillips", "locator": "GEP portal/EDI and Enverus routes; supplier training and implemented SSIS support.", "excerpt": "Exceptions to the invoicing process are identified", "access": "Editor read original live supplier page on 25 October 2026. Page retains future-tense January 2025 language, so it cannot verify completion date. No publication date displayed.", "type": "Operator supplier webpage"}], "source_type": "Operator supplier guide", "department": "Supply chain & logistics", "status": "Supplier workflow documented; transition completion unverified", "summary": "ConocoPhillips’ December 2024 guide chose GEP SMART for supplier registration, PO acknowledgement and invoicing.", "mechanism": "Suppliers register, accept PO terms, create invoices or service confirmations and inspect remittance. Current guidance offers portal, integration and Enverus submission routes.", "measure": "Faster approvals and better order accuracy are intended benefits; no measured result or ROI.", "operating_period": "Updated December 2024; January 2025 transition target. Undated live guidance still uses that future-tense wording; completion by business unit unknown.", "limit": "No transaction population, costs or actual rollout completion. Supplier fee waivers do not establish operator cost. An available Enverus route is not exclusive-platform adoption.", "ai_role": "Digital procurement and invoice handling; no AI disclosed.", "human": "Supplier acknowledgement and invoicing documented; internal automated approval authority not established.", "asset": "Supplier process; geography and business-unit rollout populations unknown.", "partner": "GEP SMART; Enverus alternative channel; Atlassian supplier-support portal.", "baseline": "Process diagram and target, not measured operating comparison.", "question": "Which units and invoice routes are live, and who owns the exceptions?", "problem": "Which units and invoice routes are live, and who owns the exceptions?", "category": "Procure-to-pay foundation", "relationship": "Supplementary platform/process context, separate from existing AI records.", "core_company": false, "source_id": "WATCH25AM-COP-GEP", "source_title": "ConocoPhillips GEP SMART Introduction", "url": "https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf", "publisher": "ConocoPhillips", "date": "2024-12", "locator": "PDF p.1 updated December 2024; p.3 transition target; p.5 intended benefits; p.6 supplier invoice flow."}]}, "watch_2026_10_25_pm": {"as_of": "2026-10-25", "run": "2026-10-25-pm", "label": "25 October 2026 / Evening research", "headline": "Ecopetrol reports an IT gain. Its wider AI rollout still has limits.", "intro": "A reported benefit, a supplier’s operating account and a software selection tell us different things. Here is what each source supports.", "count_note": "Three supplementary notes. Core profile, deployment and outcome counts are unchanged. These are earlier disclosures we had missed.", "method": "We checked the original Ecopetrol annual report, the supplier’s Dixstone announcement and the original YPF interview. None establishes complete net returns or a new failed deployment. These are selected earlier disclosures, and coverage is incomplete.", "featured_count": 3, "archive_label": "Earlier source reviews", "findings": [{"id": "W25PM-ECO-AIOPS", "company": "Ecopetrol", "title": "Ecopetrol reports IT gains while its wider AI program still faces data limits", "related_ids": ["AI-147"], "sources": [{"id": "SW6-004", "title": "Ecopetrol 2025 Integrated Management Report — digital transformation and AIOps", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "date": "2026 (exact publication day unknown)", "publisher": "Ecopetrol", "locator": "Printed p.314 / physical PDF p.315 of 482, Transformación Digital. AIOps, self-service and AI maturity are separate passages.", "excerpt": "la disponibilidad de los servicios críticos de TI al 99.52%", "access": "Editor downloaded the original report, read the complete relevant page and visually inspected it on 25 October 2026. Reuses existing source SW6-004; earlier claim-level reviews remain separate.", "format": "pdf", "existing": true}], "source_type": "Operator annual report", "department": "IT, data & enterprise transformation", "status": "AI in IT operations adopted; wider AI maturity gaps and a 2027 roadmap reported", "summary": "Ecopetrol says using AI in IT operations raised critical-service availability to 99.52% and generated more than US$1.5 million in savings.", "mechanism": "The same page says the AI ecosystem maturity review found limits in data access and autonomy. A broader operating model with integrated AI remains a 2027 goal.", "measure": "Company-reported availability level and savings; starting availability, service population, cost perimeter and savings period are not specified. An adjacent 40% digital self-service resolution rate is separate and is not explicitly attributed to AIOps.", "operating_period": "The report covers 2025; exact publication day and AIOps start/evaluation dates are unknown. The 2027 target is prospective.", "limit": "No all-in cost, cash reconciliation, comparison design or complete net ROI. The US$1.5 million-plus may overlap the existing AI-147 portfolio benefits; do not add them. Neither 160 digital initiatives nor more than 100 operating solutions is a count of AI deployments.", "ai_role": "AIOps is explicitly AI for IT operations; the report does not identify models, vendors or an LLM. Self-service and infrastructure upgrades are separate.", "human": "IT approval, escalation, override and automated remediation permissions are not disclosed.", "asset": "Critical IT services; business-unit, service, user and asset populations unspecified.", "partner": "No AIOps supplier or model named.", "baseline": "99.52% is an availability level, not a 99.52% improvement. Baseline and measurement window unknown.", "question": "Which services improved, what changed against the prior baseline, and how much of the saving remains after running costs?", "category": "IT reliability and AI rollout limits", "relationship": "Further context for AI-147, not another counted core profile. AIOps results must not be added to portfolio benefits or assigned to the upstream planner.", "core_company": true, "logo_available": true, "record": null, "problem": "Which services improved, what changed against the prior baseline, and how much of the saving remains after running costs?", "source_id": "SW6-004", "source_title": "Ecopetrol 2025 Integrated Management Report — digital transformation and AIOps", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "publisher": "Ecopetrol", "date": "2026 (exact publication day unknown)", "locator": "Printed p.314 / physical PDF p.315 of 482, Transformación Digital. AIOps, self-service and AI maturity are separate passages."}, {"id": "W25PM-YPF-EKONS", "company": "YPF", "title": "YPF’s fuel-blending tool supplies estimates to refinery operators", "related_ids": [], "sources": [{"id": "WATCH25PM-YPF-EKONS", "title": "YPF refinery blending — interview with BCG project lead Roberto Rodrigo", "url": "https://mase.lmneuquen.com/empresas/ia-ypf-la-tecnologia-que-gano-el-premio-digital-del-ano-y-ya-se-aplica-la-refineria-la-plata-n1253008", "date": "Unknown", "publisher": "Más Energía / LM Neuquén; Martina Mongelluzzo", "locator": "Sections Del laboratorio a la operación en tiempo real and Argentina, el primer caso de implementación.", "excerpt": "la solución puede reducir hasta un 50% el tiempo de ejecución de los blending", "access": "Editor read the original interview. The exact date is not displayed in its retrieved body; a syndicated copy is dated 11 September 2026. Original named supplier interview, not operator confirmation or independent performance validation."}], "source_type": "Original supplier interview", "department": "Refining, LNG & process operations", "status": "Supplier-reported use at one refinery; still being adjusted", "summary": "A Boston Consulting Group (BCG) project lead describes Ekons Matrix AI in use at YPF’s La Plata refinery in Argentina.", "mechanism": "AI and process-simulation rules estimate the properties of streams being mixed into fuel. The estimates support control-room staff; autonomous control is not established.", "measure": "The supplier says blending execution time could fall by up to 50%. This is a potential maximum, without a measured achieved average, comparison cohort or observation period.", "operating_period": "The interview describes application starting March–April 2026, about six months of implementation and continuing adjustments. A syndicated copy is dated 11 September 2026; the retrieved original does not display its exact date.", "limit": "Original interview with the supplier project lead, not an operator filing or independent result. No realized saving, quality/error measure, cost or net ROI. No stream, tank, blend, user or throughput denominator. Further rollout is an ambition. Separate from the earlier YPF cataloging and inspection profile.", "ai_role": "AI plus process simulation; no LLM disclosed.", "human": "Control-room staff use the information. Approval, override and model-validation responsibilities are not described.", "asset": "La Plata refinery, Argentina; one implementation described.", "partner": "BCG / Ekons Matrix AI and YPF, as described in the interview.", "baseline": "No measured before/after cohort; up to 50% is not an observed average.", "question": "Can YPF show achieved blend-time and quality results over enough batches to separate the tool from other operating changes?", "category": "Refinery decision support", "relationship": "Distinct from AI-149 contract cataloging, customer information and refinery inspection. Linked as company context only; it does not change that profile’s outcomes or stage.", "core_company": true, "logo_available": true, "record": null, "problem": "Can YPF show achieved blend-time and quality results over enough batches to separate the tool from other operating changes?", "source_id": "WATCH25PM-YPF-EKONS", "source_title": "YPF refinery blending — interview with BCG project lead Roberto Rodrigo", "url": "https://mase.lmneuquen.com/empresas/ia-ypf-la-tecnologia-que-gano-el-premio-digital-del-ano-y-ya-se-aplica-la-refineria-la-plata-n1253008", "publisher": "Más Energía / LM Neuquén; Martina Mongelluzzo", "date": "Unknown", "locator": "Sections Del laboratorio a la operación en tiempo real and Argentina, el primer caso de implementación."}, {"id": "W25PM-DIXSTONE-IFS", "company": "Dixstone", "title": "Dixstone selects shared project and fleet software; go-live remains unverified", "related_ids": [], "sources": [{"id": "WATCH25PM-DIXSTONE-IFS", "title": "Dixstone selects IFS Cloud for offshore services", "url": "https://www.ifs.com/pl/insights/news/dixstone-selects-ifs-cloud-to-power-global-offshore-operations", "date": "2025-12-09", "publisher": "IFS", "locator": "Dated release: selection, project controls and fleet maintenance paragraphs; About Dixstone separates affiliate status from Perenco E&P.", "excerpt": "Dixstone has selected IFS Cloud", "access": "Editor read the original IFS release dated 9 December 2025. A vendor selection announcement with a customer quote; future capabilities and company footprint are not completed rollout or outcomes."}], "source_type": "Supplier announcement", "department": "IT, data & enterprise transformation", "status": "Software selected; implementation completion and AI use unverified", "summary": "IFS says Dixstone selected IFS Cloud to manage project finances and fleet maintenance.", "mechanism": "The intended system brings budgets, schedules, resources and asset data together. The release proposes predictive maintenance for rigs, vessels and platforms.", "measure": "No measured operating result, cost or return. Ten countries and more than 2,200 professionals describe the stated company footprint, not active users, sites live or AI adoption.", "operating_period": "Selection announcement dated 9 December 2025. Go-live and implementation milestones are not disclosed.", "limit": "Vendor announcement, not proof of operating AI or delivered savings. No model, asset rollout, approval boundary or net ROI. Dixstone is a Perenco affiliate consolidating offshore services; this is not a Perenco E&P field deployment.", "ai_role": "Industrial AI is supplier positioning; no specific deployed AI function is established.", "human": "Approval, escalation and automatic maintenance-action boundaries are not disclosed.", "asset": "Offshore-services rigs, vessels and platforms; individual rollout assets unspecified.", "partner": "IFS Cloud; Dixstone is described as a Perenco affiliate and Perrodo family company.", "baseline": "Selection and intended scope, without a measured operating comparison.", "question": "Which assets and workflows are live, what still needs implementation, and where does AI actually change the work?", "category": "Offshore-services software foundation", "relationship": "Supplementary services-company transformation context, not new core AI adoption or Perenco field performance.", "core_company": false, "logo_available": false, "record": null, "problem": "Which assets and workflows are live, what still needs implementation, and where does AI actually change the work?", "source_id": "WATCH25PM-DIXSTONE-IFS", "source_title": "Dixstone selects IFS Cloud for offshore services", "url": "https://www.ifs.com/pl/insights/news/dixstone-selects-ifs-cloud-to-power-global-offshore-operations", "publisher": "IFS", "date": "2025-12-09", "locator": "Dated release: selection, project controls and fleet maintenance paragraphs; About Dixstone separates affiliate status from Perenco E&P."}]}, "operating_results_oct25": {"as_of": "2026-10-25", "run": "2026-10-25-operating-results", "label": "25 October 2026 / Operating results", "headline": "Start with the operating change, then ask what the number means.", "intro": "Four more concrete cases, plus a closer look at SLB’s chemical-dosing result. Each has a different population, period and comparison.", "count_note": "Four supplementary notes; the existing SLB chemistry note is expanded. Core profile and outcome counts are unchanged. Undated cases are labeled; this is not an exhaustive census.", "method": "Three bounded Luna searches and original-source review. AI, conventional automation and non-AI engineering remain separate. No retained case establishes independently validated, complete net ROI.", "featured_count": 4, "archive_label": "Earlier source reviews", "findings": [{"id": "RESULTS-PERENCO", "company": "Perenco", "title": "Perenco reports faster drilling across three wells", "department": "Drilling & completions", "summary": "SLB reports a 19% increase in rate of penetration in Perenco’s 2025 offshore Congo campaign.", "mechanism": "DrillOps and Neuro automated drilling and directional work; SLB describes an automatic safe-state response to a pack-off event. General override and approval rules are not supplied.", "measure": "Three wells; 97% of on-bottom operations automated. These are speed and control measures, not cash return.", "operating_period": "2025 three-well campaign. Exact publication day and campaign dates unknown.", "limit": "Raw drilling-speed baseline, matched offset population, complete costs and AI-only contribution unknown.", "ai_role": "AI-driven drilling automation; no LLM identified.", "asset": "Offshore Republic of the Congo; individual assets unnamed.", "partner": "SLB / DrillOps / Neuro", "baseline": "Manual operations are discussed; raw ROP baseline and comparison matching unknown.", "question": "Did the gain hold on more wells after accounting for geology, equipment and full deployment costs?", "sources": [{"id": "RESULTS-PERENCO", "title": "Perenco enhances drilling efficiency with advanced automation solutions", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/perenco-enhances-drilling-efficiency-with-advanced-automation-solutions", "date": "Unknown", "publisher": "SLB", "locator": "Summary and Results; three-well 2025 campaign.", "excerpt": "achieved a 19% increase in rate of penetration", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "format": "html"}], "category": "Operating result", "core_company": false, "logo_available": false, "related_ids": [], "record": null, "date": "Unknown", "status": "Supplier-reported operating deployment", "source_type": "Supplier customer case study", "relationship": "Supplementary evidence; not a new core-profile, adoption or numerical-result count. Calendar timing and evidence quality remain explicit.", "human": "DrillOps and Neuro automated drilling and directional work; SLB describes an automatic safe-state response to a pack-off event. General override and approval rules are not supplied.", "source_id": "RESULTS-PERENCO", "source_title": "Perenco enhances drilling efficiency with advanced automation solutions", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/perenco-enhances-drilling-efficiency-with-advanced-automation-solutions", "publisher": "SLB", "locator": "Summary and Results; three-well 2025 campaign.", "excerpt": "achieved a 19% increase in rate of penetration", "problem": "Did the gain hold on more wells after accounting for geology, equipment and full deployment costs?"}, {"id": "RESULTS-WAFI", "company": "Wafi Energy Pakistan", "title": "Invoice handling fell from about five minutes to under one", "department": "Finance & accounting", "summary": "SpendConsole reports invoice processing falling from about five minutes to under one in automated scenarios at Wafi Energy Pakistan.", "mechanism": "AI document capture, rules and SAP workflows handle routine invoices; finance staff handle exceptions. Autonomous payment approval is not established.", "measure": "The vendor separately reports 50–70% productivity improvement and full go-live in about 12 weeks.", "operating_period": "Publication and operating calendar dates unknown; results observed within weeks of go-live. Twelve weeks describes implementation, not the evaluation period.", "limit": "Invoice volume, automated-scenario share, precise measurement window and full costs unknown. This is the Pakistan Shell-brand licensee, not Shell-wide adoption.", "ai_role": "AI document processing plus rules and process changes; effects not isolated.", "asset": "Wafi Energy Pakistan finance function.", "partner": "SpendConsole / SAP", "baseline": "Approximately five minutes of manual handling versus under one in automated scenarios; not every invoice.", "question": "What share of invoices qualifies, and how much staff time is actually released after exception handling?", "sources": [{"id": "RESULTS-WAFI", "title": "From manual accounts payable to AI-powered finance in 12 weeks", "url": "https://spendconsole.ai/success-stories/from-manual-accounts-payable-to-ai-powered-finance-in-12-weeks/", "date": "Unknown", "publisher": "SpendConsole", "locator": "Client profile; solution; results; automated-scenario qualification.", "excerpt": "under a minute in automated scenarios", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "format": "html"}], "category": "Operating result", "core_company": false, "logo_available": false, "related_ids": [], "record": null, "date": "Unknown", "status": "Supplier-reported operating deployment", "source_type": "Supplier customer case study", "relationship": "Supplementary evidence; not a new core-profile, adoption or numerical-result count. Calendar timing and evidence quality remain explicit.", "human": "AI document capture, rules and SAP workflows handle routine invoices; finance staff handle exceptions. Autonomous payment approval is not established.", "source_id": "RESULTS-WAFI", "source_title": "From manual accounts payable to AI-powered finance in 12 weeks", "url": "https://spendconsole.ai/success-stories/from-manual-accounts-payable-to-ai-powered-finance-in-12-weeks/", "publisher": "SpendConsole", "locator": "Client profile; solution; results; automated-scenario qualification.", "excerpt": "under a minute in automated scenarios", "problem": "What share of invoices qualifies, and how much staff time is actually released after exception handling?"}, {"id": "RESULTS-APOLLO", "company": "Apollo Resources / unnamed SENM operator", "title": "A chemical-treatment redesign cut consumption without identified AI", "department": "Production & field operations", "summary": "Apollo reports 65% lower chemical consumption and $210,000 lower monthly H2S treatment cost at an unnamed southeastern New Mexico facility.", "mechanism": "A flow-meter-linked controller, new chemistry and a relocated injection point changed treatment. Key operator personnel received online write access.", "measure": "Zero out-of-spec H2S events during a 30-day observation. The monthly cost claim is not annualized here.", "operating_period": "Thirty-day observation; publication and operating calendar dates unknown.", "limit": "Population, throughput, cost calculation, installation costs and longer-term durability unknown. New chemistry and equipment prevent attributing the result to software alone.", "ai_role": "No AI identified; useful comparison for the same operating problem.", "asset": "Unnamed facility, southeastern New Mexico.", "partner": "Apollo Resources", "baseline": "Prior treatment versus redesigned treatment; comparable throughput and pre-period observation length unknown.", "question": "How much could treatment redesign achieve before adding AI, and what does it cost to sustain?", "sources": [{"id": "RESULTS-APOLLO", "title": "HSV-01 chemical treatment case study", "url": "https://apolloresourcesllc.net/apollo-resources-hsv-01-oil-treating-chemistry-effectively-eliminates-30000-ppm-of-h2s-in-senm/", "date": "Unknown", "publisher": "Apollo Resources", "locator": "Solution and Results; 30-day observation.", "excerpt": "Chemical consumption decreased by 65%", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "format": "html"}], "category": "Non-AI comparison", "core_company": false, "logo_available": false, "related_ids": [], "record": null, "date": "Unknown", "status": "Supplier-reported operating deployment", "source_type": "Supplier customer case study", "relationship": "Supplementary evidence; not a new core-profile, adoption or numerical-result count. Calendar timing and evidence quality remain explicit.", "human": "A flow-meter-linked controller, new chemistry and a relocated injection point changed treatment. Key operator personnel received online write access.", "source_id": "RESULTS-APOLLO", "source_title": "HSV-01 chemical treatment case study", "url": "https://apolloresourcesllc.net/apollo-resources-hsv-01-oil-treating-chemistry-effectively-eliminates-30000-ppm-of-h2s-in-senm/", "publisher": "Apollo Resources", "locator": "Solution and Results; 30-day observation.", "excerpt": "Chemical consumption decreased by 65%", "problem": "How much could treatment redesign achieve before adding AI, and what does it cost to sustain?"}, {"id": "RESULTS-WEATHERFORD", "company": "Weatherford / unnamed trial operators", "title": "Targeted chemical placement is another route to less waste", "department": "Production & field operations", "summary": "A Weatherford-hosted technical-paper abstract reports 65% less chemical use per well and 52% lower operating expense in capillary-injection field trials.", "mechanism": "Simulation-guided placement, a capillary string and an injection valve deliver chemicals to the treatment point. No AI is identified.", "measure": "Results are from the public abstract of SPE-227643-MS; the full paper was not reviewed.", "operating_period": "Paper dated 16 September 2025. Trial dates and evaluation duration unknown.", "limit": "Operator, geography, well population, raw baseline, cost perimeter and installation costs unknown. This does not establish all-in ROI or scaled deployment.", "ai_role": "No AI identified; precision placement and engineering comparison.", "asset": "Unnamed field trials; geography unspecified.", "partner": "Weatherford", "baseline": "Traditional cyclic treatment; raw volumes and cost baseline not supplied.", "question": "Does targeted placement outperform surface optimization after installation and maintenance costs?", "sources": [{"id": "RESULTS-WEATHERFORD", "title": "Maximizing chemical efficiency with rigless precision-based capillary injection", "url": "https://www.weatherford.com/documents/papers/2025/maximizing-chemical-efficiency-and-reducing-environmental-footprint-with-rigless,-precision-based-ca/", "date": "2025-09-16", "publisher": "Weatherford / SPE", "locator": "Public abstract, SPE-227643-MS; field-trial results.", "excerpt": "a 65% reduction in chemical usage per well", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "format": "html"}], "category": "Non-AI comparison", "core_company": false, "logo_available": false, "related_ids": [], "record": null, "date": "2025-09-16", "status": "Supplier-reported operating deployment", "source_type": "Supplier customer case study", "relationship": "Supplementary evidence; not a new core-profile, adoption or numerical-result count. Calendar timing and evidence quality remain explicit.", "human": "Simulation-guided placement, a capillary string and an injection valve deliver chemicals to the treatment point. No AI is identified.", "source_id": "RESULTS-WEATHERFORD", "source_title": "Maximizing chemical efficiency with rigless precision-based capillary injection", "url": "https://www.weatherford.com/documents/papers/2025/maximizing-chemical-efficiency-and-reducing-environmental-footprint-with-rigless,-precision-based-ca/", "publisher": "Weatherford / SPE", "locator": "Public abstract, SPE-227643-MS; field-trial results.", "excerpt": "a 65% reduction in chemical usage per well", "problem": "Does targeted placement outperform surface optimization after installation and maintenance costs?"}], "lead": {"date": "2026-10-23", "source_type": "Supplier earnings release / customer account", "partner": "SLB", "record_type": "company_adoption", "id": "W23AM-CHEMISTRY", "company": "SLB / unnamed Permian operator", "record": null, "related_ids": [], "title": "Better chemical dosing reduced waste across an existing well fleet", "department": "Production & field operations", "status": "Supplier-reported deployment; operator identity withheld", "summary": "SLB reports overtreatment falling from 17% to 3% across nearly 1,400 Permian wells over a 20-month program.", "mechanism": "Sensors feed ML detection; weekly team reviews support remote adjustments. Detailed approval and fallback permissions are not disclosed.", "measure": "SLB-reported net savings: $4.4 million in the first nine months, attributed to lower chemical spend. The meaning of net and complete deployment costs are not disclosed.", "operating_period": "Performance comparison: 20 months. Savings: first nine months. Exact operating start/end dates and case-study publication date unknown; repeated in Q3 results on 23 October 2026.", "limit": "An anonymous supplier account, without a controlled comparison or complete net ROI. The percentage change concerns the share of overtreated wells, not chemical volume or total cost. The case study and Q3 release are the same publisher, not independent corroboration.", "question": "What costs were included, and did lower overtreatment persist without compromising corrosion or scale protection?", "sources": [{"id": "RESULTS-SLB-CHEM", "title": "Permian field chemical automation case study", "url": "https://www.slb.com/resource-library/case-study-with-navigation/production-chemicals/permian-basin-field-automations-cs", "date": "Unknown", "publisher": "SLB", "locator": "Results and solution sections; 20-month dosing comparison and first-nine-month savings are different periods.", "excerpt": "USD 4.4 million in net savings in the first 9 months", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "format": "html"}, {"id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "date": "2026-10-23", "publisher": "SLB", "locator": "PDF p. 14: Technology Highlights / United States production chemistry paragraph.", "excerpt": "AI-driven injection control", "access": "Editor read the original issuer earnings-release HTML and relevant original PDF text. PDF p. 14 and p. 16 are physical page numbers (zero-based viewer P13/P15). Direct local PDF download failed; no call transcript reviewed. Supplier-reported customer claims, not independent validation.", "format": "pdf", "existing": true}], "asset": "Nearly 1,400 producing Permian wells; operator and individual assets unnamed.", "ai_role": "AI injection control plus ML anomaly detection; human review cadence disclosed, execution permissions and fallback unspecified.", "category": "Chemical dosing", "baseline": "Share of overtreated wells: 17% before, 3% after. The case study also says on- or above-target dosing; Q3 uses on-target. These labels are not silently merged.", "relationship": "Context only; anonymous supplier account excluded from named operator adoption and outcome totals.", "core_company": false, "secondary": [], "technology_note": "W23AM-CHEMISTRY", "technology_label": "Read the source and limits", "source_id": "WATCH23AM-SLB-Q3", "source_title": "SLB third-quarter 2026 results — selected operating and digital highlights", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "publisher": "SLB", "locator": "PDF p. 14: Technology Highlights / United States production chemistry paragraph.", "excerpt": "AI-driven injection control", "logo_available": false}}, "measured_results_deep26": {"as_of": "2026-10-26", "run": "2026-10-26-measured-results", "label": "26 October2026 / Results and setbacks", "headline": "A better model still has to make the job better.", "intro": "One use case was dropped, another offered little additional value, and several tests exposed errors that a headline score could hide. Those are different findings, with different consequences.", "count_note": "Four supplementary source reviews, two newly found technical accounts and two deeper readings of existing evidence. No new core deployment or ROI count. Selected coverage remains incomplete.", "method": "Three Luna assignments used 10 search clauses each. The editor read retained original sources and rechecked selected earlier cases. Tests, historical trials and live operating results remain separately labeled.", "featured_count": 4, "archive_label": "Earlier source reviews", "findings": [{"id": "D26-TOPTAL", "company": "Toptal / unnamed oil-and-gas technology company", "title": "A title model needed attorney-grounded regression tests", "department": "Land & legal", "status": "Provider-reported technical evaluation; live test population unverified", "summary": "Toptal reports false positives falling from about 27% to 2%, with recall about 93%; conveyance-party capture rose from about 85% to 96%.", "mechanism": "A reference database of more than 2,000 attorney-validated answers supports regression tests before releases across a platform handling more than 200,000 land documents.", "measure": "The source does not give evaluated decision counts, class balance or an untouched test design. Document-universe size is not the denominator of these rates.", "operating_period": "Publication, operating and evaluation dates unknown.", "limit": "Anonymous client, undated provider case. No independently confirmed live error rate, realized labor saving, full cost or ROI. Reduced false positives does not eliminate human title review.", "ai_role": "LLM title determination plus legal rules, SQL and knowledge-graph concepts.", "asset": "US land records; anonymous technology provider, no named operator.", "partner": "Toptal", "baseline": "Approximately 27% versus 2% false-positive rates; same-cohort comparison and denominator unspecified.", "question": "How does the error rate hold on unseen deeds, and who signs off on ownership conclusions?", "sources": [{"id": "D26-TOPTAL", "title": "Mineral-title evaluation and regression testing", "url": "https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy", "date": "Unknown", "publisher": "Toptal", "locator": "Challenge; Ground-truth AI Evaluation; Outcome. No publication/evaluation date displayed.", "excerpt": "approximately 27% to 2% while maintaining roughly 93% recall", "existing": false, "format": "html", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated."}], "category": "Technical test", "related_ids": [], "record": null, "core_company": false, "logo_available": false, "relationship": "Supplementary source review; no new core deployment, adoption, outcome or ROI count.", "date": "Unknown", "source_type": "Technical evaluation", "source_id": "D26-TOPTAL", "source_title": "Mineral-title evaluation and regression testing", "url": "https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy", "publisher": "Toptal", "locator": "Challenge; Ground-truth AI Evaluation; Outcome. No publication/evaluation date displayed.", "excerpt": "approximately 27% to 2% while maintaining roughly 93% recall"}, {"id": "D26-SLB-DDR", "company": "SLB", "title": "High overall accuracy hid weaker rare-event detection", "department": "Drilling & completions", "status": "Published patent-application technical examples; no verified commercial rollout", "summary": "An SLB patent application reports 89% overall accuracy but 44% equal-class recall for its oil-and-gas GPT-2 comparator.", "mechanism": "The filing describes minority-class problems in drilling-risk classification. Customized models perform better in its examples, but the results are not a current-model ceiling.", "measure": "Table 15 reports the two different measures; 44% is not overall recall on all remarks. Exact test population and independent validation for that row are not established.", "operating_period": "Application published 5 March 2026; exact underlying evaluation dates unknown.", "limit": "Applicant-authored technical examples, not an independently validated field outcome. Other examples use differing cohorts and later model configurations. No measured reduction in drilling incidents or complete ROI.", "ai_role": "Classifies risks in daily drilling remarks using language models; engineering review and operating permissions not established.", "asset": "Drilling-report examples; no named customer or live asset tied to the 89%/44% row.", "partner": "Schlumberger Technology Corporation / published US patent application", "baseline": "Different metrics for one comparator, not a before/after improvement.", "question": "Does the evaluation give rare, costly events enough weight?", "sources": [{"id": "D26-SLB-DDR", "title": "AI approach for drilling operation insights — US 20260063025 A 1", "url": "https://patents.google.com/patent/US20260063025A1/en", "date": "2026-03-05", "publisher": "Schlumberger Technology Corporation / published US patent application", "locator": "Table 15 and section 5, Model Performance Comparative Analysis; published application, not a granted-patent or commercial-deployment claim.", "excerpt": "especially for minority class when the training data is insufficient", "existing": false, "format": "html", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated."}], "category": "Technical test", "related_ids": [], "record": null, "core_company": true, "logo_available": true, "relationship": "Supplementary source review; no new core deployment, adoption, outcome or ROI count.", "date": "2026-03-05", "source_type": "Technical evaluation", "source_id": "D26-SLB-DDR", "source_title": "AI approach for drilling operation insights — US20260063025A1", "url": "https://patents.google.com/patent/US20260063025A1/en", "publisher": "Schlumberger Technology Corporation / published US patent application", "locator": "Table 15 and section 5, Model Performance Comparative Analysis; published application, not a granted-patent or commercial-deployment claim.", "excerpt": "especially for minority class when the training data is insufficient"}, {"id": "D26-SNAM", "company": "Snam / Engineering Group", "title": "Complex plant drawings were harder than simple ones", "department": "IT, data & enterprise transformation", "status": "Operator/provider research evaluation; broad production scale not established", "summary": "The Snam/Engineering paper reports 42.4% relationship recall at the top 100 predictions; a separate final-task table reports 95% component identification and 90% section assignment.", "mechanism": "The authors say complex systems remain difficult, although simpler systems work better. This constrains claims about automatic plant digitization.", "measure": "Relationship recall and final component/section measures use different tasks and denominators. The roughly 1,500-ReMi document collection is not a disclosed count of live automated plants.", "operating_period": "Preprint 5 November 2025; evaluation and operating calendar dates unknown.", "limit": "Authors provide technical tests, not realized time savings or full-cost ROI. Do not merge abstract summary accuracy with task-specific tables or label this a cancelled rollout.", "ai_role": "OCR, vision/language models, scene-graph reasoning and constrained optimization support MGM users.", "asset": "Snam gas-plant documentation; Italy; approximately 1,500 ReMi in the provided collection.", "partner": "Snam / Engineering Group", "baseline": "Table 2 relationship recall@100=.424; Table 3 final-task measures are different quantities.", "question": "Which drawings still need manual reconstruction, and how long does correction take?", "sources": [{"id": "M29-006", "title": "Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants", "url": "https://arxiv.org/html/2511.08609v1", "date": "2025-11-05", "publisher": "Snam / Engineering Group", "locator": "Sections 3.1,4(Tables 1–3) and 5; top 100 relationship recall and complex-system limitation.", "excerpt": "", "existing": true, "format": "html", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated."}], "category": "Technical test", "related_ids": ["AI-208"], "record": null, "core_company": true, "logo_available": true, "relationship": "Supplementary source review; no new core deployment, adoption, outcome or ROI count.", "date": "2025-11-05", "source_type": "Technical evaluation", "source_id": "M29-006", "source_title": "Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants", "url": "https://arxiv.org/html/2511.08609v1", "publisher": "Snam / Engineering Group", "locator": "Sections 3.1,4(Tables 1–3) and 5; top 100 relationship recall and complex-system limitation.", "excerpt": ""}, {"id": "D26-TOTAL", "company": "TotalEnergies", "title": "One useful warning model did not make every coker use case useful", "department": "Refining, LNG & process operations", "status": "Advisory pilot; one use case reportedly not retained, other work continued", "summary": "A named TotalEnergies engineer said a 75%-accuracy drum-level model was not retained because switching also depended on conditions outside its scope.", "mechanism": "The same sponsored conference report describes 10–18 minutes of steam-warning lead time and at least two acted-on alerts. These concern another use case.", "measure": "No test denominator or definition for 75% accuracy. At least two acted-on alerts is not a total-alert count, avoided-trip count or reliability rate.", "operating_period": "Joint pilot began October 2024; drum/spalling work July 2025; report 11 June 2026. Steam work began May 2025 and closed a year later; reason for closure not stated.", "limit": "Named remarks in sponsored conference coverage; original slides and later company confirmation not located. Other uses continued. No quantified loss, avoided flaring, downtime benefit or full costs; do not call the whole pilot a failure.", "ai_role": "Advisory alerts and operator decisions; no autonomous process write authority established.", "asset": "Port Arthur refinery delayed coker, Texas; alarm guidance separately covered more than 300 tags.", "partner": "Control Global / Honeywell Users Group coverage", "baseline": "Different use cases have different populations; expected benefits are not measured achieved results.", "question": "Did the prediction answer the actual operating decision?", "sources": [{"id": "WATCH16AM-TOTAL-DRUM", "title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "url": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "date": "2026-06-11", "publisher": "Control Global / Honeywell Users Group coverage", "locator": "Steam perturbance, drum cycling, heater spalling and alarm-decision-support sections; Thomas Klein remarks.", "excerpt": "", "existing": true, "format": "html", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated."}], "category": "Use case not retained", "related_ids": ["AI-207"], "record": null, "core_company": true, "logo_available": true, "relationship": "Supplementary source review; no new core deployment, adoption, outcome or ROI count.", "date": "2026-06-11", "source_type": "Source-reviewed project account", "source_id": "WATCH16AM-TOTAL-DRUM", "source_title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "url": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "publisher": "Control Global / Honeywell Users Group coverage", "locator": "Steam perturbance, drum cycling, heater spalling and alarm-decision-support sections; Thomas Klein remarks.", "excerpt": ""}]}}}