{ "updated_at": "2026-10-26", "reports": [ { "id": "CON001", "title": "Can upstream oil and gas produce value from AI’s $230 billion pay zone?", "publisher": "McKinsey & Company", "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", "published_date": "2026-08-25", "date_note": "Exact date on publisher page.", "scope": "Global upstream operators and oilfield services/equipment (OFSE); value chain life cycle.", "report_type": "model", "ai_scope": "Broad AI: ML, physics-informed ML, deep-learning surrogates, computer vision, NLP, GenAI, and agentic AI.", "access": "full text", "sample_method": "Bottom-up sizing of 550+ mapped use cases against a global annual upstream cost base (about $500–550bn capex and $400–500bn opex) and production margin base (~$2tn at $50/boe); proof points and adoption/improvement assumptions; cascade correction for overlap.", "findings": [ "McKinsey estimates $65bn/year near-term recurring value with current technologies, rising to $125bn with broader adoption and $230bn at full potential with more autonomous operations.", "Top 10 use cases represent nearly half of identified value; top 20 about two-thirds and top 60 about 95%. OFSE internal cost opportunities add a separate $10–20bn/year estimate." ], "metrics": [ { "claim": "Upstream operator value, near-term", "value": "~65", "unit": "USD bn per year recurring", "period": "near-term/current technology and adoption", "population": "global upstream oil and gas", "basis": "bottom-up 550+ use cases; net of implementation costs", "caveat": "Modeled potential, not realized market savings; scope overlaps operator/OFSE opportunity framing." }, { "claim": "Upstream operator value, full potential", "value": "~230", "unit": "USD bn per year recurring", "period": "full potential/autonomous operating modes", "population": "global upstream oil and gas", "basis": "bottom-up value-pool model with cascade overlap adjustment; >$30bn annual implementation cost netted", "caveat": "Scenario, not forecast; assumes maximum addressability and autonomous operations." }, { "claim": "Exploration success balance-sheet value", "value": ">35", "unit": "USD bn per year", "period": "full potential", "population": "global upstream", "basis": "reserves accretion from AI-enabled exploration success", "caveat": "Separate balance-sheet value, not part of recurring $230bn operating value." } ], "limitations": "Scenario-based consulting model with many assumptions; not a measured realized-impact study. $230bn depends on high adoption/autonomy. Interpret separate OFSE and exploration value carefully to avoid adding overlapping scopes.", "departments": [ "upstream operations", "subsurface", "drilling", "production", "maintenance", "procurement", "finance", "OFSE product and service teams" ], "tags": [ "upstream", "AI economics", "agentic AI", "value pools", "oilfield services" ], "why_useful": "Most explicit current-period primary-source sizing with disclosed use-case count, cost base, scenario logic, and overlap treatment.", "source_family": "original analysis", "publisher_group": "McKinsey & Company", "kind": "Opportunity model", "display_date": "2026-08-25", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "CON002", "title": "Gen AI in the OFSE industry: Progress lags behind intent", "publisher": "McKinsey & Company", "url": "https://www.mckinsey.com/industries/oil-and-gas/our-insights/gen-ai-in-the-ofse-industry-progress-lags-behind-intent", "published_date": "2026-04-01", "date_note": "Exact publication date; survey fielded May 2025.", "scope": "Global oilfield services and equipment executives: integrated service providers, EPC firms, equipment manufacturers and services firms, onshore/offshore.", "report_type": "survey", "ai_scope": "Primarily GenAI adoption and scaling; value estimate includes generative and analytical AI.", "access": "full text", "sample_method": "McKinsey OFSE Leaders GenAI Survey, May 2025, n=164 executives; company revenues from >$50m to >$5bn. The page gives categories and range but no probability sampling or response rate.", "findings": [ "More than three-quarters expected operational efficiency from GenAI, yet fewer than one-quarter of companies had moved beyond pilots and only 1% reported significant scale.", "Respondents most often cited data fragmentation and legacy integration; fewer than 30% reported a clear strategy or value-backed roadmap. McKinsey separately models $12–20bn annual EBITDA opportunity from scaling AI in current OFSE operations." ], "metrics": [ { "claim": "Companies beyond pilot phase", "value": "<25%", "unit": "share of represented companies/respondents", "period": "2025 survey", "population": "OFSE sector; survey respondents", "basis": "self-reported survey", "caveat": "Company-level representativeness and weighting are not specified. Article body describes the same threshold as some use; wording is inconsistent." }, { "claim": "Significant GenAI scale", "value": "1%", "unit": "share of respondents", "period": "2025", "population": "OFSE executives surveyed", "basis": "self-reported survey", "caveat": "Self-reported status; not independently audited." }, { "claim": "Potential EBITDA gain", "value": "$12–20bn", "unit": "USD per year", "period": "current operations scaled", "population": "OFSE industry", "basis": "McKinsey analysis, not survey result", "caveat": "Modeled opportunity, distinct from sample survey; not realized EBITDA." } ], "limitations": "Survey is an executive sample, not stated as representative; intentions and perceptions differ from observed operations. Modeled EBITDA potential should not be treated as surveyed or realized outcome. The article uses inconsistent descriptions for its <25% figure: beyond pilots in the introduction, some use in the body. Do not treat those as interchangeable adoption stages.", "departments": [ "operations", "maintenance", "digital/IT", "EPC", "equipment engineering", "service delivery", "HR/change management" ], "tags": [ "OFSE", "GenAI", "adoption", "survey", "scale-up" ], "why_useful": "Separates adoption evidence from the adjacent modeled $12–20bn potential and provides a 2023 expectation versus 2025 self-reported status comparison.", "source_family": "original survey and analysis", "publisher_group": "McKinsey & Company", "kind": "Adoption / management survey", "display_date": "2026-04-01", "period_group": "Current window", "scope_group": "Oilfield services", "reviewed_at": "2026-10-06" }, { "id": "CON003", "title": "Four shifts redefining the oil and gas operating model of the future", "publisher": "McKinsey & Company", "url": "https://www.mckinsey.com/industries/oil-and-gas/our-insights/four-shifts-redefining-the-oil-and-gas-operating-model-of-the-future", "published_date": "2026-01-05", "date_note": "Exact date on publisher page; 11-page article.", "scope": "Oil and gas operating models across upstream, midstream, downstream and corporate functions, viewed over the coming decade.", "report_type": "framework", "ai_scope": "GenAI, agentic AI, robotics, sensors, digital twins and edge computing as operating-model changes.", "access": "full text", "sample_method": "Executive synthesis drawing on prior McKinsey articles and external reporting; no new survey or quantified model disclosed.", "findings": [ "AI is framed as one of four forces requiring operating-model redesign, with agents potentially linking seismic, well, production, maintenance, and customer decisions.", "The article argues that flatter organizations and expanded roles may follow as people and agents collaborate; it does not quantify an oil-and-gas-specific financial impact." ], "metrics": [], "limitations": "Strategic framework rather than empirical study; AI claims are illustrative and not supported by a disclosed new sample or sector-specific impact model.", "departments": [ "executive leadership", "operations", "engineering", "IT", "HR", "commercial" ], "tags": [ "operating model", "agentic AI", "workforce", "strategy" ], "why_useful": "Useful for organization-design and governance implications, with an explicit boundary between strategic synthesis and quantified evidence.", "source_family": "framework/synthesis", "publisher_group": "McKinsey & Company", "kind": "Strategy / governance", "display_date": "2026-01-05", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "CON004", "title": "Beyond the hype: New opportunities for gen AI in energy and materials", "publisher": "McKinsey & Company", "url": "https://www.mckinsey.com/industries/energy-and-materials/our-insights/beyond-the-hype-new-opportunities-for-gen-ai-in-energy-and-materials", "published_date": "2024-02-05", "date_note": "Exact date on publisher page; historical context just outside requested Sep 2024–Sep 2026 priority window.", "scope": "Agriculture, chemicals, energy (including oil and gas), materials and mining; global cross-sector opportunity estimate.", "report_type": "model", "ai_scope": "Generative AI, including back-office, cross-functional and core-operations use cases; distinguishes incremental GenAI value above conventional AI/analytics.", "access": "full text", "sample_method": "McKinsey use-case analysis across energy and materials sectors; article states aggregate value range but this page does not provide a sector-by-sector oil-and-gas denominator or full modeling detail.", "findings": [ "The article estimates $390–550bn of additional value potential for the combined agriculture, chemicals, energy and materials sectors from GenAI.", "Oil and gas is one component of the combined sector set, but the headline range cannot be assigned to oil and gas alone." ], "metrics": [ { "claim": "GenAI value potential", "value": "$390–550bn", "unit": "USD (aggregate potential; period not explicitly annual in cited article excerpt)", "period": "future potential", "population": "agriculture, chemicals, energy and materials combined", "basis": "McKinsey use-case estimate", "caveat": "Not oil-and-gas-only; do not compare directly with annual upstream $230bn scenario." } ], "limitations": "Cross-sector estimate; oil-and-gas allocation and a precise time horizon/denominator are not evident in accessible article. Published before priority window, included as historical context.", "departments": [ "operations", "engineering", "R&D", "knowledge management", "corporate functions" ], "tags": [ "GenAI", "energy and materials", "historical context", "value estimate" ], "why_useful": "Provides an early GenAI-era baseline while making clear the aggregate estimate is not an O&G-specific figure.", "source_family": "original analysis", "publisher_group": "McKinsey & Company", "kind": "Opportunity model", "display_date": "2024-02-05", "period_group": "Historical context", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "CON005", "title": "The AI-First Oil and Gas Company (article and companion executive-perspectives deck)", "publisher": "Boston Consulting Group", "url": "https://www.bcg.com/publications/2025/ai-first-future-of-oil-and-gas-companies", "published_date": "2025-08-20", "date_note": "Exact date on publisher page; accompanying slide deck is downloadable.", "scope": "Global oil and gas value chain: upstream, refining and fuel retail.", "report_type": "model", "ai_scope": "AI and AI agents; operational analytics, prediction, automation, scheduling and customer applications.", "access": "full text", "sample_method": "Companion deck labels the 2030-versus-2025 EBIT bridge “Illustrative”; sources include company reports, BCG analysis, and external industry sources. A separate page-10 workplace GenAI question reports n=188 and is attributed in the footer to the BCG Deploy, Reshape, Invent Survey 2023; this respondent count does not underpin the EBIT bridge. Case studies are presented without named company identities or full baselines.", "findings": [ "The deck models five-year AI uplift as ~30% of 2025 EBIT upstream, ~50% refining, and ~70% fuel retail; this is explicitly illustrative, and each segment uses a different denominator ($/boe, $/bbl, $/m³).", "It presents separate examples/cases: one upstream condensate program reports $25–50m annual value uplift and +2–4% production versus a 2023 baseline; an emissions case reports ~30% emissions-reduction impact and ~$20m expected value. These are case claims, not the segment model." ], "metrics": [ { "claim": "Illustrative five-year AI impact / 2025 EBIT, upstream", "value": "~30%", "unit": "percent of 2025 EBIT; bridge is +$9/boe from $28/boe baseline", "period": "2025 baseline to 2030 impact", "population": "illustrative upstream business", "basis": "BCG companion deck, segment-level illustrative EBIT bridge", "caveat": "Not a forecast or realized average; per-boe denominator." }, { "claim": "Illustrative five-year AI impact / 2025 EBIT, refining", "value": "~50%", "unit": "percent of 2025 EBIT; bridge is +$1.0/bbl from $1.8/bbl baseline", "period": "2025 baseline to 2030 impact", "population": "illustrative refining business", "basis": "BCG companion deck, segment-level illustrative EBIT bridge", "caveat": "Not a forecast or realized average; per-barrel denominator." }, { "claim": "Illustrative five-year AI impact / 2025 EBIT, fuel retail", "value": "~70%", "unit": "percent of 2025 EBIT; bridge is +$24/m³ from $35/m³ baseline", "period": "2025 baseline to 2030 impact", "population": "illustrative fuel-retail business", "basis": "BCG companion deck, segment-level illustrative EBIT bridge", "caveat": "Not a forecast or realized average; per-cubic-meter denominator." }, { "claim": "Upstream condensate optimization case value uplift", "value": "$25–50m/year and +2–4% production", "unit": "USD per year; production versus baseline", "period": "annual uplift versus 2023 baseline", "population": "one unnamed ~40 MMscm/d gas-processing plant case", "basis": "BCG case study using advanced analytics and AI algorithms", "caveat": "Single unnamed case; company, counterfactual and attribution not disclosed." } ], "limitations": "All +30% upstream, +50% refining, and +70% fuel-retail figures are illustrative 2030 impacts against 2025 EBIT, not a forecast or realized average. The bridges use different denominators ($/boe, $/bbl, and $/m³). The n=188 workplace survey is a separate 2023 survey measure, not the basis for the EBIT model. Case studies are anonymized and their attribution/baselines are incomplete.", "departments": [ "upstream", "refining", "retail", "maintenance", "operations", "IT", "workforce training" ], "tags": [ "AI-first", "EBIT", "agentic AI", "value chain" ], "why_useful": "Separates the illustrative 2030-versus-2025 segment bridges from the separate 2023 workplace survey and anonymized case claims.", "source_family": "original illustrative analysis", "supplemental_urls": [ "https://www.bcg.com/assets/2025/executive-perspectives-ai-first-companies-win-the-future-oil-and-gas-4aug.pdf" ], "publisher_group": "Boston Consulting Group", "kind": "Opportunity model", "display_date": "2025-08-20", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "CON006", "title": "AI Adoption in Energy Should Focus on Agility, Not Algorithms", "publisher": "Boston Consulting Group", "url": "https://www.bcg.com/publications/2024/ai-adoption-in-energy", "published_date": "2024-12-02", "date_note": "Publisher page displays 2 December 2024; canonical article full text reviewed.", "scope": "Energy sector including oil and gas; adoption capability and implementation approach.", "report_type": "framework", "ai_scope": "AI and GenAI capability development, organizational agility and adoption.", "access": "full text", "sample_method": "Public article cites a BCG cross-industry study saying 98% of companies were experimenting with AI and 26% had capabilities to move beyond proof-of-concept; the article does not disclose the study base or method. No O&G-only survey is presented.", "findings": [ "BCG argues that oil and gas firms should emphasize adoption agility, people, processes and culture as AI tools become widely available, progressing from everyday GenAI use to process redesign and new business models.", "The article cites a separate cross-industry BCG study: 98% of companies were at least experimenting with AI, while 26% had capabilities to move beyond proof-of-concept; it also recommends a 10/20/70 effort allocation with about 70% on people, process and culture." ], "metrics": [ { "claim": "Companies with capabilities to move beyond proof of concept", "value": "26%", "unit": "percent of companies", "period": "study date not stated in article", "population": "Cross-industry companies; source phrasing says 26% had capabilities to progress beyond proof of concept among a population where 98% were experimenting; not O&G-specific.", "basis": "BCG study cited in article", "caveat": "Sample, date, and definition are not disclosed in accessible article; not an oil-and-gas adoption rate." } ], "limitations": "Full article text is accessible, but its cited cross-industry survey does not disclose sample/method in the article and must not be treated as an O&G-specific adoption measure.", "departments": [ "operations", "digital", "IT", "change management" ], "tags": [ "energy", "AI adoption", "agility", "framework" ], "why_useful": "A direct oil-and-gas adoption framework that distinguishes ubiquitous experimentation from organizational readiness to scale.", "source_family": "framework/synthesis", "publisher_group": "Boston Consulting Group", "kind": "Strategy / governance", "display_date": "2024-12-02", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-09-28" }, { "id": "CON007", "title": "Digital transformation in oil and gas—how energy companies can fix legacy data for the best AI advantage", "publisher": "Kearney", "url": "https://www.kearney.com/industry/energy/article/digital-transformation-in-oil-and-gas-how-energy-companies-can-fix-legacy-data-for-the-best-ai-advantage", "published_date": "2025-02-18", "date_note": "Exact date appears in Kearney O&G insight index; article body was not machine-readable in accessible page extraction.", "scope": "Oil and gas digital transformation, especially legacy data foundations and AI readiness.", "report_type": "framework", "ai_scope": "AI enabled by remediation and contextualization of legacy operational data.", "access": "abstract", "sample_method": "No survey/model method or quantitative evidence could be verified from accessible article text.", "findings": [ "Kearney frames legacy data as a constraint on AI advantage in oil and gas.", "No independently verifiable quantitative claim or sample was accessible; record is retained as a primary-source lead only." ], "metrics": [], "limitations": "Page text returned without body content; no method or quantitative claims verified. Do not cite numerical results from this record.", "departments": [ "data/IT", "operations", "engineering", "digital transformation" ], "tags": [ "legacy data", "AI readiness", "digital transformation" ], "why_useful": "Direct O&G-focused Kearney reference on an important practical prerequisite, with evidence status explicitly marked as incomplete.", "source_family": "framework/synthesis", "publisher_group": "Kearney", "kind": "Strategy / governance", "display_date": "2025-02-18", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "CON008", "title": "How AI Will Reshape The Frontlines Of Upstream Oil & Gas", "publisher": "Oliver Wyman, in partnership with Quantrue", "url": "https://www.oliverwyman.com/our-expertise/insights/2025/oct/how-ai-oil-gas-redefines-competitive-advantage.html", "published_date": null, "date_note": "Publisher article page is fully readable; page links to the full report download. Exact day is not stated in accessible page; URL indicates October 2025.", "scope": "Upstream workflows: exploration, drilling, production, logistics and HSE.", "report_type": "case synthesis", "ai_scope": "Reasoning-capable models, agent orchestration and multimodal vision; safely governed operational recommendations.", "access": "full text", "sample_method": "Expert synthesis and workflow examples with Quantrue; no survey, quantified model or case denominator disclosed in the accessible excerpt.", "findings": [ "The report argues that upstream AI is moving from dashboards and isolated prediction toward orchestration across asset workflows, while retaining human approvals, operating limits and audit trails.", "It recommends measuring system-level P&L outcomes and embedding AI into end-to-end workflows; no quantitative benefits are stated in the accessible text." ], "metrics": [], "limitations": "The complete publisher article page is accessible; the linked full report PDF was not separately reviewed. No survey or quantified model is disclosed in the accessible page. Partnership and workflow examples do not establish independent performance evidence.", "departments": [ "subsurface", "drilling", "production", "logistics", "HSE", "operations leadership" ], "tags": [ "upstream", "agentic AI", "multimodal", "governance", "workflow orchestration" ], "why_useful": "A focused upstream account of agentic/multimodal operational workflows and safeguards, useful as a 2025 technology-stage marker.", "source_family": "case synthesis", "publisher_group": "Oliver Wyman, in partnership with Quantrue", "kind": "Case / literature synthesis", "display_date": "2025-10", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "CON009", "title": "The Data Imperative", "publisher": "Roland Berger", "url": "https://www.rolandberger.com/en/Media/Roland-Berger-study-Many-companies-fail-to-adopt-AI-in-their-processes-in-spite.html", "published_date": "2025-05-14", "date_note": "Exact press release date; report covers cross-industry European organizations and does not include energy in the stated respondent sectors.", "scope": "150 executives in technology/IT services, healthcare/pharma, retail/consumer packaged goods, financial services and manufacturing across five European countries.", "report_type": "survey", "ai_scope": "Generative AI adoption, workflow integration, data quality, systems integration and talent.", "access": "excerpt", "sample_method": "Roland Berger survey of 150 executives across five named sectors and five European countries; sampling frame and field dates not stated in release.", "findings": [ "Among the 150 surveyed executives across five European countries and the listed non-energy sectors, 27% said AI was fully integrated into company operations and workflows.", "Respondents identified data quality/management, integration into existing systems, and talent scarcity as barriers. Energy and oil and gas are not among the sample’s stated sectors." ], "metrics": [ { "claim": "Companies with AI fully integrated into operations/workflows", "value": "27%", "unit": "share of surveyed companies", "period": "2025 publication; field period unspecified", "population": "150 executives across five European industries/countries", "basis": "Roland Berger survey", "caveat": "Oil and gas/energy is not among stated sectors; do not generalize to O&G." } ], "limitations": "The stated survey sectors exclude energy and oil and gas, so the 27% result is not O&G evidence. Sampling frame and field dates are not stated in the release.", "departments": [ "IT/data", "operations", "executive leadership", "HR" ], "tags": [ "GenAI", "data foundations", "adoption", "survey", "cross-industry" ], "why_useful": "A recent survey useful for general implementation friction, provided its non-energy sample is not used to claim O&G adoption rates.", "source_family": "original survey", "publisher_group": "Roland Berger", "kind": "Adoption / management survey", "display_date": "2025-05-14", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "CON010", "title": "A look into the future: Can autonomous AI agents manage capital projects?", "publisher": "Roland Berger", "url": "https://www.rolandberger.com/en/Insights/Publications/A-look-into-the-future-Can-autonomous-AI-agents-manage-capital-projects.html", "published_date": "2025-09-23", "date_note": "Exact date; applies to capital projects generally, with energy among relevant asset sectors.", "scope": "Large infrastructure, energy and production-asset capital projects; project owner/developer/sponsor decisions.", "report_type": "framework", "ai_scope": "Autonomous AI agents for project data integration, monitoring, coordination and decision support.", "access": "full text", "sample_method": "Scenario/framework article; cites external project-performance literature. No original O&G project sample disclosed.", "findings": [ "The article explores agent support for cost, schedule and benefit delivery in complex capital projects, highlighting data structuring and algorithm choice as prerequisites.", "It is a conceptual discussion of AI in capital projects; it provides no original oil-and-gas AI outcome metric." ], "metrics": [], "limitations": "Conceptual application to energy projects rather than a tested O&G deployment; no original AI outcome estimate is disclosed.", "departments": [ "capital projects", "project controls", "engineering", "procurement", "construction", "digital/IT" ], "tags": [ "capital projects", "agentic AI", "energy", "project controls" ], "why_useful": "Connects emerging agents to capital-project governance while flagging the borrowed 9-in-10 figure as secondary evidence.", "source_family": "framework with repeated external claim", "publisher_group": "Roland Berger", "kind": "Strategy / governance", "display_date": "2025-09-23", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "CON011", "title": "M&A in Energy and Natural Resources: Making Deal Economics Work in a Record Year", "publisher": "Bain & Company", "url": "https://www.bain.com/insights/energy-and-natural-resources-m-and-a-report-2025/", "published_date": "2025-02-04", "date_note": "Exact date on Bain page; AI is a brief deal-process mention, not the report’s main subject.", "scope": "Energy and natural-resources transactions, including oil and gas consolidation in 2024.", "report_type": "case synthesis", "ai_scope": "Generative AI as a tool for estimating deal synergies; no detailed AI deployment analysis.", "access": "excerpt", "sample_method": "Part of Bain’s 2025 M&A report; transaction-market synthesis. No AI-specific survey, sample or valuation method disclosed in the accessible excerpt.", "findings": [ "Bain says acquirers are increasingly using GenAI to develop more robust and reliable deal-synergy estimates.", "The accessible report does not quantify GenAI’s effect on synergy-estimate accuracy, diligence speed, or deal value." ], "metrics": [], "limitations": "AI claim is ancillary and qualitative; no AI-specific evidence or measurable effect is provided. The article is not a substantive AI report.", "departments": [ "corporate development", "M&A", "strategy", "finance" ], "tags": [ "Bain", "GenAI", "M&A", "synergy estimates", "peripheral source" ], "why_useful": "Documents one narrow corporate use case and establishes that Bain’s accessible 2025 energy report is not a source for sector-wide AI economics.", "source_family": "market synthesis; AI claim qualitative", "publisher_group": "Bain & Company", "kind": "Case / literature synthesis", "display_date": "2025-02-04", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-09-28" }, { "id": "CON012", "title": "Harnessing volatility: Technology transformation in oil and gas", "publisher": "McKinsey & Company", "url": "https://www.mckinsey.com/capabilities/operations/our-insights/harnessing-volatility-technology-transformation-in-oil-and-gas", "published_date": "2022-09-01", "date_note": "Exact date. Historical context, outside the Sep 2024–Sep 2026 priority window.", "scope": "Oil and gas digital transformation across upstream and downstream value chains.", "report_type": "case synthesis", "ai_scope": "Digital technologies including analytics, AI and IIoT; most quantified value is digital/analytics, not AI-only.", "access": "full text", "sample_method": "Synthesis of McKinsey research and company examples; includes anonymized Asian national oil company and refinery cases. Denominators and case selection are not fully disclosed.", "findings": [ "McKinsey says upstream companies are capturing >$5/boe from selected digital use cases, while downstream use cases deliver >$1/boe in savings; these are broad digital transformation claims, not AI-only effects.", "One refinery/petrochemical example attributes $200–300m over 18 months to advanced analytics, including an AI system; another Asian company reports >$150m new value in three years from wider digital transformation." ], "metrics": [ { "claim": "Upstream captured value from high-impact digital use cases", "value": ">$5", "unit": "USD/boe", "period": "as reported in 2022", "population": "upstream companies using exploration/drilling, well/reservoir management and condition-based maintenance", "basis": "McKinsey research synthesis", "caveat": "Realized claim in source, but digital/analytics bundle, not AI-only; no sample or company denominator supplied." }, { "claim": "Refinery/petrochemical advanced analytics case value", "value": "$200–300m", "unit": "USD over 18 months", "period": "18 months", "population": "one unnamed refinery and petrochemical complex", "basis": "anonymized case narrative; includes real-time AI system and other advanced analytics", "caveat": "Attribution and counterfactual not disclosed; not generalizable." }, { "claim": "Companies not beyond digital pilot phase", "value": "70%", "unit": "share of companies", "period": "2022 article citing McKinsey research", "population": "oil and gas companies; sample details not provided on article page", "basis": "cited McKinsey research", "caveat": "Digital initiatives broadly, not AI adoption rate; underlying sample details not surfaced." } ], "limitations": "Historical pre-GenAI context; technology bundle is broad digital/analytics, and case claims are anonymized with limited methods. Avoid comparing directly with post-2024 GenAI/AI estimates.", "departments": [ "upstream operations", "downstream operations", "IT/OT", "engineering", "workforce/change management", "executive leadership" ], "tags": [ "historical context", "digital transformation", "AI", "analytics", "value realization" ], "why_useful": "Preserves the older McKinsey baseline while clearly separating AI from broad digital value and realized case claims.", "source_family": "case synthesis and cited research", "publisher_group": "McKinsey & Company", "kind": "Case / literature synthesis", "display_date": "2022-09-01", "period_group": "Historical context", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "ADP001", "title": "Oil and gas in the AI era", "publisher": "IBM Institute for Business Value", "url": "https://www.ibm.com/thought-leadership/institute-business-value/en-us/report/oil-and-gas-in-ai-era", "published_date": "2025-05-21", "date_note": "IBM page states original publication date May 21, 2025; survey fieldwork was March–April 2025.", "scope": "Global oil and gas; upstream, midstream, downstream executives.", "report_type": "survey/model", "ai_scope": "Traditional AI and AI-enabled business models; includes operational applications.", "access": "full text", "sample_method": "IBM IBV, with Oxford Economics, surveyed 105 O&G executives in the US, UK, Germany, and Australia in March–April 2025. Respondents were equally distributed across CEO, CIO, CTO, COO, and CFO roles. Values are respondent reports about their organizations; no common baseline or causal verification is supplied.", "findings": [ "Survey respondents reported AI use across the value chain: 44% said their organizations currently use AI in exploration, while 45% expected to do so within three years; 41% reported current AI use in refining, with 52% expecting it within three years.", "Respondents reported operational improvements including 27% better production uptime, 26% better asset utilization, and 15% higher throughput; the report does not provide common baselines or independent causal verification.", "The report describes AI use in both operational efficiency and new business models; its survey is executive perception, not a sector census." ], "metrics": [ { "claim": "Organizations currently using AI in exploration", "value": 44, "unit": "percent", "period": "survey conducted March–April 2025; current at survey", "population": "Organizations represented by 105 surveyed O&G executives", "basis": "Executive survey response", "caveat": "Not a census of upstream firms; sample is four countries and five C-suite roles." }, { "claim": "Organizations expecting AI use in exploration within three years", "value": 45, "unit": "percent", "period": "expectation within three years, survey conducted March–April 2025", "population": "Organizations represented by 105 surveyed O&G executives", "basis": "Executive survey expectation", "caveat": "Expectation, not realized adoption." }, { "claim": "Reported improvement in production uptime", "value": 27, "unit": "percent", "period": "survey conducted March–April 2025; duration/baseline not stated", "population": "O&G executives reporting organizational outcomes", "basis": "Self-reported improvement attributed to AI", "caveat": "No common baseline, denominator per question, or independent causal verification supplied." }, { "claim": "Reported improvement in asset utilization", "value": 26, "unit": "percent", "period": "survey conducted March–April 2025; duration/baseline not stated", "population": "O&G executives reporting organizational outcomes", "basis": "Self-reported improvement attributed to AI", "caveat": "No common baseline, denominator per question, or independent causal verification supplied." } ], "limitations": "Executive self-report from 105 respondents in four countries; performance figures lack common baselines and independent causal verification. Adoption percentages describe surveyed organizations, not the full industry.", "departments": [ "operations", "subsurface", "maintenance", "digital/IT", "strategy" ], "tags": [ "oil-gas-specific", "operations", "traditional-ai", "uptime", "asset-utilization" ], "why_useful": "Recent O&G-specific numeric baseline for operational AI, with caveats that distinguish reported gains from verified outcomes.", "source_family": "original", "publisher_group": "IBM Institute for Business Value", "kind": "Survey & estimates", "display_date": "2025-05-21", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "ADP002", "title": "Upstream oil and gas in the AI era", "publisher": "IBM Institute for Business Value", "url": "https://www.ibm.com/thought-leadership/institute-business-value/en-us/report/upstream-oil-and-gas-in-ai-era", "published_date": "2026-09-04", "date_note": "IBM page states original publication date September 4, 2026; survey fieldwork was April–May 2026.", "scope": "Upstream oil and gas; focused on enterprise scaling of AI in workflows.", "report_type": "survey/model", "ai_scope": "AI across upstream workflows; summary frames issue as workflow/system redesign and enterprise value scaling.", "access": "full text", "sample_method": "IBM IBV, in partnership with Oxford Economics, surveyed 100 upstream O&G executives across 12 countries in April–May 2026. Respondents included CEO/CIO/CTO/COO/CFO and EVP/SVP/VP roles in upstream, subsurface, drilling, and production. The study focused on larger companies, whose average 2025 revenue exceeded US$11 billion. Findings are executive survey responses and expectations; no probability sampling or independent outcome validation is disclosed.", "findings": [ "The report says AI is being implemented within upstream workflows but has not yet translated into coordinated system-level performance; respondents identify fragmented data, unclear decision rights, and workforce constraints as scale barriers.", "Respondents report AI spend at 9% of annual IT budgets for 2023–2025 and expect 13% in 2026 and 18% in 2029; the latter are expectations, not measured sector spending.", "Respondents report AI-associated workflow improvements, including lower drilling nonproductive time (12%), shorter well-delivery cycle time (9%), and fewer safety incidents disrupting operations (16%); these are self-reported impacts, not independently established causal effects." ], "metrics": [ { "claim": "AI spend as share of annual IT budget", "value": 9, "unit": "percent", "period": "2023–2025", "population": "100 surveyed upstream O&G executives at mostly large companies", "basis": "Survey-reported spend trajectory", "caveat": "Respondent-reported historical period; not an audited industry spending statistic." }, { "claim": "AI spend as share of annual IT budget", "value": 13, "unit": "percent", "period": "2026", "population": "100 surveyed upstream O&G executives at mostly large companies", "basis": "Survey response/current-year estimate", "caveat": "Respondent-reported estimate; not an audited industry spending statistic." }, { "claim": "AI spend as share of annual IT budget", "value": 18, "unit": "percent", "period": "2029", "population": "100 surveyed upstream O&G executives at mostly large companies", "basis": "Survey expectation", "caveat": "Forecast expectation, not actual spending." }, { "claim": "Organizations citing data fragmentation as a major AI scaling barrier", "value": 57, "unit": "percent", "period": "survey conducted April–May 2026", "population": "100 surveyed upstream O&G executives", "basis": "Self-reported survey response", "caveat": "Question-specific response among a large-company-focused sample; not an industry census." }, { "claim": "Organizations reporting real-time OT/IT integration", "value": 9, "unit": "percent", "period": "survey conducted April–May 2026", "population": "100 surveyed upstream O&G executives", "basis": "Self-reported survey response", "caveat": "Question-specific response; not an independently assessed technical audit." }, { "claim": "Organizations reporting enterprise-level data integration across upstream operations", "value": 12, "unit": "percent", "period": "survey conducted April–May 2026", "population": "100 surveyed upstream O&G executives", "basis": "Self-reported survey response", "caveat": "Question-specific response; not an independently assessed technical audit." }, { "claim": "Reported reduction in drilling nonproductive time attributed to AI", "value": 12, "unit": "percent", "period": "survey conducted April–May 2026; duration not stated", "population": "Surveyed upstream O&G executives reporting AI impacts", "basis": "Self-reported AI impact response", "caveat": "No common baseline or independent causal validation; not an established sector-wide outcome." }, { "claim": "Reported reduction in well delivery cycle time attributed to AI", "value": 9, "unit": "percent", "period": "survey conducted April–May 2026; duration not stated", "population": "Surveyed upstream O&G executives reporting AI impacts", "basis": "Self-reported AI impact response", "caveat": "No common baseline or independent causal validation; not an established sector-wide outcome." }, { "claim": "Reported reduction in safety incidents disrupting operations attributed to AI", "value": 16, "unit": "percent", "period": "survey conducted April–May 2026; duration not stated", "population": "Surveyed upstream O&G executives reporting AI impacts", "basis": "Self-reported AI impact response", "caveat": "No common baseline or independent causal validation; not an established sector-wide outcome." } ], "limitations": "The sample is only 100 executives, focused on larger firms, and is not a representative census. AI impacts and integration status are self-reported; no common outcome baselines, per-question response counts, or independent causal validation are given. Do not generalize reported impacts as proven causal effects.", "departments": [ "operations", "subsurface", "engineering", "digital/IT", "organization design" ], "tags": [ "oil-gas-specific", "upstream", "scale", "workflow-redesign", "enterprise-ai" ], "why_useful": "Provides a recent upstream-specific view of spending trajectory, workflow implementation, and operating-model scale constraints; its 100-executive, large-company sample and self-reported outcomes must stay explicit.", "source_family": "original", "supplemental_urls": [ "https://www.ibm.com/downloads/documents/us-en/16ddab5bd55473bb" ], "publisher_group": "IBM Institute for Business Value", "kind": "Survey & estimates", "display_date": "2026-09-04", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "ADP003", "title": "How to seize digital opportunity in the upstream oil and gas sector", "publisher": "EY-Parthenon", "url": "https://www.ey.com/en_uk/insights/energy-resources/seizing-the-digital-opportunity-in-upstream-oil-and-gas", "published_date": "2026-06-05", "date_note": "Page publication date June 5, 2026; survey fieldwork March–April 2026.", "scope": "Global upstream oil and gas digital investment, value drivers and AI spending.", "report_type": "survey/model", "ai_scope": "Traditional AI, GenAI and agentic AI; broader digital solutions also included.", "access": "full text", "sample_method": "In March–April 2026, EY surveyed more than 200 professionals with decision-making responsibility for buying upstream digital solutions in North America, Latin America, Europe, Sub-Saharan Africa, North Africa, the Middle East, and Asia excluding China and Russia; it also interviewed 65+ upstream O&G, oilfield-services, and digital specialists. Respondents estimated the value their companies received from all digital solutions relative to prior practice; EY combined survey responses with analysis for market projections. Fieldwork overlapped the early weeks of the 2026 Iran conflict, while projections were developed before those events.", "findings": [ "The survey estimates over US$790 billion of value from all upstream digital solutions over 2026–2035, combining revenue increases and cost decreases; the figure is not AI-only and is not observed realized value.", "Respondents estimated US$2.80 of value per US$1 spent on digital solutions, expected to be realized within five to six years; this is self-estimated value, not audited ROI.", "EY projects the upstream digital-solutions market at US$38 billion in 2035. These market and value estimates describe a broader digital ecosystem, not AI software spending." ], "metrics": [ { "claim": "Estimated value from all upstream digital solutions", "value": 790, "unit": "USD billions (over US$790bn)", "period": "2026–2035", "population": "Global upstream oil and gas", "basis": "EY extrapolation informed by survey-derived value estimates; combines increased revenue and decreased costs", "caveat": "Modeled/estimated opportunity, not observed realized value; covers digital solutions broadly, not AI alone." }, { "claim": "Average estimated value per dollar spent on digital solutions", "value": 2.8, "unit": "USD value per USD spent", "period": "Respondents expected value realized within five to six years of associated spend", "population": "Survey respondents responsible for upstream digital purchases", "basis": "Respondent self-estimate relative to prior practice", "caveat": "Not audited ROI; all digital solutions, not AI-only." }, { "claim": "Projected upstream digital-solutions market size", "value": 38, "unit": "USD billions", "period": "2035", "population": "Global upstream digital solutions", "basis": "EY market projection", "caveat": "Forecast, broader than AI market or software spend." } ], "limitations": "The value estimate is survey-derived and modeled, not realized performance; covers all upstream digital solutions rather than AI alone. The public summary does not establish representative sampling. Fieldwork coincided with the early 2026 Iran conflict, but market projections predated it.", "departments": [ "subsurface", "production", "development", "procurement/logistics", "digital/IT", "capital planning" ], "tags": [ "oil-gas-specific", "upstream", "agentic-ai", "genai", "digital-value", "forecast" ], "why_useful": "Best recent O&G-specific study for separating traditional AI from early GenAI/agentic spend and distinguishing modeled digital value from realized benefits.", "source_family": "original", "publisher_group": "EY-Parthenon", "kind": "Survey & estimates", "display_date": "2026-06-05", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "ADP004", "title": "Scaling AI in upstream energy: How front runners turn intelligence into advantage", "publisher": "Accenture", "url": "https://www.accenture.com/en/insights/energy/scaling-ai-upstream-energy", "published_date": "2026-03-23", "date_note": "Accenture perspective dated March 23, 2026; the Aramco metrics are confirmed in a November 24, 2025 company release and refer to 2024 TRV and late-2025 deployment status.", "scope": "Upstream energy; includes energy respondents and Aramco case example.", "report_type": "case synthesis", "ai_scope": "Operational AI scale, workforce trust and integration in upstream workflows.", "access": "full text", "sample_method": "The public Accenture perspective does not disclose the underlying respondent N, field dates, sampling frame, or question wording for the 7%, 21%, and 64% energy statistics. Its Aramco case cites company-reported figures: Aramco’s November 2025 release confirms $1.8 billion AI-driven Technology Realized Value for 2024, 442 identified AI use cases, 200+ solutions deployed, and 100+ in development as of late 2025. Aramco defines TRV as incremental financial benefit versus a business baseline and says its TRV metric is third-party verified; the AI-specific allocation remains company-attributed.", "findings": [ "Accenture reports that 7% of energy organizations have scaled AI enterprise-wide, 21% of energy leaders cite weak strategy integration as the biggest value limiter, and 64% of energy respondents say employees trust human-colleague insights more than AI; public methodology is undisclosed.", "Aramco reported $1.8 billion in AI-driven Technology Realized Value for 2024, 442 identified use cases, 200+ deployed solutions, and 100+ in development as of late 2025; this is a single-company disclosure, not an industry benchmark.", "Aramco defines TRV as incremental benefit through avoided capex/opex or increased revenue versus a business baseline, with third-party verification of TRV; AI-specific contribution is still company-attributed." ], "metrics": [ { "claim": "Energy organizations with AI scaled across the enterprise", "value": 7, "unit": "percent", "period": "not disclosed", "population": "Energy organizations; upstream/O&G subgroup unclear", "basis": "Accenture-reported survey statistic", "caveat": "N, field dates, sampling frame, and question wording not disclosed on page; not O&G-specific." }, { "claim": "Energy leaders citing weak linkage between AI initiatives and core business strategy as biggest value limiter", "value": 21, "unit": "percent", "period": "not disclosed", "population": "Energy leaders; upstream/O&G subgroup unclear", "basis": "Accenture-reported survey statistic", "caveat": "N, field dates, sampling frame, and question wording not disclosed on page; not O&G-specific." }, { "claim": "Energy respondents saying employees trust human-colleague insights more than AI insights", "value": 64, "unit": "percent", "period": "not disclosed", "population": "Energy respondents; upstream/O&G subgroup unclear", "basis": "Accenture-reported survey statistic", "caveat": "N, field dates, sampling frame, and question wording not disclosed on page; not O&G-specific." }, { "claim": "Aramco AI-driven Technology Realized Value", "value": 1.8, "unit": "USD billions", "period": "2024", "population": "Aramco", "basis": "Company-reported TRV attributed to AI; TRV is defined as incremental benefit versus a business baseline and company says the TRV metric is third-party verified", "caveat": "Single-company disclosure; AI-specific allocation is company-attributed, not a sector benchmark or independently verified causal AI effect." } ], "limitations": "The energy survey statistics have no public sample/method detail on the cited page and are not O&G-specific. Aramco TRV is a single-company metric; although the company defines TRV against a baseline and states the metric is third-party verified, its AI-specific allocation remains company-attributed.", "departments": [ "subsurface", "wells", "production", "operations", "HR/workforce", "digital/IT", "strategy" ], "tags": [ "upstream-energy", "scale", "workforce-trust", "operations", "case-study" ], "why_useful": "Adds a scale-maturity signal and a company-specific value-capture example; keep the undisclosed cross-energy survey and Aramco-reported TRV separate.", "source_family": "original", "supplemental_urls": [ "https://europe.aramco.com/en/news-media/news/2025/aramco-deploys-ai-and-supercomputing-to-drive-digital-transformation", "https://www.aramco.com/en/investors/why-invest-in-aramco" ], "publisher_group": "Accenture", "kind": "Case / literature synthesis", "display_date": "2026-03-23", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-09-28" }, { "id": "ADP005", "title": "2026 Oil and Gas Industry Outlook", "publisher": "Deloitte Research Center for Energy & Industrials", "url": "https://www.deloitte.com/content/dam/assets-zone4/br/pt/docs/industries/energy-resources-industrials/2025/Full%20PDF%20Report%20-%202026%20Oil%20and%20Gas%20Industry%20Outlook.pdf", "published_date": "2025-10-29", "date_note": "Deloitte page publication date October 29, 2025; report marks figures as of September 2025.", "scope": "US oil and gas sector outlook, with global context.", "report_type": "model", "ai_scope": "AI/GenAI/agentic AI spending trend and operational technology examples.", "access": "full text", "sample_method": "Not a respondent survey. Deloitte analyzes IDC’s AI and generative-AI spending guide (2024–2029), Gartner IT-spending data, and other cited sources. The 2025 O&G spending growth figure is an IDC-derived market estimate for US companies combining AI, GenAI, and agentic AI; the report states data are as of September 2025.", "findings": [ "Deloitte/IDC estimate that US O&G spending on AI, GenAI, and agentic AI rose 40% year over year in 2025; this is an expenditure estimate, not a survey or deployment rate.", "The report says the combined AI/GenAI share is currently below 20% of US O&G IT spending and projects it to exceed 50% by 2029.", "The report discusses smart operations, field applications, and workforce implications as strategic directions; it does not measure industry-wide deployed operational impact." ], "metrics": [ { "claim": "US O&G spending growth on AI, GenAI, and agentic AI", "value": 40, "unit": "percent year over year", "period": "2025; report data as of September 2025", "population": "US oil and gas companies", "basis": "Deloitte analysis of IDC AI and generative-AI spending guide", "caveat": "Market/spending estimate; not survey response, deployment measure, or realized value." }, { "claim": "AI and GenAI share of US O&G IT spending", "value": "less than 20", "unit": "percent", "period": "current in 2025 outlook", "population": "US oil and gas companies", "basis": "Deloitte analysis of IDC AI and generative-AI spending guide", "caveat": "Spending estimate; combined AI and GenAI scope; not adoption rate." }, { "claim": "Projected AI and GenAI share of US O&G IT spending", "value": "more than 50", "unit": "percent", "period": "2029", "population": "US oil and gas companies", "basis": "Deloitte analysis of IDC AI and generative-AI spending guide", "caveat": "Forecast; combined AI and GenAI scope; not deployment measure." } ], "limitations": "Spending figures are IDC-derived market estimates/projections, not a respondent survey, observed company deployments, or independently measured operating outcomes. Geography is US-only.", "departments": [ "operations", "field workforce", "digital/IT", "strategy" ], "tags": [ "oil-gas-specific", "spend", "agentic-ai", "genai", "operations" ], "why_useful": "Provides a recent, explicitly modeled O&G spending measure, useful if kept separate from survey-based adoption and realized value.", "source_family": "original", "publisher_group": "Deloitte", "kind": "Opportunity model", "display_date": "2025-10-29", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "ADP006", "title": "Digital innovation in Brazil's oil and gas sector", "publisher": "Deloitte Brazil with IBP institutional support", "url": "https://www.deloitte.com/br/pt/Industries/energy-chemicals/research/inovacao-digital-oleo-gas.html", "published_date": "2025-09-09", "date_note": "Deloitte page dates the research September 9, 2025; Portuguese-language report.", "scope": "Brazilian oil and gas companies.", "report_type": "survey", "ai_scope": "Digital innovation including AI; detailed public landing page does not enumerate AI-specific questions or results.", "access": "excerpt", "sample_method": "Deloitte, with institutional support from IBP, conducted a quantitative survey of 39 Brazilian O&G companies and qualitative interviews. The public page does not specify sampling frame, field dates, or the number of interviewees.", "findings": [ "Among 39 participating Brazilian O&G companies, 69% placed their innovation maturity at intermediate level or above, while 57% reported dedicated innovation time/resources; these measures concern innovation broadly, not AI adoption.", "All participating firms said ROI was essential in post-project innovation evaluation, although Deloitte says ROI was not decisive in initial project selection.", "The source offers regional organizational context on innovation practice and cross-department integration, not an AI adoption-rate survey." ], "metrics": [ { "claim": "Companies describing innovation maturity as intermediate or above", "value": 69, "unit": "percent", "period": "2025 study", "population": "39 Brazilian oil and gas companies", "basis": "Deloitte/IBP survey", "caveat": "Innovation maturity, not AI adoption." }, { "claim": "Companies allocating dedicated resources/time to innovation", "value": 57, "unit": "percent", "period": "2025 study", "population": "39 Brazilian oil and gas companies", "basis": "Deloitte/IBP survey", "caveat": "Innovation broadly, not AI-specific." }, { "claim": "Companies considering ROI essential in post-project evaluation", "value": 100, "unit": "percent", "period": "2025 study", "population": "39 Brazilian oil and gas companies", "basis": "Deloitte/IBP survey", "caveat": "Innovation projects generally, not AI-only." } ], "limitations": "Small Brazil-specific sample; public summary omits sampling frame, field dates, and qualitative interview count. Measures innovation overall, not AI adoption.", "departments": [ "digital/IT", "operations", "R&D/innovation", "sustainability", "strategy" ], "tags": [ "oil-gas-specific", "Brazil", "digital-transformation", "innovation" ], "why_useful": "Rare country-specific O&G survey and a useful regional counterweight to global consulting narratives; direct AI claims should be checked in the full results.", "source_family": "original", "publisher_group": "Deloitte", "kind": "Adoption / management survey", "display_date": "2025-09-09", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "ADP007", "title": "Rise of agentic AI: How trust is the key to human-AI collaboration", "publisher": "Capgemini Research Institute", "url": "https://www.capgemini.com/gb-en/insights/research-library/ai-agents/", "published_date": null, "date_note": "Report is identified as a 2025 publication, but exact publication date is not confirmed on the landing page. Companion infographic gives survey fieldwork as April 2025 and N=1,522; landing page rounds to 1,500.", "scope": "Cross-industry enterprise respondents across 14 countries; not O&G-specific.", "report_type": "survey/model", "ai_scope": "Agentic AI adoption, trust, autonomy and organization design.", "access": "full text", "sample_method": "Capgemini Research Institute reports a survey of 1,500 senior executives across 14 countries; its companion infographic specifies N=1,522 and survey fieldwork in April 2025. The landing-page headline rounds the sample count. Cross-industry, not O&G-specific.", "findings": [ "The cross-industry sample reports 2% of organizations with agentic AI deployed at scale, 12% at partial scale, 23% piloting, and 61% exploring deployment.", "The US$450 billion by 2028 figure is Capgemini’s modeled opportunity across surveyed markets, not observed realized value and not O&G-specific.", "The 38% figure is an expectation that organizations will have agents as team members by 2028, not current adoption." ], "metrics": [ { "claim": "Organizations with agentic AI implemented at scale", "value": 2, "unit": "percent", "period": "April 2025", "population": "Cross-industry organizations represented by surveyed executives", "basis": "Survey, N=1,522 per infographic", "caveat": "Not O&G-specific; self-reported adoption." }, { "claim": "Potential economic value from agents", "value": 450, "unit": "USD billions", "period": "by 2028", "population": "Surveyed countries/markets", "basis": "Capgemini Research Institute estimate", "caveat": "Forecast/model combining revenue and cost impact; not observed results." }, { "claim": "Organizations expecting agents as human-team members", "value": 38, "unit": "percent", "period": "by 2028", "population": "Cross-industry surveyed organizations", "basis": "Survey expectation", "caveat": "Forecast expectation, not current use; not O&G-specific." } ], "limitations": "Cross-industry executive survey, not O&G-specific. The economic-value figure is modeled, and the 38% is future expectation. Landing page and infographic round the respondent count differently; exact publication day is not established here.", "departments": [ "IT/data", "operations", "HR/workforce", "strategy", "finance", "risk/compliance" ], "tags": [ "agentic-ai", "cross-industry", "organization", "trust", "survey" ], "why_useful": "Strong cross-industry benchmark on agent deployment stages and organizational readiness, with clear April 2025 base and denominator.", "source_family": "original", "publisher_group": "Capgemini Research Institute", "kind": "Survey & estimates", "display_date": "2025", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "ADP008", "title": "The front-runners' guide to scaling AI", "publisher": "Accenture Research", "url": "https://www.accenture.com/en/insights/ai-data/front-runners-guide-scaling-ai", "published_date": "2025-05-06", "date_note": "Page date May 6, 2025.", "scope": "Cross-industry global companies; not O&G-specific.", "report_type": "survey/model", "ai_scope": "Enterprise AI scaling, including GenAI and agents in the wider research framing.", "access": "excerpt", "sample_method": "Accenture Research says it surveyed 2,000 companies; the public page confirms the sample count and May 6, 2025 publication date but does not expose field dates, geography mix, industry breakdown, or detailed classification thresholds. Results are cross-industry.", "findings": [ "Accenture classifies 8% of surveyed companies as front-runners scaling AI enterprise-wide and 15% as reinvention-ready; these are cross-industry capability/maturity categories, not O&G adoption rates.", "Front-runners have stronger reported revenue growth and shareholder returns than comparison groups, but these are observational group comparisons and do not establish AI as the cause.", "The reported 11% cost reduction and 13% productivity increase within 18 months are company expectations after enterprise scaling, not measured outcomes." ], "metrics": [ { "claim": "Companies classified as AI front-runners scaling at enterprise level", "value": 8, "unit": "percent", "period": "2025 study", "population": "2,000 companies surveyed across industries", "basis": "Accenture Research maturity classification", "caveat": "Cross-industry and definition-dependent; not an O&G adoption rate." }, { "claim": "Companies categorized as AI reinvention-ready", "value": 15, "unit": "percent", "period": "2025 study", "population": "2,000 companies surveyed across industries", "basis": "Accenture Research capability classification", "caveat": "Not an adoption rate; thresholds and subgroup composition not fully disclosed on page." }, { "claim": "Expected cost reduction after enterprise deployment and scaling", "value": 11, "unit": "percent", "period": "within 18 months", "population": "Companies scaling AI across enterprise", "basis": "Surveyed companies’ expectation", "caveat": "Expected average, not observed causal outcome." }, { "claim": "Expected productivity increase after enterprise deployment and scaling", "value": 13, "unit": "percent", "period": "within 18 months", "population": "Companies scaling AI across enterprise", "basis": "Surveyed companies’ expectation", "caveat": "Expected average, not observed causal outcome." }, { "claim": "Revenue growth difference for front-runners with annual revenue above $10 billion", "value": 7, "unit": "percentage points faster", "period": "period not specified on page", "population": "Front-runners vs companies still experimenting with AI, revenues above $10bn", "basis": "Accenture-reported observational group comparison", "caveat": "Not a causal estimate; period and exact subgroup denominator not stated." }, { "claim": "Shareholder return difference for front-runners", "value": 6, "unit": "percentage points higher", "period": "period not specified on page", "population": "Front-runners vs other company groups, all sizes", "basis": "Accenture-reported observational group comparison", "caveat": "Not a causal estimate; period and exact group definitions/denominators not fully stated on page." } ], "limitations": "Cross-industry consulting survey with limited public sampling detail. Front-runner outcomes are observational comparisons; expected productivity and cost reductions are forecasts by surveyed companies rather than realized impact.", "departments": [ "executive leadership", "IT/data", "HR/workforce", "strategy", "operations" ], "tags": [ "cross-industry", "enterprise-ai", "scale", "organization", "survey" ], "why_useful": "Provides a scaling-maturity lens to interpret why pilot counts can coexist with limited enterprise impact.", "source_family": "original", "publisher_group": "Accenture", "kind": "Survey & estimates", "display_date": "2025-05-06", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "ADP009", "title": "Future of industrial products: Made in America", "publisher": "PwC US", "url": "https://www.pwc.com/us/en/industries/industrial-products/library/future-of-industrial-products.html", "published_date": null, "date_note": "PwC page is a 2025 report; precise publication day not shown. Survey field period is Q3 2025.", "scope": "508 US-headquartered industrials and energy companies, including oil and gas and utilities; each >$500m revenue.", "report_type": "survey", "ai_scope": "Enterprise AI investment and intelligent systems in industrial/energy sectors.", "access": "full text", "sample_method": "PwC-commissioned Q3 2025 survey of 508 C-suite leaders at US-headquartered companies in aerospace/defense, automotive, chemicals, engineering/construction, industrial manufacturing, O&G and utilities; all >$500m revenue. O&G subgroup N not stated.", "findings": [ "Executives across the pooled sample expect higher AI investment and see intelligent systems as part of industrial competitiveness.", "The published results aggregate several industries, so figures cannot be interpreted as oil-and-gas-specific.", "The report is useful for enterprise investment intent, not evidence of deployed operational AI." ], "metrics": [ { "claim": "Executives plan to increase AI investment in next three years", "value": 81, "unit": "percent", "period": "next three years from Q3 2025", "population": "508 C-suite respondents across US industrials and energy sectors", "basis": "PwC survey", "caveat": "Pooled cross-sector result; O&G subgroup rate unavailable." } ], "limitations": "US only; pooled sectors; subgroup sizes and O&G-only response not disclosed; investment intent rather than realized deployment.", "departments": [ "executive leadership", "IT/data", "operations", "finance", "strategy" ], "tags": [ "cross-industry", "energy-and-industrials", "investment-intent", "enterprise-ai" ], "why_useful": "Recent C-suite investment signal with a disclosed respondent base and O&G inclusion, while clearly illustrating the denominator limitation.", "source_family": "original", "publisher_group": "PwC US", "kind": "Adoption / management survey", "display_date": "2025", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-09-28" }, { "id": "ADP010", "title": "2025 Global Energy, Natural Resources and Chemicals CEO Outlook", "publisher": "KPMG International", "url": "https://kpmg.com/xx/en/media/press-releases/2025/10/energy-ceos-look-to-ai-to-drive-transformation-and-sustainability.html", "published_date": "2025-10-27", "date_note": "KPMG release dated October 27, 2025.", "scope": "Global ENRC CEOs; spans energy, natural resources and chemicals, including oil and gas but not separable in release statistics.", "report_type": "survey", "ai_scope": "AI strategy, workforce training and sustainability in ENRC companies.", "access": "excerpt", "sample_method": "KPMG Global CEO Outlook sector survey; exact ENRC subgroup N, field dates, country count and sampling detail not provided in press release excerpt.", "findings": [ "KPMG reports reskilling and tailored training as responses to AI-related workforce change.", "Only a minority of organizations reportedly provide AI education across the entire organization.", "The same release notes engineering talent shortages including O&G, but metrics are sector-pooled." ], "metrics": [ { "claim": "CEOs reskilling/upskilling roles impacted by AI", "value": 40, "unit": "percent", "period": "2025 survey", "population": "ENRC CEOs", "basis": "KPMG CEO Outlook survey", "caveat": "Pooled sector result; full base and field dates not in release." }, { "claim": "Organizations offering AI education across entire organization", "value": 18, "unit": "percent", "period": "2025 survey", "population": "ENRC CEOs/organizations", "basis": "KPMG CEO Outlook survey", "caveat": "Pooled sector, self-reported; not O&G-specific." } ], "limitations": "Press release summary only; ENRC umbrella denominator obscures O&G subgroup; survey size/timing need full report validation.", "departments": [ "executive leadership", "HR/workforce", "learning and development", "operations", "IT/data" ], "tags": [ "cross-industry", "energy-resources-chemicals", "organization", "skills", "survey" ], "why_useful": "Adds organization and workforce capacity evidence to operational adoption reports, with a broad-sector caveat.", "source_family": "original", "publisher_group": "KPMG International", "kind": "Adoption / management survey", "display_date": "2025-10-27", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-09-28" }, { "id": "ADP011", "title": "Building resource resilience: How oil and gas companies thrive across commodity cycles", "publisher": "PwC US", "url": "https://www.pwc.com/us/en/industries/energy-utilities-resources/resource-resilience-commodity-cycle-oil-gas.html", "published_date": null, "date_note": "Page gives March 2026; exact day not specified.", "scope": "Oil and gas company operating model and resilience across commodity cycles.", "report_type": "case synthesis", "ai_scope": "AI as decision support, enterprise data architecture and field-to-enterprise scale.", "access": "full text", "sample_method": "Advisory perspective/case synthesis; no original survey or quantified AI sample disclosed.", "findings": [ "PwC describes field-level AI decision support as more common than enterprise-level adoption.", "It positions AI value primarily in faster, better decisions and connects scale to unified data architecture and core systems.", "The page gives strategic interpretation rather than quantified adoption evidence." ], "metrics": [], "limitations": "Not a survey; claims are qualitative and illustrative; no adoption base or measured AI effect disclosed.", "departments": [ "operations", "capital planning", "IT/data", "procurement", "HR/talent", "executive leadership" ], "tags": [ "oil-gas-specific", "enterprise-ai", "decision-support", "data-foundation", "qualitative" ], "why_useful": "A fresh O&G-specific synthesis connecting AI adoption to enterprise decision speed and integration with operating processes.", "source_family": "original", "publisher_group": "PwC US", "kind": "Case / literature synthesis", "display_date": "2026-03", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "ADP012", "title": "Energizing the oil and gas value chain with AI", "publisher": "IBM Institute for Business Value", "url": "https://www.ibm.com/thought-leadership/institute-business-value/en-us/report/ai-energy-value-chain", "published_date": null, "date_note": "Older historical source; exact publication date not confirmed in accessible abstract.", "scope": "Oil and gas value chain; historical AI value/strategy benchmark.", "report_type": "survey/model", "ai_scope": "Enterprise AI strategy and value realization.", "access": "abstract", "sample_method": "IBM IBV executive research; accessible citation does not provide N, field dates or sampling detail.", "findings": [ "IBM's listed takeaway is that organizations with a well-defined enterprise-wide AI strategy see greater AI ROI.", "Use as historical framing for enterprise strategy, not a current GenAI or agentic adoption benchmark." ], "metrics": [], "limitations": "Older than priority window and accessible abstract lacks methods and verifiable metric detail.", "departments": [ "executive leadership", "IT/data", "operations", "strategy" ], "tags": [ "historical-context", "oil-gas-specific", "enterprise-ai", "strategy" ], "why_useful": "Historical context for the enterprise strategy/value argument, but not comparable to 2025–26 GenAI surveys.", "source_family": "original", "publisher_group": "IBM Institute for Business Value", "kind": "Survey & estimates", "display_date": "Date not established", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS001", "title": "Energy and AI", "publisher": "International Energy Agency (IEA)", "url": "https://www.iea.org/reports/energy-and-ai", "published_date": "2025-04-10", "date_note": "Report page gives date.", "scope": "Global AI-energy nexus, including a chapter on AI use in energy supply and oil and gas operations.", "report_type": "model", "ai_scope": "AI for energy operations, with separate treatment of energy demand from AI.", "access": "full text", "sample_method": "Global and regional modeling, new datasets, and consultation with governments, regulators, technology firms, energy companies, and experts; individual O&G cases are illustrative rather than a representative impact sample.", "findings": [ "The IEA identifies upstream subsurface processing, reservoir simulation, remote operations, predictive maintenance, compliance, leak detection, and repair automation as O&G applications.", "It cautions that broad sector-level gains in costs, reliability, and resilience are difficult to quantify beyond individual case studies.", "The report treats AI-driven data-center electricity demand separately from AI as an operational tool for energy producers." ], "metrics": [], "limitations": "Broad energy-sector synthesis; reported operating benefits are not a harmonized O&G impact estimate and are difficult to generalize from case studies.", "departments": [ "Subsurface", "Drilling", "Production", "Maintenance", "HSE", "Digital/IT" ], "tags": [ "IEA", "global", "operations", "subsurface", "maintenance", "demand-context-separated" ], "why_useful": "A credible boundary-setting source that explicitly separates AI's operational uses from data-center energy demand and acknowledges weak sector-wide impact evidence.", "source_family": "originalanalysis", "publisher_group": "International Energy Agency (IEA)", "kind": "Opportunity model", "display_date": "2025-04-10", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-09-28" }, { "id": "INS002", "title": "Oil and gas activity rises modestly as production, employment little changed (Second Quarter 2024 special questions)", "publisher": "Federal Reserve Bank of Dallas", "url": "https://www.dallasfed.org/research/surveys/des/2024/2402", "published_date": "2024-06-26", "date_note": "Published date shown on page; AI special-question responses collected June 12–20, 2024.", "scope": "AI adoption, planned use, applications, and expected/experienced benefits among oil and gas firms in the Dallas Fed district.", "report_type": "survey", "ai_scope": "Traditional AI and generative AI used or planned in O&G firms.", "access": "full text", "sample_method": "Special questions received responses from 136 O&G firms; the AI usage question had 123 responses (79 E&P, 44 services); application and benefit questions were asked of subsets who use or plan AI. Executive self-report; regional rather than global sample.", "findings": [ "26% of respondents said their firm currently uses traditional AI, generative AI, or both; 24% planned use within 12 months, while 50% reported no near-term plan.", "Among current/planned users, business or predictive analytics led (64%), followed by process automation (44%), geology/reservoir engineering (41%), and predictive maintenance (41%).", "Benefits were self-reported or expected: productivity (62%), better/timelier information (53%), and lower costs (47%)." ], "metrics": [ { "claim": "Currently using any AI", "value": 26, "unit": "percent", "period": "June 12–20 2024", "population": "123 responding O&G firms", "basis": "survey respondent shares", "caveat": "Self-report; regional and unequal-size firm mix; includes traditional and generative AI." }, { "claim": "Plans AI adoption within 12 months", "value": 24, "unit": "percent", "period": "June 12–20 2024", "population": "123 responding O&G firms", "basis": "survey respondent shares", "caveat": "Intent, not realized deployment." }, { "claim": "Business/predictive analytics among current/planned users", "value": 64, "unit": "percent", "period": "June 12–20 2024", "population": "59 respondents using or planning AI", "basis": "multiple-choice application responses", "caveat": "Subset denominator; multiple selections allowed." } ], "limitations": "Survey perceptions, not independently verified adoption or realized productivity effects; district-based frame; tiny large-E&P subgroup (16 firms).", "departments": [ "Subsurface", "Drilling", "Production", "Maintenance", "Supply chain", "Finance" ], "tags": [ "Dallas Fed", "survey", "adoption", "United States", "E&P", "services" ], "why_useful": "One of the clearest public denominator-based snapshots of actual O&G AI uptake, planned adoption, and workflow priorities.", "source_family": "originalanalysis", "publisher_group": "Federal Reserve Bank of Dallas", "kind": "Adoption / management survey", "display_date": "2024-06-26", "period_group": "Historical context", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS003", "title": "AI spells opportunity and manageable risk for the oil and gas industry", "publisher": "DNV", "url": "https://www.dnv.com/article/ai-spells-opportunity-and-manageable-risk-for-the-oil-and-gas-industry/", "published_date": "2024-10-01", "date_note": "DNV page metadata shows 1 October 2024; it summarizes Transforming Through Uncertainty: Energy Industry Insights 2024, cited as April 2024. Survey timing described as early 2024.", "scope": "AI adoption stage, applications, opportunities, and risks in oil and gas.", "report_type": "survey", "ai_scope": "O&G operations and strategy; includes AI/ML operational applications and assurance concerns.", "access": "excerpt", "sample_method": "DNV article summarizes its 2024 Energy Industry Insights survey. Public article does not provide the O&G subgroup n, full questionnaire, or weighting. Crucially, its 47% table value is net agreement (% agreeing minus % disagreeing), not the gross share of respondents.", "findings": [ "DNV says 12% of O&G respondents report live day-to-day AI use and 3% highly integrated/advanced use; the page also phrases their combined low-base figure as 15%.", "Separately, 47% of respondents is the net agreement (share agreeing minus share disagreeing) that their organization will use AI-driven applications in operations in the year ahead; this is not a gross adoption/planning share.", "DNV reports >99% correct identification in simulated mooring-line test cases only; this is not field validation." ], "metrics": [ { "claim": "Live day-to-day AI use (DNV text gives 12%; same sentence also states 15% as combined low-base figure)", "value": 12, "unit": "percent", "period": "early 2024 survey", "population": "DNV O&G respondents; subgroup n not given on public article", "basis": "reported survey share", "caveat": "DNV article wording is internally ambiguous: it says 15% using AI in live day-to-day operations (12%), then separately 3% highly integrated/advanced. Preserve 12% and 3% as the split; do not describe 15% as day-to-day alone. Self-report; O&G subgroup N and weights undisclosed." }, { "claim": "Highly integrated and/or advanced AI use (reported as separate 3% category)", "value": 3, "unit": "percent", "period": "early 2024 survey", "population": "DNV O&G respondents; subgroup n not given on public article", "basis": "reported survey share", "caveat": "The article presents this separately after the 12% live-use figure; O&G subgroup N, category definitions and weighting are not exposed." }, { "claim": "AI use at planning or piloting stage", "value": 47, "unit": "percent", "period": "early 2024 survey", "population": "DNV O&G respondents; subgroup n not given on public article", "basis": "reported survey share", "caveat": "Separate current-stage measure; the same rounded 47% also appears as a forward-use net-agreement statistic. Do not conflate the measures. Self-report; denominator and weighting not public." }, { "claim": "Net agreement: expected operational AI use in the coming year", "value": 47, "unit": "percentage points (share agreeing minus share disagreeing)", "period": "early 2024 survey", "population": "DNV O&G respondents; subgroup n not given on public article", "basis": "percent agree minus percent disagree", "caveat": "Net agreement, not gross share agreeing and not adoption/planning-stage share. O&G subgroup N and weighting not exposed." }, { "claim": "Mooring-line condition correctly identified in simulated test cases", "value": ">99", "unit": "percent", "period": "DNV simulation; date not stated in article", "population": "simulated offshore installation mooring-line cases", "basis": "DNV-stated simulation performance", "caveat": "DNV says simulated test cases only. Test-set size, class balance, split, baseline and external validation are not disclosed; no field or safety-outcome inference." } ], "limitations": "Public article summary omits survey sample size, questionnaire, subgroup denominator, and weighting. Net agreement must not be read as a gross share. Technical performance statement is simulation-only.", "departments": [ "Subsurface", "Production", "Maintenance", "Pipeline integrity", "HSE", "Cybersecurity", "Digital/IT" ], "tags": [ "DNV", "survey", "O&G-specific", "adoption-stages", "net-agreement", "assurance", "safety" ], "why_useful": "Separates low live-use prevalence from pilots and future intent, and provides an important denominator lesson for comparing survey headlines.", "source_family": "originalanalysis", "publisher_group": "DNV", "kind": "Adoption / management survey", "display_date": "2024-10-01", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS004", "title": "State of the art for responsible use of artificial intelligence in the petroleum sector", "publisher": "Norwegian Ocean Industry Authority (Havtil), prepared by DNV", "url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/reports-from-projects/2024/responsible-use-of-artificial-intelligence-in-the-petroleum-sector/", "published_date": "2024-11-04", "date_note": "Havtil report page gives publication date; attached PDF is the commissioned knowledge overview.", "scope": "AI risks and opportunities in petroleum operations, with emphasis on major accident risk and safety barriers.", "report_type": "framework", "ai_scope": "Safety-critical AI alongside traditional IT/automation in petroleum settings.", "access": "full text", "sample_method": "Regulator-commissioned state-of-the-art knowledge overview; synthesis of technical literature and sector context, not a controlled field trial or prevalence survey.", "findings": [ "The report's purpose is to map fundamental risk factors in AI development and use in petroleum, especially major-accident risk.", "It considers potential efficiency and safety improvements alongside risks that differ from conventional IT and automation.", "It is useful for governance and hazard analysis, but does not establish measured frequency or magnitude of AI-related safety benefits." ], "metrics": [], "limitations": "Evidence synthesis, not an empirical estimate of incident reduction or AI deployment prevalence; findings depend on maturity and context of underlying literature.", "departments": [ "HSE", "Operations", "Automation", "Cybersecurity", "Risk management" ], "tags": [ "Havtil", "DNV", "safety", "major-accident-risk", "framework", "Norway" ], "why_useful": "A rare petroleum-specific regulator-commissioned source for safety assurance, barrier management, and responsible deployment questions.", "source_family": "originalanalysis", "publisher_group": "Norwegian Ocean Industry Authority (Havtil), prepared by DNV", "kind": "Strategy / governance", "display_date": "2024-11-04", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS005", "title": "Artificial intelligence in geoenergy: bridging petroleum engineering and future-oriented applications", "publisher": "Journal of Petroleum Exploration and Production Technology (Springer Nature)", "url": "https://link.springer.com/article/10.1007/s13202-025-01939-3", "published_date": "2025-02-10", "date_note": "Published online 10 February 2025.", "scope": "Systematic mapping of AI research in conventional petroleum (reservoir, production, drilling) and emerging geoenergy fields.", "report_type": "case synthesis", "ai_scope": "ML/DL application research, including proxy modeling, dimension reduction, data generation, and optimization.", "access": "full text", "sample_method": "Review of 254 papers; analyzed publication trends, input-output relationships, AI methods, and reported training sample sizes. Search/screening criteria should be consulted in full text for reproducibility details.", "findings": [ "This 254-paper review maps conventional petroleum applications alongside future-oriented geoenergy topics such as geological carbon storage, gas hydrates and underground hydrogen.", "Proxy modeling is prominent; the review highlights heterogeneous datasets, validation practices and a research-heavy evidence base.", "Its training-sample counts describe papers, not deployed assets or operational adoption, and are omitted from the adoption metrics." ], "metrics": [], "limitations": "Literature review, not a survey or meta-analysis of field performance. Reported study sample sizes and validation designs vary; geoenergy scope extends beyond oil and gas.", "departments": [ "Subsurface", "Reservoir engineering", "Production", "Drilling", "CCUS" ], "tags": [ "peer-reviewed", "systematic-review", "petroleum", "ML", "DL", "training-data" ], "why_useful": "Provides a counted literature base and concrete evidence about data scale, while making clear that research volume is not proof of field value.", "source_family": "originalanalysis", "publisher_group": "Journal of Petroleum Exploration and Production Technology (Springer Nature)", "kind": "Case / literature synthesis", "display_date": "2025-02-10", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS006", "title": "Artificial Intelligence in the Oil and Gas Industry: Applications, Challenges, and Future Directions", "publisher": "Applied Sciences (MDPI)", "url": "https://www.mdpi.com/2076-3417/15/14/7918", "published_date": "2025-07-16", "date_note": "Article page gives 16 July 2025.", "scope": "Review and bibliometric mapping of AI applications, constraints, and research in the O&G sector.", "report_type": "case synthesis", "ai_scope": "AI/ML applications across O&G value chain and literature trends.", "access": "full text", "sample_method": "Literature review plus VOSviewer bibliometric analysis; reports selected article corpus and publication trends, with SPE conference papers prominent. Not a survey of operating companies.", "findings": [ "This review maps O&G AI applications and research themes across the literature, including exploration, production, maintenance and operational analytics.", "It identifies data quality, interoperability, skills and explainability as recurring implementation challenges.", "Publication counts and conference-proceeding shares reflect the authors’ selected corpus, not company adoption or operational impact, and are omitted from adoption metrics." ], "metrics": [], "limitations": "Bibliometric/review synthesis; counts depend on databases, search terms and inclusion criteria. It does not measure deployment or independently validate field impact.", "departments": [ "Subsurface", "Drilling", "Production", "Maintenance", "Digital/IT" ], "tags": [ "peer-reviewed", "bibliometric", "review", "SPE", "implementation-barriers" ], "why_useful": "Maps where the technical conversation is concentrated and helps distinguish publication momentum from operational evidence.", "source_family": "originalanalysis", "publisher_group": "Applied Sciences (MDPI)", "kind": "Case / literature synthesis", "display_date": "2025-07-16", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS007", "title": "Artificial intelligence in shale gas and oil: A comprehensive review of applications and challenges", "publisher": "Green and Smart Mining Engineering / ScienceDirect", "url": "https://www.sciencedirect.com/science/article/pii/S2950555025000412", "published_date": null, "date_note": "Published in Volume 2, Issue 3, September 2025; exact day not stated in accessible article record.", "scope": "AI in shale oil and gas resource management, including characterization, fracture prediction, production forecasting, and hydraulic-fracturing optimization.", "report_type": "case synthesis", "ai_scope": "ML/DL and hybrid/physics-informed methods for shale subsurface and operations.", "access": "full text", "sample_method": "Narrative comprehensive review of published studies; not a pooled meta-analysis or deployment survey.", "findings": [ "The review finds applications in reservoir characterization, fracture prediction, production forecasting, and hydraulic-fracturing optimization.", "Hybrid approaches combining AI with physical simulation are a recurring direction, especially where data are heterogeneous.", "Data sparsity, interpretability, real-time updating, and uncertainty quantification remain barriers to scalable use." ], "metrics": [], "limitations": "Narrative synthesis with no pooled effect estimates; application accuracy in published studies may not transfer across basins or operating conditions.", "departments": [ "Subsurface", "Completions", "Production", "Reservoir engineering" ], "tags": [ "peer-reviewed", "review", "shale", "fracturing", "physics-informed", "uncertainty" ], "why_useful": "A focused synthesis for unconventional assets and a good counterweight to generic claims about AI across the full value chain.", "source_family": "originalanalysis", "publisher_group": "Green and Smart Mining Engineering / ScienceDirect", "kind": "Case / literature synthesis", "display_date": "2025-09", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "INS008", "title": "Artificial intelligence and its performance impacts in the oil and gas industry: Challenges, insights, and evaluation approaches", "publisher": "Results in Engineering / ScienceDirect", "url": "https://www.sciencedirect.com/science/article/pii/S2590123025036898", "published_date": null, "date_note": "Volume 28, December 2025; exact day not stated in accessible record.", "scope": "Conceptual review of how O&G companies can measure AI realization and performance impacts.", "report_type": "framework", "ai_scope": "AI value realization across exploration, production, and trading; adoption vs sustained embedded use.", "access": "abstract", "sample_method": "Thematic review and conceptual model; accessible abstract does not provide a systematic search protocol or pooled quantitative evidence.", "findings": [ "The authors distinguish initial AI adoption from AI realization: sustained embedding, scaling, and value capture through data, skills, and routines.", "They argue that value metrics must account for value-chain stage and ownership model, which can affect measurement validity and outcomes.", "They recommend linking operational, financial, and sustainability data and adopting common reporting rules; this is proposed guidance, not a validated standard." ], "metrics": [], "limitations": "Accessible evidence is abstract-level; conceptual recommendations are not validated against a common multi-company outcomes dataset.", "departments": [ "Executive", "Digital/IT", "Operations", "Finance", "Sustainability" ], "tags": [ "peer-reviewed", "AI-value", "measurement", "realization", "governance" ], "why_useful": "Directly addresses the central evidence gap: companies announce pilots, while consistent proof of sustained enterprise value is uncommon.", "source_family": "originalanalysis", "publisher_group": "Results in Engineering / ScienceDirect", "kind": "Strategy / governance", "display_date": "2025-12", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "INS009", "title": "Gas Lift Optimization in the Permian Uses Machine Learning, Artificial Intelligence", "publisher": "Journal of Petroleum Technology (SPE), reporting SPE-219553", "url": "https://jpt.spe.org/gas-lift-optimization-in-the-permian-uses-machine-learning-artificial-intelligence-restricted", "published_date": "2025-10-01", "date_note": "JPT page dated 1 October 2025; paper abstracted as SPE 219553.", "scope": "Closed-loop, well-by-well gas-lift optimization in a Permian operating portfolio.", "report_type": "technical study", "ai_scope": "ML/AI-assisted iterative rate testing with remote gas-lift injection setpoint control and automated data acquisition.", "access": "excerpt", "sample_method": "Operator workflow case reported by JPT; deployed across more than 1,300 wells. Full SPE paper is restricted; outcome details cannot be validated from the accessible summary.", "findings": [ "The workflow combines remote setpoint changes, multirate tests, and automated well data collection in a repeated well-level optimization loop.", "The described deployment scale exceeds 1,300 operator wells, making this a notable operational deployment rather than a lab-only model.", "The accessible summary does not publish a quantified production uplift, baseline, counterfactual, or duration, so efficacy cannot be inferred from scale alone." ], "metrics": [ { "claim": "Wells in reported deployment", "value": ">1300", "unit": "wells", "period": "reported by October 2025", "population": "Operator Permian Basin wells", "basis": "deployment count stated in JPT summary", "caveat": "No publicly accessible measured uplift, control group, or realized value in summary." } ], "limitations": "Restricted full paper; public summary omits outcome magnitude, uncertainty, selection effects, and independent verification.", "departments": [ "Production", "Artificial lift", "Automation", "Data/analytics" ], "tags": [ "SPE", "JPT", "case", "Permian", "gas-lift", "closed-loop", "deployment-scale" ], "why_useful": "A rare public description of AI-linked workflow actually deployed at asset scale, while illustrating how even deployment evidence can lack public outcome data.", "source_family": "originalanalysis", "publisher_group": "Journal of Petroleum Technology (SPE), reporting SPE-219553", "kind": "Technical case", "display_date": "2025-10-01", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-09-28" }, { "id": "INS010", "title": "Oil companies race for AI edge in upstream projects", "publisher": "S&P Global Commodity Insights", "url": "https://www.spglobal.com/energy/en/news-research/latest-news/crude-oil/071025-oil-companies-race-for-ai-edge-in-upstream-projects", "published_date": "2025-07-10", "date_note": "Article dated 10 July 2025; operator statements were made at the OPEC International Seminar on 9–10 July.", "scope": "Reported operator claims about AI in exploration, drilling, production, and maintenance, alongside separate discussion of data-center energy demand.", "report_type": "case synthesis", "ai_scope": "Reported upstream applications and claimed operator outcomes; AI demand context is a distinct article thread.", "access": "full text", "sample_method": "Journalistic synthesis of public remarks by executives at an industry conference; not an independent audit or representative company sample.", "findings": [ "S&P reports Aramco CEO Amin Nasser said the company estimated $4bn in 2024 Technology Realized Value (TRV) across deployed technologies; the article associates this with around 500 AI use cases, but the $4bn is not an AI-only figure.", "Aramco separately states $1.8bn of 2024 AI-driven TRV and 442 identified AI use cases; this is a company-reported AI-specific value claim, not an independent audit.", "S&P attributes BP’s 98% kick-detection and 90%-faster well-placement figures to CEO Murray Auchincloss; methods and validation denominators are not provided." ], "metrics": [ { "claim": "Technology Realized Value (TRV), all technology; S&P links the claim to ~500 AI use cases", "value": 4, "unit": "USD billion", "period": "2024", "population": "Aramco operations", "basis": "Aramco CEO statement reported by S&P Global; Aramco sources describe $4bn as company-wide technology TRV, over half associated with digital and AI solutions", "caveat": "Not AI-only savings. S&P article wording may imply the full $4bn came from AI; Aramco separately reports $1.8bn AI-driven TRV for 2024. Company-reported value; no independent audit or public attribution method." }, { "claim": "AI-driven Technology Realized Value (TRV)", "value": 1.8, "unit": "USD billion", "period": "2024", "population": "Aramco AI use cases across operations", "basis": "Aramco company-reported AI-driven TRV; official disclosure: https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/building-the-ai-future", "caveat": "Company-reported realized-value metric; no independent audit or public case-level attribution method in cited source. Keep separate from $4bn total technology TRV." }, { "claim": "AI kick detection accuracy", "value": 98, "unit": "percent", "period": "reported July 2025", "population": "BP production-well kick detection", "basis": "CEO statement reported by S&P Global", "caveat": "No dataset size, class balance, precision/recall definition, or external validation given." }, { "claim": "Faster new-well placement in Azerbaijan hot spots", "value": 90, "unit": "percent faster", "period": "reported July 2025", "population": "BP well placement workflow", "basis": "CEO statement reported by S&P Global", "caveat": "Company claim; baseline and number of wells not stated." } ], "limitations": "Executive claims reported at a conference and company TRV disclosures; not independent measurement. Aramco $4bn refers to broader all-technology TRV, while Aramco separately reports $1.8bn AI-driven TRV for 2024. BP accuracy/speed claims have no disclosed denominators or validation method. Keep separate from the article’s AI-related power-demand narrative.", "departments": [ "Subsurface", "Drilling", "Production", "Maintenance", "Operations" ], "tags": [ "S&P Global", "operator-claims", "upstream", "seismic", "well-placement", "demand-context-separated" ], "why_useful": "Captures high-profile claimed outcomes and helps show exactly why reported benefit metrics require transparent denominators and independent validation.", "source_family": "commoncitedsource", "publisher_group": "S&P Global Commodity Insights", "kind": "Case / literature synthesis", "display_date": "2025-07-10", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-09-28" }, { "id": "GAPM001", "title": "The $500 billion case for digital and AI in upstream oil and gas", "publisher": "Rystad Energy Advisory", "url": "https://www.rystadenergy.com/insights/digital-and-ai-in-upstream-whitepaper", "published_date": "2026-06-03", "date_note": "Publication date shown on the publisher's whitepaper page.", "scope": "Global E&P operators; upstream surface and subsurface workflows, with operator cohorts by digital/AI adoption.", "report_type": "model", "ai_scope": "Digitalization and AI together; the page distinguishes traditional ML from emerging agentic AI.", "access": "Publisher whitepaper summary and full companion article reviewed. Full whitepaper download requires a form; not obtained.", "sample_method": "Publisher says analysis uses global asset-level data, operator case studies and expert interviews; the public page does not disclose the full model, assumptions, or detailed interview/sample counts.", "findings": [ "Rystad models nearly $500bn in cumulative value for E&Ps investing in digital and AI from 2026–30, from cost reduction, production/uplift and shorter development timelines; cost and production pools contribute roughly equally through 2030.", "It frames AI primarily as a way for followers and laggards to approach frontier-operator performance, with surface operations/maintenance nearer-term and subsurface workflows the largest untapped pool. It says scaling and integration, not tool availability, are the central barriers." ], "metrics": [ { "claim": "Cumulative E&P value opportunity", "value": "~500", "unit": "USD bn", "period": "2026–2030 cumulative", "population": "E&Ps currently investing in digital and AI", "basis": "Rystad estimate across upstream workflows using global asset-level data, cases and interviews", "caveat": "Modeled digital-plus-AI opportunity; not an AI-only or realized savings figure. Public methods and assumption detail are limited." }, { "claim": "Annual tools and services market", "value": ">35", "unit": "USD bn per year", "period": "by 2030; approaching 50 by 2035", "population": "Providers of upstream digital and AI tools/services", "basis": "Rystad market estimate; page says E&Ps spent about $25bn on purchases in the prior year", "caveat": "Market spend is not operator value creation; estimate definitions and boundaries are not fully disclosed." } ], "limitations": "Broad digital-plus-AI scope and opaque public model detail make direct comparison with McKinsey's AI-only use-case model or EY's digital-solution estimate unsafe. Scenarios and potential are not observed industry-wide returns.", "departments": [ "upstream operations", "maintenance", "asset development", "subsurface", "drilling", "wells", "production", "digital and IT" ], "tags": [ "upstream", "digitalization", "AI economics", "market sizing", "adoption cohorts" ], "why_useful": "Adds an explicit supplier-market estimate, adopter-cohort view, and a catch-up rather than frontier-uplift framing absent from the existing quantified studies.", "source_family": "original analysis", "gap_filled": "Upstream operator cohorts and supplier-market estimates, using a different model from McKinsey.", "reviewed_at": "2026-09-29", "scope_group": "Upstream", "supplemental_urls": [ "https://www.rystadenergy.com/insights/ai-in-upstream-oil-and-gas" ], "company_match_ids": [], "publisher_group": "Rystad Energy Advisory", "kind": "Opportunity model", "display_date": "2026-06-03", "period_group": "Current window" }, { "id": "GAPM002", "title": "Every last drop: AI-powered analysis reveals oil industry's trillion-barrel opportunity in existing fields", "publisher": "Wood Mackenzie", "url": "https://www.woodmac.com/press-releases/every-last-drop-ai-powered-analysis-reveals-oil-industrys-trillion-barrel-opportunity-in-existing-fields/", "published_date": "2025-09-18", "date_note": "Exact date shown on publisher release.", "scope": "More than 30,000 oil fields worldwide; scenario analysis for more than 2,500 conventional fields, with recovery upside segmented by operator type and geography.", "report_type": "model", "ai_scope": "Machine-learning analogues tool compares fields across 60+ parameters; AI supports identifying recovery analogues, not producing the physical recovery uplift itself.", "access": "Full publisher release reviewed; proprietary model and underlying field data not available.", "sample_method": "Wood Mackenzie proprietary Synoptic data and Analogues feature; similarity analysis uses reservoir geology, hydrocarbon quality, in-place resources, operator access to capital/technology, costs and fiscal terms. Public release does not provide validation statistics or complete scenario tables.", "findings": [ "The analysis identifies 470bn to more than 1tn barrels of additional potential from proven recovery practices in existing fields; top-quartile recovery-factor performance moves the industry average from 29% to 34% and yields the 470bn-barrel estimate.", "Nearly 70% of the modeled upside is in NOC/state-controlled fields; upside is concentrated onshore and in shallow offshore assets. The study's AI-enabled contribution is improved analogue screening, while the recovery barrels depend on field investment and established techniques." ], "metrics": [ { "claim": "Additional reserves at industry top-quartile recovery factor", "value": "470", "unit": "billion barrels", "period": "scenario; potential in existing producing fields", "population": "Global oil fields", "basis": "Average recovery factor rises from 29% to 34%", "caveat": "Modeled reserve potential, not observed AI-attributable production or an AI market/value estimate." }, { "claim": "NOC/state-controlled share of best-in-class recovery upside", "value": "~70", "unit": "percent of global opportunity", "period": "scenario", "population": "Fields operated by NOCs and state-controlled enterprises", "basis": "Wood Mackenzie analogue scenarios; >320bn bbl at top-quartile and 700bn bbl best-in-class", "caveat": "Scenario upside is conditional on investment, technology access, fiscal terms and successful application of recovery practices." } ], "limitations": "Publisher release gives high-level method but not full technical validation, uncertainty distributions or the proprietary underlying field data. Do not present identified barrels as realized gains caused by AI.", "departments": [ "reservoir engineering", "subsurface", "field development", "production", "strategy", "partnerships and technology access" ], "tags": [ "upstream", "recovery factor", "AI-enabled analytics", "NOCs", "market opportunity" ], "why_useful": "Introduces a concrete physical-resource opportunity and NOC partnership implication, while clearly separating an AI-enabled analytical tool from AI-caused production gains.", "source_family": "original analysis", "gap_filled": "AI-enabled reservoir analogue analysis tied to asset-level recovery scenarios; distinct from operating-cost/productivity value pools.", "reviewed_at": "2026-09-29", "scope_group": "Upstream", "supplemental_urls": [ "https://www.woodmac.com/siteassets/horizons/2025/september-2025/horizons_sept-2025-whitepaper_1009.pdf" ], "company_match_ids": [], "publisher_group": "Wood Mackenzie", "kind": "Opportunity model", "display_date": "2025-09-18", "period_group": "Current window" }, { "id": "GAPM003", "title": "AI Breaks Out: From Pilot Projects to Energy’s Digital Engine", "publisher": "L.E.K. Consulting", "url": "https://www.lek.com/insights/energy-environment/ai-breaks-out-pilot-projects-energys-digital-engine", "published_date": "2026-02-27", "date_note": "Exact date shown in L.E.K. article index.", "scope": "Global energy organizations including oil and gas, utilities and renewables; O&G-specific responses are discussed but no O&G subgroup denominator is disclosed.", "report_type": "survey", "ai_scope": "AI use cases and maturity, including analytics, operational support and automation; organized through L.E.K.'s performance/competitiveness/differentiation lens.", "access": "Full publisher survey article reviewed; underlying questionnaire and microdata not obtained.", "sample_method": "L.E.K. 2025 Global Energy Study: more than 300 senior energy executives and 25+ in-depth industry-leader interviews; companies above $50m revenue and respondents with at least five years' industry experience. Seven global regions; subgroup sizes and response rates not stated in article.", "findings": [ "L.E.K. sees energy AI deployment concentrated in performance improvement, with less use for competitive advantage and new business differentiation. For oil and gas, analytics and planning uses are more prevalent than field automation.", "More than half of O&G respondents rely on AI for market-sentiment modeling and performance analytics. Respondents see implementation difficulties and data/IT modernization as the bottleneck to broader realization." ], "metrics": [ { "claim": "O&G respondents who believe AI's full potential is realized today", "value": "44", "unit": "percent", "period": "2025 survey; current view", "population": "Oil and gas respondents; n not stated", "basis": "Self-reported perception in L.E.K. Global Energy Study", "caveat": "Perception, not measured business impact; O&G subgroup denominator is not reported." }, { "claim": "O&G respondents expecting AI's full potential to be realized in 10 years", "value": "78", "unit": "percent", "period": "10-year expectation from 2025 survey", "population": "Oil and gas respondents; n not stated", "basis": "Self-reported expectation", "caveat": "Forecast belief, not an independently validated outcome." } ], "limitations": "Mixed energy-sector survey and interviews; O&G subgroup size, recruitment method, response rate and detailed questionnaire are not disclosed in the public article. Reported adoption is self-reported.", "departments": [ "trading", "planning and scheduling", "inventory", "supplier risk", "drilling", "operations", "IT and data" ], "tags": [ "survey", "adoption", "energy sector", "AI maturity", "operating workflows" ], "why_useful": "Adds a qualitative and survey-based adoption taxonomy that separates efficiency, competitive advantage and business-model differentiation, plus a caution against reading general energy counts as O&G rates.", "source_family": "original analysis", "gap_filled": "O&G-specific maturity/use-case evidence organized by strategic purpose, beyond generic adoption-rate reporting.", "reviewed_at": "2026-09-29", "scope_group": "Broader energy", "company_match_ids": [], "publisher_group": "L.E.K. Consulting", "kind": "Adoption / management survey", "display_date": "2026-02-27", "period_group": "Current window" }, { "id": "GAPM004", "title": "The State of AI in Energy, Resources, and Industrials", "publisher": "Deloitte AI Institute", "url": "https://www.deloitte.com/us/en/industries/energy/articles/state-of-ai-energy-sector.html", "published_date": null, "date_note": "Publisher labels this the 2026 report but the public page does not show an exact publication day; survey fieldwork was August–September 2025.", "scope": "Global energy, resources and industrials (ER&I), not oil and gas alone; cross-sector metrics must not be presented as O&G-specific.", "report_type": "survey", "ai_scope": "Enterprise AI including process redesign, agentic AI, physical AI and sovereign/vendor-origin considerations.", "access": "Full publisher sector analysis reviewed; underlying dataset not obtained.", "sample_method": "Deloitte surveyed 3,235 senior leaders globally during August–September 2025, including board/C-suite, president, VP and director levels. Public page does not disclose the O&G subgroup size, selection method or response rate.", "findings": [ "Across ER&I, 40% report AI use with little process change, 37% are redesigning key processes, and 23% report deep business-model transformation; 84% have not redesigned jobs around AI.", "The study covers emerging physical/agentic AI as well as sovereign procurement: cost is the most cited physical-AI barrier, 72% report at least limited physical-AI use, and 83% consider vendor country of origin." ], "metrics": [ { "claim": "ER&I companies reporting deep business-model transformation with AI", "value": "23", "unit": "percent", "period": "2025 survey", "population": "ER&I survey respondents", "basis": "Self-reported Deloitte survey", "caveat": "Broad ER&I subgroup, not O&G-only; does not establish causation or audited financial return." }, { "claim": "ER&I companies reporting at least limited physical AI use", "value": "72", "unit": "percent", "period": "2025 survey", "population": "ER&I survey respondents", "basis": "Self-reported survey", "caveat": "Physical AI spans ER&I and controlled industrial settings; not equivalent to oilfield autonomous deployment." } ], "limitations": "The sample is described as organizations on the leading edge of AI, which may overstate maturity relative to the full industry. ER&I aggregation and undisclosed O&G base limit sector-specific inference.", "departments": [ "operations", "workforce and HR", "IT and AI platforms", "procurement", "manufacturing and facilities", "risk and governance" ], "tags": [ "survey", "physical AI", "agentic AI", "workforce redesign", "sovereign AI" ], "why_useful": "Adds evidence on physical AI, process redesign, job redesign and procurement sovereignty; useful when carefully labeled as cross-industry ER&I evidence rather than an oil-and-gas adoption survey.", "source_family": "original analysis", "gap_filled": "Broadest current sample among the omissions, with distinct workforce, physical-AI and vendor-selection dimensions not captured in upstream value models.", "reviewed_at": "2026-09-29", "scope_group": "Broader energy", "display_date": "2026", "company_match_ids": [], "publisher_group": "Deloitte", "kind": "Adoption / management survey", "period_group": "Current window" }, { "id": "GAPM005", "title": "GETI 2026: The workforce trends reshaping the energy industry", "publisher": "Airswift / Energy Jobline", "url": "https://www.airswift.com/blog/geti-workforce-trends", "published_date": "2026-02-04", "date_note": "Exact date shown on Airswift publisher page.", "scope": "Global traditional energy workforce, including oil and gas and petrochemicals, plus transitional and renewable energy. Traditional-energy figures are not O&G-only.", "report_type": "survey", "ai_scope": "Worker use of AI, career pathways, skills and impacts on the traditional energy talent pipeline.", "access": "excerpt", "sample_method": "Tenth annual GETI survey; publisher describes responses from over 9,000 respondents across energy. The public article does not detail the O&G-only sample, survey instrument, recruitment or response rate; downloadable report is gated.", "findings": [ "Airswift reports AI use by 45% of traditional-energy professionals, up from 24% in 2024, while noting adoption still trails other sectors. It presents AI alongside a shrinking/aging traditional-energy talent pool and changing career pathways.", "For oil and gas employers, the signal is workforce capability and adoption exposure rather than realized productivity: worker usage does not show whether use is sanctioned, embedded in operating processes or generating verified returns." ], "metrics": [ { "claim": "Traditional-energy professionals using AI in their roles", "value": "45", "unit": "percent", "period": "GETI 2026; compared with 24% in 2024", "population": "Traditional energy respondents, including oil and gas and petrochemicals", "basis": "Airswift GETI workforce survey; report said to draw on over 9,000 total energy respondents", "caveat": "Not O&G-only and not company-level deployment; subgroup base/methodology unavailable in public excerpt." } ], "limitations": "The public material provides only selected findings and not the full instrument or subgroup bases. Self-reported individual usage cannot be equated to enterprise adoption, approved access or business value.", "departments": [ "human resources", "learning and development", "operations leadership", "digital transformation", "workforce planning" ], "tags": [ "workforce", "survey", "skills", "AI adoption", "oil and gas talent" ], "why_useful": "Provides a workforce-side check on enterprise AI adoption claims and links usage to skills/career pathways, a dimension thinly covered in operator economics studies.", "source_family": "original analysis", "gap_filled": "Adds individual worker adoption and talent-pipeline perspective; denominators are explicitly broader than oil and gas.", "reviewed_at": "2026-09-29", "scope_group": "Broader energy", "company_match_ids": [], "publisher_group": "Airswift / Energy Jobline", "kind": "Adoption / management survey", "display_date": "2026-02-04", "period_group": "Current window" }, { "id": "GAPF001", "title": "Operationalizing AI in Midstream Oil and Gas: Use Cases, Benefits and Barriers", "publisher": "Alvarez & Marsal", "url": "https://www.alvarezandmarsal.com/sites/default/files/2025-05/Operationalizing%20AI%20in%20Midstream%20Oil%20and%20Gas%20%E2%80%93%20Use%20Cases%2C%20Benefits%20and%20Barriers.pdf", "published_date": null, "date_note": "PDF footer specifies May 2025; day not stated.", "scope": "Midstream oil and gas, spanning pipelines, terminals, transport-to-invoice, trade capture, procure-to-pay, record-to-report, maintenance, and field workforce knowledge.", "report_type": "case synthesis/framework", "ai_scope": "AI and GenAI use-case mapping across operations and functional back office.", "access": "full text", "sample_method": "Consulting perspective with seven workflow/use-case sketches and illustrative company cases; no survey or systematic case sample disclosed.", "findings": [ "The report maps use cases across commercial settlement/BOL validation, pipeline and terminal throughput, predictive maintenance, procurement, invoice reconciliation and field knowledge management.", "It argues that midstream scaling is constrained by fragmented IT/OT, legacy assets and weak data foundations.", "A Shell predictive-maintenance case claims operational improvements, but the PDF gives no asset count, baseline, measurement period or traceable primary citation for the 25% downtime claim; treat it as unverified vendor/consulting case evidence." ], "metrics": [], "limitations": "No methodology or evidence corpus disclosed. Shell case metrics are not sourced in the paper's references and should not be presented as verified. ADNOC value figure is attributed to Reuters and concerns 2023, not a 2025 measurement.", "departments": [ "Pipeline operations", "Terminals", "Commercial settlement", "Supply chain and procurement", "Maintenance", "Finance and accounting", "Field workforce" ], "tags": [ "midstream", "pipeline", "terminal", "procurement", "back-office", "use-case-map" ], "why_useful": "Fills midstream/back-office functional breadth that is thin in the current reports; also surfaces the need to reject unsourced vendor/consulting case metrics. The asset register already has examples for pipeline commercial (AI-170), document checks (AI-171), and Enbridge ILI prioritization (AI-172).", "source_family": "original analysis", "gap_filled": "Broad midstream and back-office workflow taxonomy; connects AI use cases to commercial, procurement and finance processes beyond pipeline integrity.", "company_match_ids": [ "AI-170", "AI-171", "AI-172", "AI-140" ], "reviewed_at": "2026-09-29", "display_date": "2025-05", "publisher_group": "Alvarez & Marsal", "kind": "Case / workflow framework", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF002", "title": "Case Study: AI-Driven Optimization of SWD Pumping Efficiency Using KPI-Based Time Series Analytics", "publisher": "Journal of Petroleum Technology / SPE (case authors from Neuralix)", "url": "https://jpt.spe.org/case-study-ai-driven-optimization-of-swd-pumping-efficiency-using-kpi-based-time-series-analytics", "published_date": "2025-07-01", "date_note": "JPT page dated July 1, 2025.", "scope": "Produced-water saltwater disposal pumping at an anonymized midstream operator in Oklahoma and Texas.", "report_type": "case study", "ai_scope": "AI and first-principles time-series analytics for pump energy optimization and maintenance timing.", "access": "full text", "sample_method": "Vendor-authored case in SPE's Journal of Petroleum Technology; Phase 1 subset and site counts are not disclosed. The client operator is anonymized.", "findings": [ "A KPI-monitoring system ingests SCADA frequency, flow, pressure, electrical readings and power prices, and surfaces site-level kWh/bbl and cost/bbl for field teams.", "The article reports energy reduction and lower cost per barrel at Phase 1 sites, plus projected network-wide gains; the source does not disclose the site denominator or comparison design.", "It also describes an initial proof-of-concept maintenance-timing estimate, separate from the observed Phase 1 pump-energy results." ], "metrics": [ { "claim": "Phase 1 energy reduction", "value": 30, "unit": "percent", "period": "Phase 1; dates not stated", "population": "Subset of injection sites at anonymized midstream operator", "basis": "Vendor-reported case outcome in JPT", "caveat": "Site count, baseline, duration and independent verification not stated; page headline separately says up to 40% electricity savings across a subset." }, { "claim": "Phase 1 reduction in cost per barrel", "value": 14, "unit": "percent", "period": "Phase 1; dates not stated", "population": "Participating SWD sites, number undisclosed", "basis": "Vendor-reported case outcome", "caveat": "No denominator or counterfactual disclosed." }, { "claim": "Projected annualized optimization at full scale", "value": ">40", "unit": "percent", "period": "Projected after scaling; date not stated", "population": "All operator sites, count not stated", "basis": "Vendor projection", "caveat": "Forecast, not achieved network-wide savings." } ], "limitations": "Vendor-authored case; client and sample sizes withheld. The page's headline up-to-40% electricity saving differs from the detailed Phase 1 outcome of 30% energy reduction; preserve the distinction. Maintenance revenue uplift is an initial proof-of-concept estimate, not field-proven.", "departments": [ "Midstream operations", "Produced-water management", "Pumping", "Maintenance", "Energy management" ], "tags": [ "midstream", "saltwater-disposal", "SCADA", "energy-efficiency", "pump-optimization" ], "why_useful": "A rare quantified midstream AI case tied to power and operating KPIs, complementing the current library's pipeline commercial and integrity examples. It is stronger on functional detail than on independent validation.", "source_family": "original case study", "gap_filled": "Operating performance evidence for produced-water disposal and compression/pumping energy, a midstream asset class not represented in existing detailed reports.", "company_match_ids": [], "reviewed_at": "2026-09-29", "publisher_group": "Journal of Petroleum Technology / SPE (case authors from Neuralix)", "kind": "Case / workflow framework", "display_date": "2025-07-01", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF003", "title": "Risk Analysis of Flowlines in the Oil and Gas Sector: A Geographic Information Systems and Machine Learning Approach", "publisher": "SPE Journal / OnePetro", "url": "https://onepetro.org/SJ/article-abstract/30/08/4518/687158/Risk-Analysis-of-Flowlines-in-the-Oil-and-Gas?redirectedFrom=fulltext", "published_date": "2025-05-23", "date_note": "Published online May 23, 2025; SPE Journal 30(8), paper SPE-228286-PA.", "scope": "Oil and gas flowline risk ranking; study data from Colorado Energy and Carbon Management Commission records.", "report_type": "research study", "ai_scope": "GIS plus machine-learning risk classification using logistic regression, support vector machines and gradient-boosted decision trees.", "access": "abstract", "sample_method": "Peer-reviewed study using regulator-sourced flowline data, spatial matching, feature engineering, and reduced-feature analyses (PCA/CCA); full article is purchase/access controlled in OnePetro.", "findings": [ "The paper addresses smaller flowlines between wellheads and facilities, which the authors say receive less risk assessment than transmission pipelines.", "It combines spatial and operating features to identify higher-risk zones for prioritized monitoring and mitigation.", "The accessible abstract describes the modeling pipeline but does not report validated classification performance or field outcomes." ], "metrics": [], "limitations": "Only the abstract was accessible during review; training/test split, regulator dataset size, class balance, sensitivity/false-negative rate and external validation could not be verified. Risk-ranking output is not evidence of prevented failures.", "departments": [ "Flowline integrity", "Pipeline operations", "HSE", "Risk management", "Inspection and maintenance" ], "tags": [ "flowlines", "pipeline-integrity", "GIS", "machine-learning", "peer-reviewed" ], "why_useful": "Adds a peer-reviewed, flowline-specific integrity method to complement the current Enbridge ILI-prioritization case (AI-172), while making its validation gap explicit.", "source_family": "original research", "gap_filled": "Flowline risk screening evidence; differentiates small gathering/production flowlines from larger transmission-pipeline programs.", "company_match_ids": [], "reviewed_at": "2026-09-29", "publisher_group": "SPE Journal / OnePetro", "kind": "Case / workflow framework", "display_date": "2025-05-23", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF005", "title": "AGM 2025 CEO and Chair speeches: AI in LNG operations and energy trading", "publisher": "Shell", "url": "https://www.shell.com/investors/shareholder-meetings/_jcr_content/root/main/section/simple/list/list_item_1094738722.multi.stream/1749109727574/15efecffebe3e1aa420a725ff455aaa9b3c356cd/ceo-and-chair-speeches-agm-2025.pdf", "published_date": null, "date_note": "2025 AGM source materials; exact publication day not verified.", "scope": "Shell LNG facilities and energy trading, with a specific offshore-wind forecasting example.", "report_type": "company disclosure", "ai_scope": "AI for LNG emissions monitoring, predictive maintenance and electricity trading support.", "access": "full text", "sample_method": "Executive AGM remarks, not a research study. Company describes active applications without disclosing site counts, model design or measured outcomes.", "findings": [ "Shell says AI monitors sensor data at LNG facilities to reduce emissions and supports predictive maintenance.", "Shell also says it is applying AI to electricity trading by improving management of power flows and response to price volatility; ML forecasting for offshore wind is one cited example.", "The disclosure identifies functions and operational direction but offers no quantified impact or LNG-specific deployment scale." ], "metrics": [], "limitations": "Executive statements without site-by-site evidence, baselines or outcome metrics. The trading example is electricity/offshore-wind oriented, not a quantified hydrocarbon or LNG cargo trading result.", "departments": [ "LNG operations", "Emissions management", "Maintenance", "Energy trading", "Power forecasting" ], "tags": [ "LNG", "trading", "emissions", "predictive-maintenance", "company-disclosure" ], "why_useful": "Adds dated primary-source evidence for LNG and trading workflows, where the current case library has sparse energy-process details. Avoid reading intent/application descriptions as measured value.", "source_family": "original company disclosure", "gap_filled": "LNG process monitoring and downstream commercial/trading workflow coverage.", "company_match_ids": [ "AI-025" ], "reviewed_at": "2026-09-29", "display_date": "2025", "publisher_group": "Shell", "kind": "Case / workflow framework", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF006", "title": "Plantwide AI optimisation and beyond: Big West Oil refinery case", "publisher": "Petroleum Technology Quarterly (PTQ), with Big West Oil and Imubit authors", "url": "https://ptqmagazines.digitalrefining.com/PTQ-Q3-2025/57/", "published_date": null, "date_note": "PTQ Q3 2025 issue; exact day not stated. Web issue page number 57.", "scope": "Big West Oil's 36,000 b/d refinery in North Salt Lake City, Utah.", "report_type": "case study", "ai_scope": "Closed-loop AI optimization across FCC/debutanizer, diesel pool, hydroprocessing, CDU molecular management and feedstock trade-offs.", "access": "full text", "sample_method": "Named refinery case written by company and solution-provider personnel; no independent evaluator or published controlled outcome series.", "findings": [ "Deployment began with an FCC/debutanizer bottleneck and extended to ULSD flash giveaway, hydroprocessing, crude-distillation molecular management and seasonal feedstock trade-offs.", "The case emphasizes operator/engineer training alongside closed-loop process control, and describes data quality, governance, infrastructure and change management as implementation hurdles.", "Public issue text reviewed here does not provide a quantified margin/yield effect or baseline, despite the headline's margin-improvement framing." ], "metrics": [], "limitations": "Co-authored by operator and vendor; no counterfactual, independent audit or quantified improvement series in the accessible page. Do not infer value from claimed potential or rollout breadth.", "departments": [ "Refining", "FCC operations", "Crude distillation", "Hydroprocessing", "Supply planning", "Workforce learning" ], "tags": [ "refining", "FCC", "closed-loop-control", "yield-optimization", "workforce" ], "why_useful": "Fills a detailed independent-operator refinery workflow description and highlights adoption/change needs beyond AI for upstream production. It is distinct from the current Preem FCC announcement (AI-169).", "source_family": "original case study", "gap_filled": "Refinery AI across multiple process units and operator learning, with explicit disclosure that public outcomes are not quantified.", "company_match_ids": [], "reviewed_at": "2026-09-29", "display_date": "2025 Q3", "publisher_group": "Petroleum Technology Quarterly (PTQ), with Big West Oil and Imubit authors", "kind": "Case / workflow framework", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF007", "title": "Implementation of reinforcement learning for enhanced pressure control in a 190,000-barrel crude distillation unit: The first full-scale commercial deployment", "publisher": "Engineering Applications of Artificial Intelligence / Elsevier", "url": "https://www.sciencedirect.com/science/article/pii/S0952197625009078", "published_date": "2025-09-15", "date_note": "Volume 156, Part A, article 110907, published September 15, 2025.", "scope": "Full-scale pressure control in a 190,000 b/d crude distillation unit; likely Korean refinery context, but the article abstract does not identify the operator by name.", "report_type": "research study and industrial case", "ai_scope": "Offline-to-online reinforcement learning controller (Soft Actor-Critic with conservative loss and Return-to-Go) connected to a legacy control system.", "access": "abstract", "sample_method": "Peer-reviewed paper reports deployment in a real CDU; accessible abstract gives no site/operator name, observation period, error definition, intervention denominator or independent replication.", "findings": [ "The controller learned from historical manual-operation logs and automated pressure-control decisions in a 190,000 b/d CDU; authors describe this as permanent-operation deployment, not a simulation demonstration.", "The paper reports fewer operator interventions and lower cumulative pressure-control errors.", "It is a rare downstream AI outcome report, but operator identity and details needed to evaluate comparability are not visible in the abstract." ], "metrics": [ { "claim": "Reduction in operator intervention time", "value": 84, "unit": "percent", "period": "deployment period not specified in abstract", "population": "190,000 b/d CDU control workflow", "basis": "Peer-reviewed case result as reported by authors", "caveat": "Intervention-time measure and observation window not specified in accessible abstract; full-paper audit required." }, { "claim": "Reduction in cumulative pressure-control errors", "value": 12.8, "unit": "percent", "period": "deployment period not specified in abstract", "population": "190,000 b/d CDU pressure control", "basis": "Peer-reviewed case result as reported by authors", "caveat": "Error definition, baseline, uncertainty and independent validation not specified in accessible abstract." } ], "limitations": "Abstract-level review only. Full text is needed to verify operator, controls/safety envelope, time windows, error calculation, sustained performance and economics. Do not infer enterprise-scale refinery gains from one unit.", "departments": [ "Refining", "Crude distillation", "Process control", "Operations engineering", "Automation" ], "tags": [ "refining", "reinforcement-learning", "process-control", "closed-loop", "peer-reviewed" ], "why_useful": "The strongest identified public quantitative example of AI closed-loop control in an actual refinery unit; helps distinguish field deployment from simulation-based process optimization.", "source_family": "original peer-reviewed study", "gap_filled": "Measured refinery process-control performance rather than only broad downstream roadmaps or unquantified company announcements.", "company_match_ids": [], "reviewed_at": "2026-09-29", "publisher_group": "Engineering Applications of Artificial Intelligence / Elsevier", "kind": "Case / workflow framework", "display_date": "2025-09-15", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPF008", "title": "The Real Cost Advantage in Oil and Gas", "publisher": "Boston Consulting Group", "url": "https://www.bcg.com/publications/2025/real-cost-advantage-oil-gas", "published_date": "2025-08-26", "date_note": "BCG page dated August 26, 2025.", "scope": "Oil and gas procurement and supply chain, especially operator sourcing and supplier risk.", "report_type": "model and case synthesis", "ai_scope": "GenAI/AI tender assistants, supplier analysis, price/should-cost modeling and supply-chain risk control towers.", "access": "full text", "sample_method": "Consulting analysis and illustrative cases; published page does not disclose a survey sample or measured cross-company intervention dataset.", "findings": [ "BCG estimates procurement and supply-chain functions offer material AI/GenAI automation and augmentation potential, covering RFP drafting, bid analysis, supplier engagement and risk monitoring.", "It describes AI-enabled procurement copilots and control towers as ways to shift resources to strategic work and improve supplier-risk visibility.", "Most numerical productivity and cost figures are estimates/potential, not observed O&G adoption averages or verified company outcomes." ], "metrics": [ { "claim": "Procurement/supply-chain activities AI/GenAI can automate or augment", "value": "around 75", "unit": "percent", "period": "current estimate in 2025 article", "population": "Oil and gas procurement and supply chain activities", "basis": "BCG estimate", "caveat": "Addressability estimate, not observed automation or O&G survey statistic." }, { "claim": "Potential productivity cost savings on core procurement workflows", "value": "10–15", "unit": "percent", "period": "potential; timing not specified", "population": "Oil and gas procurement functions", "basis": "BCG estimate based on described AI opportunities", "caveat": "Potential, not realized savings; method and company sample not disclosed." } ], "limitations": "The page is a consulting analysis rather than a survey or outcome evaluation. Estimates are not audited; examples do not establish typical adoption. Its separate peer advantage estimates should not be conflated with AI-specific productivity estimates.", "departments": [ "Procurement", "Supply chain", "Supplier management", "Commercial", "Logistics", "Risk management" ], "tags": [ "procurement", "supply-chain", "GenAI", "supplier-risk", "value-potential" ], "why_useful": "Adds an O&G-specific procurement and supply-chain framework to a collection where detailed cases are scarce; keep its potential estimates separate from deployments such as ExxonMobil ERP/logistics (AI-024), TotalEnergies BuyerCompanion (AI-167) and Baker Hughes sourcing (AI-165).", "source_family": "original analysis", "gap_filled": "Functional research coverage for procurement, supplier risk, tendering, sourcing and supply-chain value capture.", "company_match_ids": [ "AI-024", "AI-165", "AI-167" ], "reviewed_at": "2026-09-29", "publisher_group": "Boston Consulting Group", "kind": "Case / workflow framework", "display_date": "2025-08-26", "period_group": "Current window", "scope_group": "Oil & gas" }, { "id": "GAPE004", "title": "Full-field petroleum production forecasting: Is deep learning all you need?", "publisher": "Petroleum Research / Chinese Petroleum Society (Elsevier)", "url": "https://www.sciencedirect.com/science/article/pii/S2096249526000074", "published_date": "2026-01-21", "date_note": "ScienceDirect gives online availability date 21 January 2026; volume issue metadata may postdate online publication.", "scope": "Full-field petroleum production forecasting, including oilfield management scenarios such as waterflood scheduling and production loss from schedule delays; real-world datasets including Volve.", "report_type": "technical study", "ai_scope": "Comparison of machine-learning and deep-learning proxy models across four datasets and three forecast horizons for field-scale production.", "access": "Published online 21 January 2026. We reviewed the publisher’s indexed abstract; direct full-article access was unavailable.", "sample_method": "Abstract describes a peer-reviewed empirical comparison using four real-world datasets and three forecast horizons. Abstract reports validation and comparative analysis; exact model list, split protocol, sample counts, and detailed metrics should be taken from the article tables before quantitative reuse.", "findings": [ "Across the study's multi-dataset and multi-horizon comparison, kernel-based methods consistently outperform neural architectures; the authors report no tangible accuracy benefit from deeper models in this full-field forecasting setting.", "The authors find generalization appears more sensitive to spatial density of production data than to total dataset size, a practical warning against reading raw well or row counts as sufficient training coverage.", "The study argues compact models can offer useful accuracy at lower computational cost, supporting fit-for-purpose validation rather than defaulting to more complex deep learning." ], "metrics": [], "limitations": "A comparative technical study, not an operating deployment or ROI evaluation. Abstract-level findings do not supply cost amounts or field decision outcomes. Validate exact model, split, and error metrics from the full article before citing a numeric accuracy difference.", "departments": [ "reservoir engineering", "production engineering", "waterflood planning", "digital and data science" ], "tags": [ "model validation", "production forecast", "algorithm comparison", "data density", "Green AI" ], "why_useful": "Adds an unusually systematic validation counterweight to deployment claims and directly tests whether complex models outperform simpler alternatives under sparse field data.", "source_family": "original analysis", "gap_filled": "Independent model-comparison and generalization evidence for model choice/validation; exposes the lack of measured production economics in algorithm studies.", "company_ids": [], "companies_mentioned": [ "Volve field" ], "access_note": "Publisher marks article open access under a Creative Commons license; the abstract is accessible and full text is linked by publisher.", "reviewed_at": "2026-10-14", "scope_group": "Upstream", "company_match_ids": [], "publisher_group": "Petroleum Research / Chinese Petroleum Society (Elsevier)", "kind": "Technical case", "display_date": "2026-01-21", "period_group": "Current window" }, { "id": "GAPE005", "title": "Federated Transformer-Based ROP Prediction Under Geological Heterogeneity and Privacy Constraints", "publisher": "Results in Engineering / Elsevier", "url": "https://doi.org/10.1016/j.rineng.2025.107924", "published_date": null, "date_note": "Volume 28 bibliographic metadata gives December 2025; online publication page should be checked for exact first-online date before day-level use.", "scope": "Cross-field rate-of-penetration prediction using Volve, XJ, and BH oilfield datasets with differing geological and operating conditions.", "report_type": "technical study", "ai_scope": "Horizontal federated learning using a Transformer and sample-weighted aggregation to forecast drilling ROP without pooling raw field data.", "access": "Abstract and bibliographic record reviewed; linked full paper not reviewed.", "sample_method": "Peer-reviewed experiment on real-world data from three oilfields. Reported comparisons include local and centralized baselines and long-sequence forecasting; the abstract does not disclose well counts, train/test partition, or a blinded prospective field trial.", "findings": [ "The proposed federated Transformer combines data across Volve, XJ, and BH while keeping raw data local, addressing cross-company privacy and geological heterogeneity constraints.", "The abstract reports up to 72% lower mean squared error than comparison baselines in long-sequence forecasting, with gains particularly noted in data-scarce single-well settings.", "This is model-performance evidence, not evidence of reduced nonproductive time, lower drilling cost, or sustained rig deployment." ], "metrics": [], "limitations": "Abstract and public metadata establish a promising privacy-preserving validation setup, but do not establish a production deployment, rig-level economics, or robustness under prospective field shifts. Treat the 72% as a task-specific maximum, not expected operational uplift.", "departments": [ "drilling engineering", "well planning", "digital and data science", "cybersecurity and data governance" ], "tags": [ "drilling", "ROP", "federated learning", "cross-field validation", "data privacy" ], "why_useful": "Directly addresses data-sharing constraints that impede small operators and regional transfer, while offering measurable technical validation instead of only a proposed architecture.", "source_family": "original analysis", "gap_filled": "Cross-field model validation, privacy-preserving collaboration, and geographic/geological transfer; does not prove cost savings or field adoption.", "company_ids": [], "companies_mentioned": [ "Volve field", "XJ oilfield", "BH oilfield" ], "access_note": "Publisher/secondary scholarly metadata identifies the article as open access; DOI points to publisher. Use publisher full text for detailed method and exact publication date.", "reviewed_at": "2026-09-29", "display_date": "2025-12", "scope_group": "Upstream", "company_match_ids": [], "publisher_group": "Results in Engineering / Elsevier", "kind": "Technical case", "period_group": "Current window" }, { "id": "GAPR001", "title": "Energy and utilities firms investing aggressively in AI despite mixed returns", "publisher": "S&P Global / 451 Research", "url": "https://www.spglobal.com/energy/en/technology-ai-research-insights/blogs/energy-and-utilities-firms-investing-aggressively-in-ai-despite-mixed-returns", "published_date": "2026-03-31", "date_note": "Published 31 March 2026; looks back 12 months.", "scope": "Utilities and oil & gas combined.", "scope_group": "Broader energy", "report_type": "survey", "ai_scope": "Multiple AI classes.", "access": "Public survey analysis reviewed; underlying microdata not reviewed.", "sample_method": "Voice of the Enterprise: AI & Machine Learning, Use Cases 2026. IT/business respondents; sector N and field dates absent from public article.", "findings": [ "Respondents describe terminated projects and post-deployment shutdowns, addressing a gap in success-heavy case collections.", "Sourcing includes internal builds, third-party custom solutions and application add-ons; categories overlap.", "The survey separately reports 37% of initiatives started in the prior 12 months as operational and delivering value, and 63% of respondents expect most projects not to return positively within 12 months; these are different denominators and neither is O&G-only." ], "metrics": [ { "claim": "Projects terminated, including 9% decommissioned after deployment", "value": "34", "unit": "%", "period": "Initiated in preceding 12 months", "population": "Energy/utilities survey projects", "basis": "Respondent-reported", "caveat": "Not an O&G-only failure rate." }, { "claim": "Projects expected to achieve ROI within 12 months", "value": "43", "unit": "%", "period": "Previous-year initiatives", "population": "Same survey", "basis": "Expectation", "caveat": "Not realized ROI." }, { "claim": "Initiatives started in the past 12 months perceived by respondents as successfully deployed, currently operational and delivering business value", "value": "37", "unit": "%", "period": "Initiated in preceding 12 months", "population": "AI initiatives in the combined utilities and oil-and-gas survey; initiative-level denominator as classified/perceived by respondents", "basis": "Respondent classification / perception", "caveat": "Not share of companies or O&G-only. Survey does not provide sector subsample N or microdata." }, { "claim": "Respondents forecasting that most of their projects will not deliver positive ROI within 12 months of launch", "value": "63", "unit": "%", "period": "12 months from project launch", "population": "Respondents in the combined utilities and oil-and-gas survey; respondent-level denominator, each assessing most of their own projects", "basis": "Respondent expectation / forecast", "caveat": "Not share of projects, not failure rate, and not complement of the 43% project measure. S&P says this does not necessarily mean projects are considered failed." } ], "limitations": "Unknown sector sample; self-reports; mixed utilities and oil-and-gas population. The 37% is an initiative-status classification, while 63% is a respondent-level expectation about most projects. Termination is distinct from delayed payback; no sector subsample N or microdata is public.", "departments": [ "IT", "Supply chain" ], "tags": [ "failures", "sourcing", "survey" ], "why_useful": "Adds project attrition to the investment case.", "source_family": "451 Research Use Cases 2026", "gap_filled": "Project failures and sourcing routes", "reviewed_at": "2026-10-06", "company_match_ids": [], "publisher_group": "S&P Global / 451 Research", "kind": "Adoption / management survey", "display_date": "2026-03-31", "period_group": "Current window" }, { "id": "GAPR002", "title": "From Paradox to Progress: A Net-Positive AI Energy Framework", "publisher": "World Economic Forum / Accenture", "url": "https://www.weforum.org/publications/from-paradox-to-progress-a-net-positive-ai-energy-framework/", "published_date": "2025-12-11", "date_note": "Publisher date: 11 December 2025.", "scope": "Energy systems globally; broader than oil and gas.", "scope_group": "Broader energy", "report_type": "framework", "ai_scope": "AI energy consumption and operational applications.", "access": "Publisher summary reviewed. Linked PDF timed out; interactive catalog identified but not extracted.", "sample_method": "Framework referencing a repository of over 130 energy use cases; not an adoption survey.", "findings": [ "A broad external case inventory can help check omissions from our company register.", "The framework considers efficiency, application impact and demand management together." ], "metrics": [], "limitations": "The 130+ cases are not all O&G, nor verified additions to this register. Case-level provenance still needs review.", "departments": [ "Energy management", "IT" ], "tags": [ "case-catalog", "energy-systems" ], "why_useful": "Provides an external inventory for a systematic case-by-case reconciliation.", "source_family": "WEF / Accenture AI Energy 2025", "gap_filled": "External case inventory and energy-system boundary", "supplemental_urls": [ "https://public.tableau.com/app/profile/world.economic.forum.data.analytics/viz/AIEnergy2025/AIEnergy", "https://reports.weforum.org/docs/WEF_From_Paradox_to_Progress_A_Net_Positive_AI_Energy_Framework_2025.pdf" ], "reviewed_at": "2026-09-29", "company_match_ids": [], "publisher_group": "World Economic Forum / Accenture", "kind": "Strategy / governance", "display_date": "2025-12-11", "period_group": "Current window" }, { "id": "GAPR003", "title": "AIASM: gas-asset surveillance feasibility, study costs and Alpha funding decision", "publisher": "National Gas / Ofgem", "url": "https://www.nationalgas.com/sites/default/files/documents/AIASM%20SIF%20Discovery%20Final%20Report%20November%202025.pdf", "published_date": "2025-11", "display_date": "2025–2026", "date_note": "2025 Discovery final report; February 2026 Alpha decision. Ofgem authorized Discovery for May–September 2025; the report includes August/September materials.", "scope": "UK gas transmission surveillance research; one named innovation programme.", "scope_group": "Broader energy", "report_type": "innovation feasibility study / regulator funding decision", "ai_scope": "AI-assisted vegetation and third-party-interference surveillance; planned multimodal sensing.", "access": "Full operator final report and regulator decision reviewed.", "sample_method": "Discovery work packages, feasibility/design assessment and finance update; no live-network outcome evaluation.", "findings": [ "The completed Discovery study assessed surveillance workflows, sensing and AI architecture. Legacy integration, data formats and APIs constrain implementation.", "The final report discloses actual study expenditure; operating detection performance and cash ROI are not demonstrated.", "Ofgem declined successor Alpha application 10179002 because it did not meet the challenge’s transition-planning objective. The regulator acknowledged possible operational efficiencies." ], "metrics": [ { "value": "91,153.23", "unit": "GBP", "claim": "Actual Discovery study expenditure", "period": "2025 Discovery phase", "population": "Project 10157410", "basis": "Reported finance update, physical p. 19", "caveat": "Study spend, not deployment cost or payback. Approved funding £97,650; reported underspend £6,496.77." }, { "value": "497,092", "unit": "GBP", "claim": "Alpha funding requested, not awarded", "period": "February 2026 decision", "population": "Successor application 10179002", "basis": "Ofgem table physical p. 37; AIASM-specific decision physical pp. 58–64", "caveat": "Proposed total project cost £552,435; request was declined. Not actual spending or proof of technical failure." } ], "limitations": [ "No production rollout or measured ROI demonstrated. A rejected funding application does not prove technical failure, permanent cancellation or absence of other funding.", "Funding rationale is specific to this challenge; do not substitute the preceding Digital Decommissioning project’s objections." ], "departments": [ "Maintenance & reliability", "HSE & emissions", "IT & data" ], "tags": [ "implementation cost", "feasibility", "funding declined", "gas transmission", "data integration" ], "why_useful": "A named example showing what feasibility costs and why an AI proposal can miss its funding criteria.", "source_family": "AIASM Discovery 10157410 / Alpha 10179002", "publisher_group": "National Gas / Ofgem", "kind": "Feasibility / funding decision", "period_group": "Current window", "reviewed_at": "2026-10-06", "supplemental_urls": [ { "url": "https://www.ofgem.gov.uk/sites/default/files/2026-02/Cycle_4_Alpha_Funding_Decision_and_Summary.pdf", "label": "Ofgem Alpha decision · table p. 37, AIASM section pp. 58–64" }, { "url": "https://www.ofgem.gov.uk/sites/default/files/2025-04/10157410%20Gas%20transmission%20asset%20resilience%20through%20network%20transitions%20PD.pdf", "label": "Ofgem Discovery direction · project 10157410" } ], "company_match_ids": [] }, { "id": "TRNREP001", "title": "Putting AI to work: The operational excellence imperative", "publisher": "McKinsey & Company", "publisher_group": "McKinsey & Company", "url": "https://www.mckinsey.com/capabilities/operations/our-insights/putting-ai-to-work-the-operational-excellence-imperative", "published_date": "2026-06-19", "display_date": "2026-06-19", "date_note": "2026-06-19 (survey conducted Oct 7–Nov 10,2025)", "report_type": "survey", "kind": "Adoption survey", "scope": "Worldwide firms with at least $500 million revenue and 100 employees; respondents were managers and executives.", "scope_group": "Cross-industry", "period_group": "Current window", "ai_scope": "AI within broader operating and business-process changes; scope varies by source.", "access": "Original publisher page and cited methodology reviewed.", "sample_method": "October 7–November 10,2025 survey: 1,000 managers/executives representing 696 companies; 677 matched to 2014–2024 financial data. Study associations are not causal. The headline AI adoption estimates are separately cited background research.", "findings": [ "Operational excellence and AI adoption are associated with stronger financial performance in a cross-industry study. The article separately cites adoption estimates of almost 90% experimenting and 7% scaling; these are not an oil-and-gas failure rate." ], "metrics": [], "limitations": "Self-reported cross-sectional survey. Enterprise-scale definition and causal direction should not be inferred. Performance comparison may reflect other organizational differences.", "departments": [ "operations", "IT", "executive leadership", "HR" ], "tags": [ "business transformation", "workflow redesign", "operating model", "implementation" ], "why_useful": "Useful cross-industry adoption benchmark, not an O&G deployment failure rate.", "source_family": "Original consulting survey analysis", "reviewed_at": "2026-10-07" }, { "id": "TRNREP002", "title": "The state of AI: How organizations are rewiring to capture value", "publisher": "McKinsey & Company", "publisher_group": "McKinsey & Company", "url": "https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-how-organizations-are-rewiring-to-capture-value", "published_date": "2025-03-12", "display_date": "2025-03-12", "date_note": "Published 12 March 2025; survey conducted 16–31 July 2024.", "report_type": "survey", "kind": "Adoption survey", "scope": "Global cross-industry survey of organizations using AI/gen AI; this particular 21% denominator is respondents reporting gen-AI use.", "scope_group": "Cross-industry", "period_group": "Current window", "ai_scope": "AI within broader operating and business-process changes; scope varies by source.", "access": "Original publisher page and cited methodology reviewed.", "sample_method": "July 16–31,2024 global survey, 1,491 respondents in 101 nations. The 21% figure uses the subset reporting GenAI use; this is older observation evidence published within the requested disclosure window.", "findings": [ "21% of respondents reporting GenAI use say at least some workflows were fundamentally redesigned. Redesign was strongly associated with self-reported EBIT impact among the tested organizational practices." ], "metrics": [], "limitations": "Self-reported survey and association, not proof that redesign caused EBIT impact. It does not measure oil-and-gas firms specifically or establish that unchanged workflows mean a failed deployment.", "departments": [ "operations", "IT", "executive leadership", "HR" ], "tags": [ "business transformation", "workflow redesign", "operating model", "implementation" ], "why_useful": "Supports workflow redesign as a likely scale/value enabler, with survey-level rather than operator outcome evidence.", "source_family": "Original consulting survey analysis", "reviewed_at": "2026-10-07" }, { "id": "TRNREP003", "title": "From pilots to performance: How COOs can scale AI in manufacturing", "publisher": "McKinsey & Company", "publisher_group": "McKinsey & Company", "url": "https://www.mckinsey.com/capabilities/operations/our-insights/from-pilots-to-performance-how-coos-can-scale-ai-in-manufacturing", "published_date": "2025-12-15", "display_date": "2025-12-15", "date_note": "2025-12-15 (survey conducted June–July 2025)", "report_type": "survey", "kind": "Adoption survey", "scope": "Senior operating executives at manufacturers with at least $1 billion revenue; 17% of respondent organizations were in energy and materials, but results are not separately reported for them.", "scope_group": "Cross-industry", "period_group": "Current window", "ai_scope": "AI within broader operating and business-process changes; scope varies by source.", "access": "Original publisher page and cited methodology reviewed.", "sample_method": "June–July 2025 survey of 101 COOs at manufacturers with at least $1 billion revenue. Energy and materials represented 17% of respondents; sector-specific results were not reported.", "findings": [ "Only 2% of surveyed manufacturing COOs describe AI as fully embedded across operations; 46% cite data or IT/OT limitations. The article discusses reusable software and frontline capabilities as deployment constraints." ], "metrics": [], "limitations": "Small executive survey and self-reported maturity/barriers. These are not observed failure rates. Recommendations and SQM miner illustration are consulting case reporting, not a controlled comparison.", "departments": [ "operations", "IT", "executive leadership", "HR" ], "tags": [ "business transformation", "workflow redesign", "operating model", "implementation" ], "why_useful": "Useful implementation analog for change capability, IT/OT integration, and reusable deployment design.", "source_family": "Original consulting survey analysis", "reviewed_at": "2026-10-07" }, { "id": "TRNREP004", "title": "5 takeaways from the world’s largest dataset on industrial transformation", "publisher": "World Economic Forum", "publisher_group": "World Economic Forum", "url": "https://www.weforum.org/stories/emerging-technologies/takeaways-from-the-world-largest-dataset-industrial-transformation/", "published_date": "2025-10-29", "display_date": "2025-10-29", "date_note": "2025-10-29 (dataset spans 2017–2025)", "report_type": "case synthesis", "kind": "Selected industrial cases", "scope": "1,085 transformation solutions across 32 countries and 35 industries from the WEF Lighthouse Network and partner submissions. The case examples are selected industrial sites, not a representative sample of all firms.", "scope_group": "Cross-industry", "period_group": "Current window", "ai_scope": "AI within broader operating and business-process changes; scope varies by source.", "access": "Original publisher page and cited methodology reviewed.", "sample_method": "1,085 transformation solutions across 32 countries and 35 industries from the WEF Lighthouse Network and partner submissions. The case examples are selected industrial sites, not a representative sample of all firms.", "findings": [ "The selected industrial dataset links mature operating processes and workforce capabilities to faster deployment. Most transformation programs combine several technologies; standalone-model results are a poor proxy for whole-program outcomes." ], "metrics": [], "limitations": "The Lighthouse network selects advanced sites; reported KPI deltas and correlations are not independently audited causal estimates. The 50% pilot-to-scale result is an aggregate comparison and may reflect selection, definitions, or co-occurring management practices. The metals producer is not named in the article.", "departments": [ "operations", "IT", "executive leadership", "HR" ], "tags": [ "business transformation", "workflow redesign", "operating model", "implementation" ], "why_useful": "Strong cross-industry analog for process maturity, integration, and workforce capability; selected-case evidence must be labeled.", "source_family": "Selected submitted site cases", "reviewed_at": "2026-10-07" }, { "id": "TRNREP005", "title": "New Global Lighthouse Sites Demonstrate How AI Is Rewiring Manufacturing and Supply Chains", "publisher": "World Economic Forum", "publisher_group": "World Economic Forum", "url": "https://www.weforum.org/media/new-global-lighthouse-sites-demonstrate-how-ai-is-rewiring-manufacturing-and-supply-chains/", "published_date": "2026-06-22", "display_date": "2026-06-22", "date_note": "2026-06-22", "report_type": "case synthesis", "kind": "Selected industrial cases", "scope": "Two selected lighthouse sites; outcomes cover whole transformation packages involving process redesign, automation, robotics, data systems, and AI, not isolated AI effects.", "scope_group": "Cross-industry", "period_group": "Current window", "ai_scope": "AI within broader operating and business-process changes; scope varies by source.", "access": "Original publisher page and cited methodology reviewed.", "sample_method": "Two selected lighthouse sites; outcomes cover whole transformation packages involving process redesign, automation, robotics, data systems, and AI, not isolated AI effects.", "findings": [ "Named cases include Merck’s shift to continuous biologics processing and NIO’s connected product-development system. Their outcomes concern broad operating changes, not isolated AI contributions." ], "metrics": [], "limitations": "WEF press release and site-submitted outcomes; no counterfactual, independent audit, baseline details, or AI-only attribution. Merck is life sciences manufacturing and NIO automotive; these are implementation analogs rather than oilfield results.", "departments": [ "operations", "IT", "executive leadership", "HR" ], "tags": [ "business transformation", "workflow redesign", "operating model", "implementation" ], "why_useful": "Concrete successful analogs that show measurable gains accompany end-to-end operating-system change and integration, not a stand-alone model.", "source_family": "Selected submitted site cases", "reviewed_at": "2026-10-07" }, { "id": "O14B-KPMG-GLOBAL", "title": "KPMG Global tech report 2026: Energy, Natural Resources and Chemicals", "publisher": "KPMG International", "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", "published_date": "2026", "display_date": "2026 edition · exact day unknown", "date_note": "2026 report edition; exact publication day and survey fieldwork dates were not established in this review.", "scope": "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.", "report_type": "survey", "ai_scope": "AI and automation; multiple technology classes", "access": "Original PDF text reviewed; Printed pages 6–7: priorities and AI strategy; page 16: methodology", "sample_method": "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.", "findings": [ "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.", "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." ], "metrics": [ { "claim": "AI/automation investment priority", "value": "79", "unit": "percent of surveyed respondents", "period": "2026 edition; fieldwork dates not established in this review", "population": "58 oil-and-gas technology leaders; broader energy/extractives comparison 69%; cross-sector 77%", "basis": "Self-reported survey responses", "caveat": "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." } ], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "survey", "technology maturity", "AI priorities" ], "why_useful": "Do the priorities translate into sustained use and measured outcomes on comparable assets?", "source_family": "original survey analysis", "publisher_group": "KPMG International", "kind": "Adoption survey", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-10-14" }, { "id": "O14B-KPMG-US", "title": "How energy can scale advanced tech safely", "publisher": "KPMG US", "url": "https://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2026/how-energy-can-scale-advanced-tech-safely.pdf", "published_date": "2026", "display_date": "2026 edition · exact day unknown", "date_note": "2026 report edition; exact publication day not established.", "scope": "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.", "report_type": "survey", "ai_scope": "AI and automation; multiple technology classes", "access": "Original PDF text reviewed; PDF page 3 / printed page 1: population; PDF page 5 / printed page 3: maturity comparisons", "sample_method": "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.", "findings": [ "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.", "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." ], "metrics": [ { "claim": "Advanced AI/automation implementation", "value": "52", "unit": "percent of surveyed respondents", "period": "2026 edition; fieldwork dates not established in this review", "population": "US energy comparison: 57% elsewhere", "basis": "Self-reported survey responses", "caveat": "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." }, { "claim": "Advanced data/analytics implementation", "value": "45", "unit": "percent of surveyed respondents", "period": "2026 edition; fieldwork dates not established in this review", "population": "US energy comparison: 60% elsewhere", "basis": "Self-reported survey responses", "caveat": "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." } ], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "survey", "technology maturity", "AI priorities" ], "why_useful": "Which data and integration bottlenecks prevent useful systems from becoming part of everyday work?", "source_family": "original survey analysis", "publisher_group": "KPMG US", "kind": "Adoption survey", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-10-14" }, { "id": "O14C-SURFACE", "title": "AI and ML for production prediction and optimization of oil and gas surface networks", "publisher": "Energy and AI / Elsevier", "url": "https://www.sciencedirect.com/science/article/pii/S2666546826000984", "published_date": "2026-09", "display_date": "2026-09", "date_note": "September 2026 journal edition; literature search through March 2026. Relevant publisher text was available through the web index.", "scope": "Oil and gas surface production networks", "report_type": "literature review", "ai_scope": "Numerical ML and deep learning; not an LLM deployment study", "access": "September 2026 journal edition; literature search through March 2026. Relevant publisher text was available through the web index.", "sample_method": "The authors reviewed prediction, monitoring and control literature from January 2015 through March 2026. They argue for physically defensible recommendations, oversight and protected actuation.", "findings": [ "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.", "This review does not measure incident frequency, failure prevalence or the performance of a named operator’s deployed system." ], "metrics": [], "limitations": "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.", "departments": [ "Production & field operations", "Reservoir engineering & field development" ], "tags": [ "technical evidence", "model validation", "limits" ], "why_useful": "What detects drift, handles unfamiliar conditions and prevents a recommendation from becoming an unsafe action?", "source_family": "Original technical research", "publisher_group": "Energy and AI / Elsevier", "kind": "Literature review", "period_group": "Current window", "scope_group": "Upstream oil & gas", "reviewed_at": "2026-10-14" }, { "id": "WATCH14PM-BCG", "title": "The Formula for Agentic AI Value / Applied AI Index 2026", "publisher": "Boston Consulting Group", "url": "https://www.bcg.com/publications/2026/the-formula-for-agentic-ai-value", "published_date": "2026-09-30", "display_date": "2026-09-30", "date_note": "Published 30 September 2026. Survey fieldwork dates and an oil-and-gas denominator were not established in the reviewed material.", "scope": "Cross-sector; no upstream subgroup disclosed", "report_type": "survey commentary", "ai_scope": "Enterprise AI deployment, governance and planned investment", "access": "Original publisher article and relevant method notes read", "sample_method": "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.", "findings": [ "BCG’s survey finds that 42% expect agents to have decision authority by 2030, while 5% report having all the relevant controls today.", "42% is a future intention; 5% is a present control-readiness claim. They are not the same measure or an estimate of unsafe deployments." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation", "Finance, investment & strategy" ], "tags": [ "survey", "governance", "adjacent sectors" ], "why_useful": "Before granting an agent authority, who owns its decision, what limits its actions, and how can a person reverse them?", "source_family": "Publisher survey summary", "publisher_group": "Boston Consulting Group", "kind": "Adoption survey", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-10-14" }, { "id": "WATCH14PM-BX", "title": "Pattern Recognition / AI Adoption Is Accelerating", "publisher": "Blackstone", "url": "https://www.blackstone.com/pattern-recognition/", "published_date": "2026-08-05", "display_date": "2026-08-05", "date_note": "Published 5 August 2026. The 2Q26 survey ran 9–30 June 2026; 2Q25 ran 5–19 June 2025.", "scope": "Cross-sector; no upstream subgroup disclosed", "report_type": "survey commentary", "ai_scope": "Enterprise AI deployment, governance and planned investment", "access": "Original publisher article and relevant method notes read", "sample_method": "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.", "findings": [ "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.", "Deployment stage and planned spending are reported; productivity, cash returns and energy-company adoption are not measured here." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation", "Finance, investment & strategy" ], "tags": [ "survey", "governance", "adjacent sectors" ], "why_useful": "Which portfolio workflows are in use, and what changes in operating or financial results after deployment?", "source_family": "Publisher survey summary", "publisher_group": "Blackstone", "kind": "Adoption survey", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-10-14" }, { "id": "WATCH15PM-FORECAST", "title": "Data-driven production forecast for oil and gas reservoirs - Advantages and limitations", "publisher": "Geoenergy Science and Engineering / Elsevier", "url": "https://www.sciencedirect.com/science/article/pii/S2949891026002137", "published_date": "2026-10", "display_date": "2026-10 (issue month; exact online date unknown)", "date_note": "October 2026 issue; exact online publication date unknown. It cannot be dated after the previous update.", "scope": "Synthetic reservoir benchmark; no operator deployment", "report_type": "technical study", "ai_scope": "Production forecasting and operating-context limits", "access": "Editor independently read publisher-indexed abstract and section text; direct retrieval failed, and no full paper or numerical tables were obtained.", "sample_method": "UNISIM-IV synthetic benchmark; BiLSTM illustration. Full methods and test split not retrieved.", "findings": [ "A UNISIM-IV study finds that production-history-only forecasts miss changes driven by operating events, interactions between wells and constraints." ], "metrics": [], "limitations": "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.", "departments": [ "Reservoir engineering & field development" ], "tags": [ "forecasting", "validation", "operating context" ], "why_useful": "Does the forecast see interventions and constraints, and how does it behave when the operating regime changes?", "source_family": "Publisher technical paper", "publisher_group": "Geoenergy Science and Engineering / Elsevier", "kind": "Technical study", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-10-15", "doi": "10.1016/j.geoen.2026.214569", "authors": [ "Luis Oliveira Pires", "Bahram Lavi", "João Roberto Bertini Junior", "Denis José Schiozer" ] }, { "id": "WATCH15PM-OFGEM", "title": "AI Reg Lab: High-level findings for licensees and stakeholders — February 2026", "publisher": "Ofgem", "url": "https://www.ofgem.gov.uk/sites/default/files/2026-04/AI-Reg-Lab-Februray-2026.pdf", "published_date": "2026-04-29", "display_date": "2026-04-29", "date_note": "Lab held 25 February 2026; summary published 29 April 2026.", "scope": "One anonymized energy-infrastructure pilot; no petroleum operator identified", "report_type": "regulatory lab discussion", "ai_scope": "GenAI asset-data inference and assurance", "access": "Editor read the original four-page PDF on 15 October 2026. © Crown copyright 2026; document reused under the Open Government Licence.", "sample_method": "One company and innovation partner, 25 February 2026 discussion with experts. Anonymized summary, not a representative study.", "findings": [ "A February lab examined a GenAI pilot that structures legacy asset records, fills gaps and builds an asset hierarchy." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "asset data", "provenance", "governance" ], "why_useful": "Can an engineer tell what came from a record, what was inferred, and who accepted the change?", "source_family": "Official regulator discussion summary", "publisher_group": "Ofgem", "kind": "Governance and assurance", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-10-15" }, { "id": "WATCH16PM-IEA", "title": "Key Questions on Energy and AI", "publisher": "International Energy Agency", "url": "https://www.iea.org/reports/key-questions-on-energy-and-ai", "published_date": "2026-04-16", "display_date": "2026-04-16", "date_note": "Published 16 April 2026, verified on the official page. Survey vintage: 2025; forecast horizon: 2030.", "scope": "Global energy and AI; no upstream-only survey subgroup", "report_type": "energy-system outlook and survey synthesis", "ai_scope": "Data-centre electricity demand and AI use in energy operations", "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.", "sample_method": "IEA modelling and datasets; adoption barriers draw on Industry Employment Survey 2025. Respondent count is not stated in the reviewed chapter.", "findings": [ "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation", "Workforce & learning", "Finance, investment & strategy" ], "tags": [ "skills", "data infrastructure", "AI electricity demand" ], "why_useful": "Before buying another model, which missing skills, data connections or equipment would prevent people from using it?", "source_family": "Intergovernmental report", "publisher_group": "IEA", "kind": "System outlook and adoption barriers", "period_group": "Current window", "scope_group": "Cross-industry", "reviewed_at": "2026-10-16", "related_report_ids": [ "INS001" ], "supplemental_urls": [ { "url": "https://iea.blob.core.windows.net/assets/3179f7f8-01f6-4dd6-bffa-c9f7b73f1dc9/KeyQuestionsonEnergyandAI.pdf", "label": "Full report · demand outlook p. 10; adoption barriers and Figure 9.4 pp. 80–85" } ], "source_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" }, { "id": "WATCH17PM-FDP", "title": "Field Development Plan Optimization Under Uncertainty Using Physics-Informed Neural Networks and Generative AI: A Real-Field Case Study in North America", "publisher": "SPE / OnePetro; authors affiliated with OriGen.AI", "url": "https://onepetro.org/SPEADIP/proceedings-abstract/25ADIP/25ADIP/792786", "published_date": "2025-11-03", "display_date": "2025-11-03", "date_note": "Published 3 November 2025, according to the publisher’s indexed metadata. Field-data and test dates are not supplied.", "scope": "One unnamed North American reservoir, 10 completed wells and four reservoir units; model-based development-plan study.", "report_type": "technical modeling study", "ai_scope": "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.", "access": "Publisher-indexed metadata and complete abstract inspected by editor. Direct OnePetro page failed; full paper, figures and validation protocol not reviewed.", "sample_method": "OriGen.AI-affiliated authors describe a generative well-proposal model, physics-informed simulator surrogates and metaheuristic optimization. Publisher abstract only; full validation protocol unavailable.", "findings": [ "An OriGen.AI-authored study models development choices for one unnamed North American reservoir represented by 10 completed wells and four reservoir units.", "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.", "Authors report about 24% higher expected NPV and roughly 3% surrogate error. These are modeled comparisons, not realized production, cash flow or ROI." ], "metrics": [], "limitations": "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.", "departments": [ "Reservoir engineering & field development", "Finance, investment & strategy" ], "tags": [ "field development", "surrogate models", "generative AI", "modeled NPV" ], "why_useful": "Would the preferred plan survive an independent simulator, economic assumptions and engineering review before capital is committed?", "source_family": "SPE technical conference paper", "publisher_group": "SPE / OnePetro", "kind": "Technical study · modeled economics", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-17", "source_locator": "SPE-229333-MS, abstract; publisher metadata: Published November 03 2025; conference 3–6 November 2025", "doi": "10.2118/229333-MS", "authors": [ "R. Rodriguez Torrado", "A. Pumar Jiménez", "M. Shishehbor", "D. Badawi" ], "reviewed_scope": "Publisher-indexed abstract and metadata only; not a field-outcome evaluation." }, { "id": "WATCH18PM-JPT-IOCAAS", "title": "Case Study: Field Deployments of AI-Based IOCaaS Advancing Artificial Lift and Flow Assurance", "publisher": "JPT / SPE; OPX Ai coauthors", "url": "https://jpt.spe.org/case-study-field-deployments-of-ai-based-iocaas-advancing-artificial-lift-and-flow-assurance", "published_date": "2026-04-01", "display_date": "2026-04-01", "date_note": "2026 publication; Kaybob deployment began 2020 and was live by early 2022. Montney calendar dates unspecified.", "scope": "Two named Canadian assets, Kaybob and Montney, already represented by AI-201 and AI-202.", "report_type": "Supplier-authored technical field cases", "ai_scope": "Exception-based surveillance, existing SCADA/historian/ERP integration and bounded lift/control actions.", "access": "Full original publisher text read; not an independent outcome audit.", "sample_method": "Two developer-coauthored case narratives; no controlled study or complete well denominators.", "findings": [ "Kaybob shifted from engineer-approved recommendations to selected adjustments within safety limits.", "Reported production and LOE changes lack full cost and causal validation." ], "metrics": [], "limitations": "OPX Ai founders coauthor. No full costs, controlled counterfactual or independently verified net ROI. Historical field timing is separate from publication. Both cases already counted in the register; indexing adds no deployment or result.", "departments": [ "Production & field operations", "Maintenance & reliability" ], "tags": [ "artificial lift", "flow assurance", "human approval", "SCADA" ], "why_useful": "Shows the integration work and approval progression behind two existing field cases.", "source_family": "Technical field-case account", "publisher_group": "SPE / JPT", "kind": "Developer-coauthored field cases", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-18", "source_locator": "Architecture; Kaybob; Montney; author biographies", "authors": [ "Yogashri Pradhan", "Jai Joon" ], "record_ids": [ "AI-201", "AI-202" ] }, { "id": "WATCH19AM-HAVTIL", "title": "Can artificial intelligence help improve learning from incidents?", "publisher": "Havtil", "publisher_full": "Norwegian Ocean Industry Authority (Havtil)", "url": "https://www.havtil.no/en/explore-technical-subjects2/technical-competence/news/2026/can-artificial-intelligence-help-improve-learning-from-incidents/", "published_date": "2026-01-07", "display_date": "2026-01-07", "date_note": "Workshop: 28 October 2025. Article: 7 January 2026. The tool was developed in 2024; exact development and slide-posting dates unknown.", "scope": "Norwegian petroleum incident-learning workshop; 40 representatives from authorities, operators and suppliers. Not a survey or prevalence estimate.", "report_type": "Regulator workshop account with technical demonstration slides", "ai_scope": "LLM-supported incident-data extraction, shared concepts and searchable contextual data; not automatic high-severity root-cause investigation.", "access": "Original article and 36-page slides read; workflow and investigation-boundary pages rendered locally.", "sample_method": "Qualitative workshop account; attendee concerns are general experience, not a controlled evaluation of the demonstrated DNV tool. Company and participant counts are different.", "findings": [ "Havtil’s account of a 40-person petroleum workshop describes shared incident-learning work and concerns about unreliable or untraceable AI answers.", "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.", "No accuracy score, incident reduction, cost or ROI. Participant concerns are general experience, not a measured evaluation of the demonstrated tool." ], "metrics": [], "limitations": "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). No accuracy benchmark, causal incident-reduction study or independently established net return. The 2024 tool-development year does not establish a deployment inside the rolling window.", "departments": [ "HSE & emissions", "IT, data & enterprise transformation" ], "tags": [ "incident learning", "traceability", "human review", "shared definitions" ], "why_useful": "Can reviewers trace each extracted cause to the original passage and correct a classification?", "source_family": "Regulator workshop and technical demonstration", "publisher_group": "Havtil / DNV", "kind": "Qualitative workshop / demonstration", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-19", "source_locator": "Workshop date/attendance; How AI can help; Quality and trust; Obstacles; The way forward", "companion_sources": [ { "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." } ], "supplemental_urls": [ { "url": "https://www.havtil.no/contentassets/2152e5ff08734a219a61315d4a50464c/havtil-arbeidsmote-20251028-ki-for-laring-av-hendelser.pdf#page=15", "label": "Workshop slides · workflow p. 15; investigation boundary p. 20" } ], "related_reports": [ "INS004" ] }, { "id": "WATCH19EX-MURPHY", "title": "Framework for AI- and ML-Based Predictive Maintenance for Offshore Rotating Equipment", "publisher": "SPE / JPT", "publisher_full": "Society of Petroleum Engineers / Journal of Petroleum Technology", "url": "https://jpt.spe.org/framework-for-ai-and-ml-based-predictive-maintenance-for-offshore-rotating-equipment", "published_date": "2025-01-20", "display_date": "2025-01-20", "date_note": "Article: 20 January 2025. Alert records begin January/March 2023, before this report’s rolling window; exact project start/end dates unknown.", "scope": "Retrospective Murphy Oil predictive-maintenance project on two unnamed deepwater U.S. Gulf platforms; 46 models and 24-month deployment period.", "report_type": "Technical implementation and lessons-learned account", "ai_scope": "Machine-learning anomaly alerts, human disposition/retraining and CMMS maintenance notifications; not autonomous equipment maintenance.", "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.", "sample_method": "Single operator project account. Known-failure replay used in early vendor evaluation; not a blind prospective live-warning test. No controlled comparator or measured review-labor denominator.", "findings": [ "Murphy’s two-platform, 24-month project deployed 46 rotating-equipment models. Missing data delayed the first model until six months after kickoff.", "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.", "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." ], "metrics": [], "limitations": "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. The publication is in-window, but 2023 alert records are older context; current operation is unknown. False-positive improvement is reported qualitatively, not an independently established performance measure.", "departments": [ "Maintenance & reliability" ], "tags": [ "predictive maintenance", "false positives", "data readiness", "human review", "implementation lessons" ], "why_useful": "How many alerts lead to useful work, and what does reviewing the rest cost?", "source_family": "Technical operator project account", "publisher_group": "SPE / JPT", "kind": "Retrospective technical case", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-19", "source_locator": "Methodology; Observations; Discussion: Project Planning and Management, Data Prerequisites, Domain Knowledge, Model Development, Performance, and User Experience; Figs. 4–6", "related_reports": [] }, { "id": "WATCH19PM-METERING", "title": "From Measurement to Commercial Confidence: An Operator’s View of Custody Transfer—and a Responsible Role for AI", "publisher": "SPE / JPT", "publisher_full": "Society of Petroleum Engineers / Journal of Petroleum Technology", "url": "https://jpt.spe.org/from-measurement-to-commercial-confidence-an-operators-view-of-custody-transfer-and-a-responsible-role-for-ai", "published_date": "2026-10-12", "display_date": "2026-10-12", "date_note": "Published 12 October 2026. Illustrative workflows; no named asset, operating period or installed system.", "scope": "Expert editorial on the measurement-to-settlement evidence chain; no sampled installations.", "report_type": "Technical-section editorial", "ai_scope": "Separate rules, ML diagnostics and generative evidence support from authority to change a commercial record.", "access": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "sample_method": "Technical-section expert guidance and illustrative examples, not a measured deployment or independent outcome evaluation.", "findings": [ "SPE’s flow-measurement editorial describes AI support for investigating discrepancies in the quantities used for commercial settlement.", "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.", "No deployment count, recovered revenue or ROI. Measurement uncertainty multiplied by volume and price is not a demonstrated loss." ], "metrics": [], "limitations": "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. No quantified commercial recovery, cost or net return.", "departments": [ "Finance, investment & strategy" ], "tags": [ "measurement", "commercial settlement", "approval", "evidence trail" ], "why_useful": "Can an investigator reproduce the correction and see who approved it?", "source_family": "Technical-section expert editorial", "publisher_group": "SPE / JPT", "kind": "Expert guidance", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-19", "source_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", "related_reports": [] }, { "id": "WATCH19PM-STIMULATION", "title": "A systematic and critical review of artificial intelligence applications for enhancing well stimulation operations and sustainable reservoir management", "publisher": "Discover AI", "publisher_full": "Discover Artificial Intelligence / Springer Nature", "url": "https://link.springer.com/article/10.1007/s44163-026-02455-x", "published_date": "2026-10-15", "display_date": "2026-10-15", "date_note": "Published 15 October 2026. Literature focus: 2020–2025, with earlier foundational studies; much of the underlying work predates the rolling window.", "scope": "Stimulation literature focused on 2020–2025, with older context. Qualitative set of 53 studies and quantitative subset of 23; neither is a deployment population.", "report_type": "Academic literature review", "ai_scope": "Conventional ML, hybrid methods and emerging control approaches; review-author assessments, not verified company maturity.", "access": "Original publisher HTML and relevant passages read by editor 19 October 2026; guidance/review, not independently validated operating outcomes.", "sample_method": "Review describes five databases, two-reviewer screening and a quality rubric (§§2.1–2.3). Descriptive synthesis, not a formal pooled effect. Underlying experiments were not independently re-audited by Sunya.", "findings": [ "A stimulation-AI review highlights weak external testing, small datasets and inconsistent reporting in the published research it examines.", "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.", "The quantitative subset contains 23 studies, not 23 deployments. No independently established project return; broad cost and benefit estimates are not admitted as benchmarks." ], "metrics": [], "limitations": "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. Broad implementation budgets, financial benefits, performance averages and readiness scores are excluded from our numerical comparisons.", "departments": [ "Drilling & completions" ], "tags": [ "stimulation", "external validation", "data leakage", "field transfer" ], "why_useful": "Was the final model tested on a different field, with tuning kept separate from the test?", "source_family": "Academic literature synthesis", "publisher_group": "Springer Nature", "kind": "Literature review", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-19", "source_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", "related_reports": [] }, { "id": "WATCH20UP-SPE-JPT", "title": "Industry Executives Offer Practical Advice on How To Use Modern AI Models Wisely", "publisher": "SPE / JPT", "publisher_full": "Society of Petroleum Engineers / JPT", "url": "https://jpt.spe.org/industry-executives-offer-practical-advice-on-how-to-use-modern-ai-models-wisely", "published_date": "2026-09-10", "display_date": "2026-09-10", "date_note": "September10 publication header; current article includes an undated following-conference Landmark update. Later paragraph is not backdated.", "scope": "SPE content/model licensing experience plus expert comments from Oceaneering and an offshore engineering consultancy; not an operator deployment census.", "report_type": "First-hand conference panel account", "ai_scope": "Technical knowledge retrieval, metadata, model upkeep and content licensing; no field control or CEO authority.", "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.", "sample_method": "Publisher report of practitioner remarks; no measured deployment sample or independently audited commercial results.", "findings": [ "SPE heard more interest in licensing OnePetro knowledge than another standalone AI service; EnRG-LLM remains available.", "Current original identifies Halliburton Landmark as the first content-license customer in an undated follow-up.", "SPE describes variable AI costs, metadata quality and governance as practical implementation concerns." ], "metrics": [], "limitations": "No license value, user population, measured retrieval gain or net ROI. Commercial adjustment is not product termination. October20 classification announcement is a separate source, not evidence that the old model was replaced.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "technical knowledge", "content licensing", "metadata", "AI costs", "product demand" ], "why_useful": "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.", "source_family": "First-hand practitioner panel report", "publisher_group": "SPE / JPT", "kind": "Practitioner account", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-20", "source_locator": "SPE’s AI Journey; licensing-demand discussion; following-conference Landmark paragraph; cost and metadata discussion", "related_reports": [] }, { "id": "WATCH20PM-USEDC", "title": "The Bottleneck Isn’t the Model. It’s the Execution.", "publisher": "AOGR; operator/vendor coauthors", "publisher_full": "The American Oil & Gas Reporter; ROAM-AI and U.S. Energy Development Corp. coauthors", "url": "https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution", "published_date": "2026-07", "display_date": "July 2026", "date_note": "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.", "scope": "12 Permian wells; December 2025 start; six-month author account. A different unnamed 250-plus-well operator is excluded from this named cohort.", "report_type": "Operator/vendor-coauthored field account", "ai_scope": "Physics-aware AI, edge control and approved operating envelopes; no LLM identified.", "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.", "sample_method": "Operator/vendor-coauthored account; production compared with 90-day pre-activation baseline. No matched control or adjustment methodology disclosed.", "findings": [ "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.", "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." ], "metrics": [], "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. No full-cost return definition or independent audit. Does not validate the later October 20 platform launch.", "departments": [ "Production & field operations" ], "tags": [ "ESP", "closed-loop control", "human boundaries", "pre-post baseline", "reported return" ], "why_useful": "Ask for the all-in cost and benefit bridge, the baseline calculation and the log of exceptions, overrides and rejected adjustments.", "source_family": "Operator/vendor-coauthored field account", "publisher_group": "AOGR", "kind": "Field account", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-20", "source_locator": "July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies", "record_ids": [ "AI-296" ], "related_reports": [] }, { "id": "WATCH21AM-MNVR-CASE", "title": "Five Failures Across Five Systems", "publisher": "MNVR AI / Energy Solutions", "publisher_full": "MNVR AI and maintenance partner Energy Solutions", "url": "https://www.mnvrai.ai/s/MNVR-Customer-Story-Energy-Solutions-5-Examples-09222026.pdf", "published_date": null, "display_date": "Undated · checked 21 October 2026", "date_note": "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.", "scope": "Seven gas generators at Muskrat, North Dakota; five selected faults over six weeks. Adjacent gas-to-power context.", "report_type": "Vendor/maintenance-partner customer case", "ai_scope": "Diagnostic agents, operator questions and retained confirmed diagnoses; no automatic control established.", "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.", "sample_method": "Vendor/maintenance-partner case selection, with author-reported inspection confirmation. No full alert population or matched comparison.", "findings": [ "An undated MNVR and Energy Solutions case describes ODIE supporting seven gas generators at the Muskrat microgrid in North Dakota’s Bakken Basin.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Maintenance & reliability" ], "tags": [ "maintenance", "diagnosis", "human confirmation", "estimated avoided costs", "adjacent industry" ], "why_useful": "Can the team show every alert, inspection and resulting work order—including alerts that turned out to be wrong?", "source_family": "Vendor/maintenance-partner customer case", "publisher_group": "MNVR AI", "kind": "Customer case", "period_group": "Date not established", "scope_group": "Adjacent operating analog", "reviewed_at": "2026-10-21", "source_locator": "PDF pp. 1–3: seven generators; five fault cases; estimated costs; six-week period and retained diagnoses", "record_ids": [], "related_reports": [], "supplemental_urls": [ { "url": "https://www.mnvrai.ai/odie", "label": "ODIE product page · human decision boundary" } ] }, { "id": "WATCH21PM-IEA", "title": "Powering Artificial Intelligence: Strategies for integrating data centres into electricity systems", "publisher": "IEA", "publisher_full": "International Energy Agency", "url": "https://www.iea.org/reports/powering-artificial-intelligence", "published_date": "2026-10-14", "display_date": "14 October 2026", "date_note": "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.", "scope": "Abstract: 34 policy jurisdictions covering about 90% of global data-centre capacity. Gas-generation pipeline is global project capacity, not an oil-and-gas AI adoption sample.", "report_type": "Institutional policy and project-pipeline report", "ai_scope": "Electricity-system integration and energy supply for data centres; distinct from AI use inside oil-and-gas operators.", "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.", "sample_method": "Policy inventory and project-pipeline analysis. The report identifies 134 policies/regulations and uses Global Energy Monitor August 2026 data for the gas-generation pipeline; this is not a representative operator deployment or ROI survey.", "findings": [ "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.", "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.", "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." ], "metrics": [ { "value": "~210", "unit": "GW", "claim": "Planned onsite natural-gas generation", "period": "2026 pipeline; Global Energy Monitor data August 2026", "population": "Global project pipeline; all stages", "basis": "IEA report p.60; rounded project-generation capacity", "caveat": "Overlapping total/subsets, not additive figures. Not operating load, gas volumes, binding offtake or verified project completion; no residual operating capacity inferred." }, { "value": "~190", "unit": "GW", "claim": "Capacity at proposal stage", "period": "2026 pipeline; Global Energy Monitor data August 2026", "population": "Subset of the same global project pipeline", "basis": "IEA report p.60; rounded project-generation capacity", "caveat": "Overlapping total/subsets, not additive figures. Not operating load, gas volumes, binding offtake or verified project completion; no residual operating capacity inferred." }, { "value": "17", "unit": "GW", "claim": "Capacity under construction", "period": "2026 pipeline; Global Energy Monitor data August 2026", "population": "Subset of the same global project pipeline", "basis": "IEA report p.60; rounded project-generation capacity", "caveat": "Overlapping total/subsets, not additive figures. Not operating load, gas volumes, binding offtake or verified project completion; no residual operating capacity inferred." } ], "limitations": "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. Do not add these pipeline figures to those in the separate April IEA report or infer gas demand without utilization, conversion efficiency, dispatch and fuel terms.", "departments": [ "Finance, investment & strategy" ], "tags": [ "data centres", "gas demand", "project maturity", "grid access", "onsite generation" ], "why_useful": "Before relying on a project’s gas demand, ask for its customer commitment, generation equipment, permits, grid-access terms and phased load schedule.", "source_family": "Institutional policy and project-pipeline analysis", "publisher_group": "IEA", "kind": "Policy and project-pipeline study", "period_group": "Current window", "scope_group": "Energy and data centres", "reviewed_at": "2026-10-21", "source_locator": "Dated landing page; executive summary; PDF pp.6–10, 36–47 and 58–62; gas-generation pipeline on p.60", "record_ids": [], "related_reports": [ "WATCH16PM-IEA", "INS001" ], "supplemental_urls": [ { "label": "Full report · onsite gas pipeline on page 60", "url": "https://iea.blob.core.windows.net/assets/cd360a30-aaef-43cd-84e6-a125c3a0bc67/PoweringArtificialIntelligence.pdf#page=60" }, { "label": "Executive summary", "url": "https://www.iea.org/reports/powering-artificial-intelligence/executive-summary" } ] }, { "id": "WATCH22FAIL-SPE", "title": "Evaluation of Influential Parameters for Prediction of Flowing Bottom Hole Pressure in Dry Gas Wells Using Machine Learning", "publisher": "SPE", "publisher_full": "Society of Petroleum Engineers / OnePetro", "url": "https://onepetro.org/SPENAIC/proceedings-abstract/26NAIC/26NAIC/SPE-234889-MS/802231", "doi": "10.2118/234889-MS", "authors": [ "Fred Akpososo", "Victor Aimikhe", "David Kalu", "Precious Okolaa" ], "published_date": "2026-08-10", "display_date": "10 August 2026", "date_note": "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.", "scope": "Historical 206-sample ProBHP pressure dataset; number of wells and geography undisclosed.", "report_type": "technical study", "ai_scope": "Random forest versus artificial neural network for pressure prediction; no LLM or field control.", "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.", "sample_method": "Ten inputs; 80/20 sample split; Z-score outlier removal reduced training data. Full split protocol and independence by well not verified.", "findings": [ "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.", "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." ], "metrics": [ { "value": "209.55", "unit": "psig MAE", "claim": "Neural-network mean absolute pressure error", "period": "Historical dataset; test chronology unknown", "population": "Reported test split from 206 samples; exact test count and well independence unknown", "basis": "SPE-234889-MS publisher-indexed abstract; 80/20 split", "caveat": "Author-reported model error, lower is better in this experiment. No raw data/full methods, field deployment or net ROI verified." }, { "value": "93.73", "unit": "psig MAE", "claim": "Random-forest mean absolute pressure error", "period": "Historical dataset; test chronology unknown", "population": "Reported test split from 206 samples; exact test count and well independence unknown", "basis": "SPE-234889-MS publisher-indexed abstract; 80/20 split", "caveat": "Author-reported model error, lower is better in this experiment. No raw data/full methods, field deployment or net ROI verified." } ], "limitations": "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. An 80/20 sample split does not by itself establish testing on new wells. Sample count is not a well count.", "departments": [ "Reservoir engineering & field development" ], "tags": [ "pressure prediction", "neural network", "random forest", "model validation", "negative results" ], "why_useful": "Would performance hold on entirely new wells, with the full error distribution and an operating decision attached?", "source_family": "Publisher-indexed technical abstract; full methods unreviewed", "publisher_group": "SPE / OnePetro", "kind": "Technical study", "period_group": "Current window", "scope_group": "Upstream", "reviewed_at": "2026-10-22", "source_locator": "SPE-234889-MS; publisher-indexed abstract: dataset/split and Random Forest/ANN comparison paragraphs. Published 10 August 2026; conference 3–5 August 2026.", "record_ids": [], "related_reports": [ "WATCH15PM-FORECAST" ], "supplemental_urls": [ { "label": "DOI record · access may be restricted", "url": "https://doi.org/10.2118/234889-MS" } ] }, { "id": "WATCH22ROI-EY", "title": "Technology Adoption and Workforce Readiness in Oil & Gas and Chemicals — EY Future of Energy Survey", "publisher": "EY", "publisher_full": "EY / TRUE Global Intelligence", "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", "published_date": "2025-03", "display_date": "March 2025", "date_note": "Publication day unknown. Fieldwork: 4 November–3 December 2024.", "scope": "Oil, gas and chemicals pooled; Americas, EMEA and APAC.", "report_type": "survey", "ai_scope": "AI expectations; technology scaling.", "access": "Editor read the publisher-hosted PDF, including findings and methodology. Exact item bases and respondent microdata are not supplied.", "sample_method": "Online survey of 508 employees. Technology questions went to directors and above; exact item bases are undisclosed.", "findings": [ "27% of surveyed executives said AI was meeting expectations.", "85% reported difficulty scaling technology pilots. This covers digital technology broadly." ], "metrics": [ { "value": "27", "unit": "% of executives surveyed", "claim": "AI met expectations", "period": "November–December 2024 fieldwork", "population": "Director-and-above subset; exact base unknown", "basis": "Printed p. 7", "caveat": "Self-reported expectations, not cash ROI." }, { "value": "85", "unit": "% of executives surveyed", "claim": "Difficulty scaling technology pilots", "period": "November–December 2024 fieldwork", "population": "Director-and-above subset; exact base unknown", "basis": "Printed p. 8; major and minor challenges combined", "caveat": "Digital technology broadly; not 85% of pilots failing." } ], "limitations": "Pooled sectors; no O&G-only result, project failure rate, costs or measured cash returns. Total sample 508 is not the denominator for these executive-only items. EY sells consulting services.", "departments": [ "Enterprise AI & knowledge" ], "tags": [ "scaling", "expectations", "survey", "workforce" ], "why_useful": "Ask whether the pilot changes everyday work and meets a defined business target.", "source_family": "Original survey", "publisher_group": "EY", "kind": "Survey", "period_group": "Current window", "scope_group": "Oil, gas & chemicals", "reviewed_at": "2026-10-22", "source_locator": "Printed pp. 7–8: expectations and scaling; pp. 27–29: methodology and profile (PDF indices 6–7, 26–28).", "record_ids": [], "related_reports": [ "GAPR001" ], "supplemental_urls": [] }, { "id": "WATCH23AM-SLB-Q3", "title": "SLB third-quarter 2026 results — selected operating and digital highlights", "publisher": "SLB", "publisher_full": "SLB supplier earnings disclosure", "url": "https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f", "published_date": "2026-10-23", "display_date": "23 October 2026", "date_note": "Publication date verified. Customer operating dates are separately unknown.", "scope": "Selected customer disclosures, PDF pp. 14 and 16; not an earnings-call transcript review.", "report_type": "Supplier earnings release", "ai_scope": "Named operating examples, software awards and an anonymous chemistry-control account; AI demand for data centers is excluded.", "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.", "sample_method": "Supplier-selected customer examples; no representative market sample or independent result validation.", "findings": [ "Azule: connected planning and rig execution; BSP: awarded software contract, not verified go-live.", "Eni and ADNOC: updates to existing records, without another deployment count.", "Anonymous chemistry claim remains outside named-company and ROI totals." ], "metrics": [], "limitations": "Case study and earnings release are the same supplier. Different stages, populations and baselines cannot be pooled. Net costs avoided is undefined; no full-cost cash ROI established. No Q3 call transcript reviewed. 24 October economics review: 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.", "departments": [ "Drilling & completions", "Production & field operations", "IT, data & enterprise transformation" ], "tags": [ "supplier evidence", "execution", "contract versus use", "anonymous operator" ], "why_useful": "Compare what the system does with who acts, then inspect population and measurement boundaries.", "source_family": "Supplier earnings release", "publisher_group": "SLB", "kind": "Company disclosure", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_locator": "PDF p. 16: Eni, Azule, ADNOC and BSP Digital highlights; PDF p. 14: production chemistry. HTML release has the same paragraphs.", "record_ids": [ "AI-301", "AI-302", "AI-027", "AI-226" ], "related_reports": [ "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-third-quarter-2026-results", "https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-fourth-quarter-and-full-year-2025-results", "https://investorcenter.slb.com/static-files/cf773c3d-348f-479f-aeb0-81f5418844ca", "https://investorcenter.slb.com/static-files/0835f6dc-d2fb-4bc0-b85e-afecadcf36fe", "https://www.software.slb.com/slb-online-services-terms-and-conditions", "https://www.software.slb.com/slb-online-services-service-specific-terms", "https://georgenoble.substack.com/p/35-of-the-s-and-p-20-of-the-cash", "https://investorcenter.slb.com/static-files/7aa731c6-2f38-47e4-9aa2-473a8d04ec6e", "https://energyai.sunya.ai/assets/slb-data-ai.html" ] }, { "id": "WATCH24PM-WFRD-Q1", "title": "Weatherford Announces First Quarter 2026 Results — PDO ESP analytics", "publisher": "Weatherford", "publisher_full": "Weatherford supplier earnings disclosure", "url": "https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491", "published_date": "2026-04-21", "display_date": "2026-04-21", "date_note": "Publication date verified; exact operating start date unknown. Missed April evidence added in October.", "scope": "Named PDO Oman ESP predictive analytics; asset and pump population unknown.", "report_type": "Supplier earnings release", "ai_scope": "ForeSite integrated with PetroVisor machine learning for failure forecasting and run-life estimation.", "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.", "sample_method": "Supplier-selected deployment description, without operator confirmation, controls or population.", "findings": [ "Weatherford states PDO moved ESP Predictive Analytics from pilot to operating use.", "On-premise deployment meets cybersecurity requirements; no automatic control or maintenance approval boundary disclosed.", "No quantified performance or economics supplied." ], "metrics": [], "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.", "departments": [ "Production & field operations", "Maintenance & reliability" ], "tags": [ "ESP", "named deployment", "on-premise", "supplier evidence" ], "why_useful": "A named move beyond pilot, with deployment architecture and explicit gaps in scale and economics.", "source_family": "Supplier earnings release", "publisher_group": "Weatherford", "kind": "Company disclosure", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_locator": "Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph.", "record_ids": [ "AI-305" ], "related_reports": [] }, { "id": "DEEP24-MARS", "title": "Operational machine learning for remote spectroscopic detection of CH4 point sources", "publisher": "Růžička and coauthors / arXiv", "publisher_full": "Růžička and coauthors / arXiv", "url": "https://arxiv.org/html/2511.07719v2", "published_date": "2026-03-24", "display_date": "2026-03-24", "date_note": "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.", "scope": "Cross-sector satellite methane monitoring; no new named O&G operator.", "report_type": "Technical study", "ai_scope": "Computer vision ensembles for candidate plume detection; human verification retained.", "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.", "sample_method": "Single-model mean across five runs versus ensemble on the same 200-granule, 311-event EMIT test. Counted objects; not percentages or rates.", "findings": [ "MARS researchers found that testing on small image chips hid the false-alarm burden in full satellite scenes. 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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "HSE & emissions" ], "tags": [ "methane", "false alarms", "model validation" ], "why_useful": "Before scaling, how many false alarms and missed events will a team handle on the full live workload?", "source_family": "Original author study", "publisher_group": "Research authors", "kind": "Technical study", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-10-24", "source_locator": "Version 2, 24 March 2026; sections 3.1, 5.1, 6.2 and Table 2; section 6.4 and Table 5.", "record_ids": [], "related_reports": [ "DEEP24-SPRINGER" ] }, { "id": "DEEP24-SPRINGER", "title": "Reducing False Positives in Operational Remote Sensing: Lessons from Methane Plume Detection", "publisher": "Springer Nature / Discovery Science 2026", "publisher_full": "Springer Nature / Discovery Science 2026", "url": "https://doi.org/10.1007/978-3-032-41086-3_46", "published_date": "2026-10-17", "display_date": "2026-10-17", "date_note": "Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass.", "scope": "Atmospheric methane screening; no named petroleum deployment or project stop.", "report_type": "Technical study", "ai_scope": "Computer vision ensembles for candidate plume detection; human verification retained.", "access": "Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass.", "sample_method": "Publisher abstract only; experimental population and full methods not reviewed.", "findings": [ "Authors report competitive end-to-end benchmark performance but more operational false positives than a domain-designed pipeline.", "The abstract favors constraining the task and using domain features before applying machine learning. No numerical comparison is admitted from this abstract." ], "metrics": [], "limitations": "Full chapter unavailable. No measured customer loss, cancelled deployment or industry failure rate. This is a separate study, not a replication of MARS.", "departments": [ "HSE & emissions" ], "tags": [ "methane", "false alarms", "model validation" ], "why_useful": "Before scaling, how many false alarms and missed events will a team handle on the full live workload?", "source_family": "Original author study", "publisher_group": "Research authors", "kind": "Technical study", "period_group": "Current window", "scope_group": "Broader energy", "reviewed_at": "2026-10-24", "source_locator": "Publisher abstract and First Online metadata; chapter pp. 721–736. Full subscription chapter not reviewed.", "record_ids": [], "related_reports": [ "DEEP24-MARS" ] }, { "id": "DEEP24-HAVTIL", "title": "Artificial intelligence (AI) in the petroleum industry is in need of better oversight and management", "publisher": "Norwegian Ocean Industry Authority (Havtil)", "publisher_full": "Norwegian Ocean Industry Authority (Havtil)", "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/", "published_date": "2026-10-23", "display_date": "2026-10-23", "date_note": "Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page.", "scope": "Norwegian petroleum operating companies; identities and response count undisclosed.", "report_type": "Regulator survey summary", "ai_scope": "Inventory and oversight of AI in planning, maintenance and operational decision support.", "access": "Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page.", "sample_method": "Regulator summary of company responses; questionnaire, dates, response rate and item bases absent from reviewed page.", "findings": [ "Norway’s petroleum regulator says operators vary widely in how they oversee AI. Many lack a complete inventory, including AI inside supplier products.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "operationalization", "oversight" ], "why_useful": "Can the operating team name every AI system it relies on, including supplier features, and who owns its review?", "source_family": "Original regulator summary", "publisher_group": "Havtil", "kind": "Regulator report", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_locator": "Varying levels of maturity; An overview is essential for effective risk management; Prudent development and use.", "record_ids": [], "related_reports": [] }, { "id": "SOL5-WESTBRAE", "title": "Well Intelligence – Optimised Production", "publisher": "Net Zero Technology Centre", "publisher_full": "Net Zero Technology Centre", "url": "https://www.netzerotc.com/case-studies/well-intelligence-optimised-production/", "published_date": null, "display_date": "Unknown", "date_note": "Undated historical account. Exact trial, publication and end dates unknown; not a new deployment in the rolling 24-month window.", "scope": "West Brae; NZTC describes eight production wells and roughly 8,000 bbl/day. These are historical scope figures.", "report_type": "Trial case", "ai_scope": "Machine learning uses historical production data to recommend valve movements.", "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.", "sample_method": "One field; five live usable test points and 12 historical points, not 17 independent projects. No confirmed realized uplift.", "findings": [ "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.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Production & field operations" ], "tags": [ "asset fit", "trial", "value verification" ], "why_useful": "Is there enough improvement left in the chosen asset to measure and pay for this tool?", "source_family": "Original project assessment and supplier case", "publisher_group": "Net Zero Technology Centre", "kind": "Field account", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_locator": "Project overview; Industry value; Lessons learned. Undated historical account naming Marathon and West Brae.", "record_ids": [], "related_reports": [ "SOL5-WELLINT-SUPPLIER" ] }, { "id": "SOL5-WELLINT-SUPPLIER", "title": "Case Study: Enhancing Production with Well Intelligence AI Model", "publisher": "Intelligent Plant", "publisher_full": "Intelligent Plant", "url": "https://www.intelligentplant.com/case-study-well-intelligence", "published_date": null, "display_date": "Unknown", "date_note": "Undated historical account. Exact trial, publication and end dates unknown; not a new deployment in the rolling 24-month window.", "scope": "West Brae; NZTC describes eight production wells and roughly 8,000 bbl/day. These are historical scope figures.", "report_type": "Trial case", "ai_scope": "Machine learning uses historical production data to recommend valve movements.", "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.", "sample_method": "Undated selected supplier case; anonymous operator. A possible retelling, not another independent trial.", "findings": [ "Supplier describes an anonymous four-week trial around 8,000 bbl/day and claims valve discovery, fewer shutdowns and improved asset economics.", "The account shares distinctive details with NZTC’s West Brae review, but identity and dates are unconfirmed. No quantified benefits, full costs or independent operator confirmation." ], "metrics": [], "limitations": "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.", "departments": [ "Production & field operations" ], "tags": [ "asset fit", "trial", "value verification" ], "why_useful": "Is there enough improvement left in the chosen asset to measure and pay for this tool?", "source_family": "Original project assessment and supplier case", "publisher_group": "Intelligent Plant", "kind": "Field account", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_locator": "Challenge and Testing; Historical Data Review; Results and Impact. Anonymous client, no publication date.", "record_ids": [], "related_reports": [ "SOL5-WESTBRAE" ] }, { "id": "SOL5-WELLRECORDS", "title": "Information extraction from historical well records using a large language model", "publisher": "Ma, Santos, Lackey, Viswanathan and O’Malley / Scientific Reports", "publisher_full": "Ma, Santos, Lackey, Viswanathan and O’Malley / Scientific Reports", "url": "https://www.nature.com/articles/s41598-024-81846-5", "published_date": "2024-12-30", "display_date": "2024-12-30", "date_note": "Published 30 December 2024. Historical documents and 2024-era models; no prospective deployment date.", "scope": "150 Colorado completion reports and 10 Pennsylvania records; research corpus, not 160 operating wells.", "report_type": "Technical study", "ai_scope": "OCR/text conversion plus Llama-family extraction of location and true vertical depth; prompt variations compared.", "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.", "sample_method": "Accuracy accepts location within 10 m and depth within 10 ft. Same-corpus prompt comparison, not independently validated production reliability. 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.", "findings": [ "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.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "model validation", "deployment limits", "source quality" ], "why_useful": "Does the test check the right depth, coordinates, units and source quality before values enter an operating record?", "source_family": "Original peer-reviewed technical study", "publisher_group": "Ma, Santos, Lackey, Viswanathan and O’Malley / Scientific Reports", "kind": "Technical study", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_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.", "record_ids": [], "related_reports": [] }, { "id": "SOL5-RAG", "title": "Evaluating Chunking Strategies For Retrieval-Augmented Generation in Oil and Gas Enterprise Documents", "publisher": "Taiwo and Yusoff / CS&IT proceedings author manuscript", "publisher_full": "Taiwo and Yusoff / CS&IT proceedings author manuscript", "url": "https://arxiv.org/pdf/2603.24556", "published_date": "2026-03-25", "display_date": "2026-03-25", "date_note": "Author manuscript posted 25 March 2026; proceedings marked 2026, exact issue day unknown.", "scope": "60 documents and 60 queries;20 P&ID files, not an operator or asset population.", "report_type": "Technical study", "ai_scope": "Azure document extraction; OpenAI text-embedding-3-large; Azure CosmosDB vector retrieval. Answer generation was not evaluated.", "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.", "sample_method": "Four separately indexed methods, pooled top 10 results, human relevance grades 0/1/2. Corpus and judgments not independently replicated. 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.", "findings": [ "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.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "model validation", "deployment limits", "source quality" ], "why_useful": "Can the system retrieve the right information from drawings and tables, as well as prose, before anyone relies on its answers?", "source_family": "Original technical proceedings author manuscript", "publisher_group": "Taiwo and Yusoff / CS&IT proceedings author manuscript", "kind": "Technical study", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-24", "source_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.", "record_ids": [], "related_reports": [] }, { "id": "WATCH25AM-PTTEP-XBOT", "title": "PTTEP 56-1 One Report 2025 — Procurement X.Bot", "publisher": "PTTEP", "publisher_full": "PTTEP", "url": "https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025", "published_date": null, "display_date": "Unknown", "date_note": "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.", "scope": "Procurement process; geography, business-unit and supplier coverage unknown.", "report_type": "Operator annual report", "ai_scope": "AI plus automation; no LLM, autonomous purchasing or X.brain integration established.", "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.", "sample_method": "Average per-PO time and six-month backlog change; baseline values and cohort unknown.", "findings": [ "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.", "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." ], "metrics": [], "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.", "departments": [ "Supply chain & logistics" ], "tags": [ "Procurement AI and automation", "source limits", "process evidence" ], "why_useful": "Which steps changed, who approves orders, and do the time gains persist after all operating costs?", "source_family": "Operator annual report", "publisher_group": "PTTEP", "kind": "Company disclosure", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "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.", "record_ids": [ "AI-306" ], "related_reports": [] }, { "id": "WATCH25AM-BKV-RUNWAY", "title": "BKV 2025 Sustainability Report — Project Runway", "publisher": "BKV Corporation", "publisher_full": "BKV Corporation", "url": "https://www.bkv.com/2025-sustainability-report-governance", "published_date": null, "display_date": "Unknown", "date_note": "Go-live 19 November 2025. Report letter is dated September 2026; exact report publication day unknown.", "scope": "Barnett Shale; broader portfolio and acquisition populations differ.", "report_type": "Operator annual report", "ai_scope": "Conventional operating-data infrastructure; not AI adoption.", "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.", "sample_method": "Four legacy environments; timings are reported process comparisons.", "findings": [ "BKV consolidated four inherited SCADA/CygNet environments. Its Barnett platform went live 19 November 2025.", "Field staff reviewed terminology and screens. Peer-led training and field ride-alongs supported the change.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "Operating-system transformation", "source limits", "process evidence" ], "why_useful": "Does the common data model survive the next acquisition, and how much ongoing support does it need?", "source_family": "Operator annual report", "publisher_group": "BKV Corporation", "kind": "Company disclosure", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Governance: Project Runway; Informed by field operations; Collaboration beyond implementation.", "record_ids": [], "related_reports": [] }, { "id": "WATCH25AM-HARBOUR-COUPA", "title": "Harbour Energy Coupa Supplier Portal FAQ, version 4", "publisher": "Harbour Energy", "publisher_full": "Harbour Energy", "url": "https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf", "published_date": "2025-06", "display_date": "2025-06", "date_note": "Guide version 4, June 2025; exact day, system start and entity rollout unknown.", "scope": "Affected supplier-channel scope; no asset, user or invoice denominator.", "report_type": "Operator supplier guide", "ai_scope": "Supplier process digitization; no AI established.", "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.", "sample_method": "Published process instructions, not before/after evaluation.", "findings": [ "Harbour’s June 2025 Coupa guide covers purchase orders, confirmations, invoices and status tracking.", "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.", "No measured time, error, cash saving or ROI result. Free supplier access does not establish the operator’s total cost." ], "metrics": [], "limitations": "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.", "departments": [ "Supply chain & logistics" ], "tags": [ "Procure-to-pay foundation", "source limits", "process evidence" ], "why_useful": "Where do invoices stall, which exceptions require people, and which supplier channel applies?", "source_family": "Operator supplier guide", "publisher_group": "Harbour Energy", "kind": "Company disclosure", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_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.", "record_ids": [], "related_reports": [] }, { "id": "WATCH25AM-COP-GEP", "title": "ConocoPhillips GEP SMART Introduction", "publisher": "ConocoPhillips", "publisher_full": "ConocoPhillips", "url": "https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf", "published_date": "2024-12", "display_date": "2024-12", "date_note": "Updated December 2024; January 2025 transition target. Undated live guidance still uses that future-tense wording; completion by business unit unknown.", "scope": "Supplier process; geography and business-unit rollout populations unknown.", "report_type": "Operator supplier guide", "ai_scope": "Digital procurement and invoice handling; no AI disclosed.", "access": "Editor read original 7-page issuer PDF. January 2025 is a target, not verified unit-by-unit completion. Exact revision day unknown.", "sample_method": "Process diagram and target, not measured operating comparison.", "findings": [ "ConocoPhillips’ December 2024 guide chose GEP SMART for supplier registration, PO acknowledgement and invoicing.", "Suppliers register, accept PO terms, create invoices or service confirmations and inspect remittance. Current guidance offers portal, integration and Enverus submission routes.", "Faster approvals and better order accuracy are intended benefits; no measured result or ROI." ], "metrics": [], "limitations": "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.", "departments": [ "Supply chain & logistics" ], "tags": [ "Procure-to-pay foundation", "source limits", "process evidence" ], "why_useful": "Which units and invoice routes are live, and who owns the exceptions?", "source_family": "Operator supplier guide", "publisher_group": "ConocoPhillips", "kind": "Company disclosure", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "PDF p.1 updated December 2024; p.3 transition target; p.5 intended benefits; p.6 supplier invoice flow.", "record_ids": [], "related_reports": [] }, { "id": "W25PM-ECO-AIOPS", "title": "Ecopetrol 2025 Integrated Management Report — digital transformation and AIOps", "publisher": "Ecopetrol", "publisher_full": "Ecopetrol", "url": "https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf", "published_date": "2026", "display_date": "2026 (exact publication day unknown)", "date_note": "The report covers 2025; exact publication day and AIOps start/evaluation dates are unknown. The 2027 target is prospective.", "scope": "Critical IT services; business-unit, service, user and asset populations unspecified.", "report_type": "Operator annual report", "ai_scope": "AIOps is explicitly AI for IT operations; the report does not identify models, vendors or an LLM. Self-service and infrastructure upgrades are separate.", "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.", "sample_method": "99.52% is an availability level, not a 99.52% improvement. Baseline and measurement window unknown.", "findings": [ "Ecopetrol says using AI in IT operations raised critical-service availability to 99.52% and generated more than US$1.5 million in savings.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "IT reliability and AI rollout limits", "evidence limits" ], "why_useful": "Which services improved, what changed against the prior baseline, and how much of the saving remains after running costs?", "source_family": "Operator annual report", "publisher_group": "Ecopetrol", "kind": "Operator annual report", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Printed p.314 / physical PDF p.315 of 482, Transformación Digital. AIOps, self-service and AI maturity are separate passages.", "record_ids": [ "AI-147" ], "related_reports": [] }, { "id": "W25PM-YPF-EKONS", "title": "YPF refinery blending — interview with BCG project lead Roberto Rodrigo", "publisher": "Más Energía / LM Neuquén; Martina Mongelluzzo", "publisher_full": "Más Energía / LM Neuquén; Martina Mongelluzzo", "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", "published_date": null, "display_date": "Unknown", "date_note": "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.", "scope": "La Plata refinery, Argentina; one implementation described.", "report_type": "Original supplier interview", "ai_scope": "AI plus process simulation; no LLM disclosed.", "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.", "sample_method": "No measured before/after cohort; up to 50% is not an observed average.", "findings": [ "A Boston Consulting Group (BCG) project lead describes Ekons Matrix AI in use at YPF’s La Plata refinery in Argentina.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Refining, LNG & process operations" ], "tags": [ "Refinery decision support", "evidence limits" ], "why_useful": "Can YPF show achieved blend-time and quality results over enough batches to separate the tool from other operating changes?", "source_family": "Original supplier interview", "publisher_group": "Más Energía / LM Neuquén; Martina Mongelluzzo", "kind": "Original supplier interview", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Sections Del laboratorio a la operación en tiempo real and Argentina, el primer caso de implementación.", "record_ids": [], "related_reports": [] }, { "id": "W25PM-DIXSTONE-IFS", "title": "Dixstone selects IFS Cloud for offshore services", "publisher": "IFS", "publisher_full": "IFS", "url": "https://www.ifs.com/pl/insights/news/dixstone-selects-ifs-cloud-to-power-global-offshore-operations", "published_date": "2025-12-09", "display_date": "2025-12-09", "date_note": "Selection announcement dated 9 December 2025. Go-live and implementation milestones are not disclosed.", "scope": "Offshore-services rigs, vessels and platforms; individual rollout assets unspecified.", "report_type": "Supplier announcement", "ai_scope": "Industrial AI is supplier positioning; no specific deployed AI function is established.", "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.", "sample_method": "Selection and intended scope, without a measured operating comparison.", "findings": [ "IFS says Dixstone selected IFS Cloud to manage project finances and fleet maintenance.", "The intended system brings budgets, schedules, resources and asset data together. The release proposes predictive maintenance for rigs, vessels and platforms.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "Offshore-services software foundation", "evidence limits" ], "why_useful": "Which assets and workflows are live, what still needs implementation, and where does AI actually change the work?", "source_family": "Supplier announcement", "publisher_group": "IFS", "kind": "Supplier announcement", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Dated release: selection, project controls and fleet maintenance paragraphs; About Dixstone separates affiliate status from Perenco E&P.", "record_ids": [], "related_reports": [] }, { "id": "RESULTS-PERENCO", "title": "Perenco enhances drilling efficiency with advanced automation solutions", "publisher": "SLB", "publisher_full": "SLB", "url": "https://www.slb.com/resource-library/case-study-with-navigation/di/perenco-enhances-drilling-efficiency-with-advanced-automation-solutions", "published_date": null, "display_date": "Unknown", "date_note": "2025 three-well campaign. Exact publication day and campaign dates unknown.", "scope": "Offshore Republic of the Congo; individual assets unnamed.", "report_type": "Supplier customer case study", "ai_scope": "AI-driven drilling automation; no LLM identified.", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "sample_method": "Manual operations are discussed; raw ROP baseline and comparison matching unknown.", "findings": [ "SLB reports a 19% increase in rate of penetration in Perenco’s 2025 offshore Congo campaign.", "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.", "Three wells; 97% of on-bottom operations automated. These are speed and control measures, not cash return." ], "metrics": [], "limitations": "Raw drilling-speed baseline, matched offset population, complete costs and AI-only contribution unknown.", "departments": [ "Drilling & completions" ], "tags": [ "Operating result", "evidence limits" ], "why_useful": "Did the gain hold on more wells after accounting for geology, equipment and full deployment costs?", "source_family": "Supplier customer case study", "publisher_group": "SLB", "kind": "Supplier customer case study", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Summary and Results; three-well 2025 campaign.", "record_ids": [], "related_reports": [] }, { "id": "RESULTS-WAFI", "title": "From manual accounts payable to AI-powered finance in 12 weeks", "publisher": "SpendConsole", "publisher_full": "SpendConsole", "url": "https://spendconsole.ai/success-stories/from-manual-accounts-payable-to-ai-powered-finance-in-12-weeks/", "published_date": null, "display_date": "Unknown", "date_note": "Publication and operating calendar dates unknown; results observed within weeks of go-live. Twelve weeks describes implementation, not the evaluation period.", "scope": "Wafi Energy Pakistan finance function.", "report_type": "Supplier customer case study", "ai_scope": "AI document processing plus rules and process changes; effects not isolated.", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "sample_method": "Approximately five minutes of manual handling versus under one in automated scenarios; not every invoice.", "findings": [ "SpendConsole reports invoice processing falling from about five minutes to under one in automated scenarios at Wafi Energy Pakistan.", "AI document capture, rules and SAP workflows handle routine invoices; finance staff handle exceptions. Autonomous payment approval is not established.", "The vendor separately reports 50–70% productivity improvement and full go-live in about 12 weeks." ], "metrics": [], "limitations": "Invoice volume, automated-scenario share, precise measurement window and full costs unknown. This is the Pakistan Shell-brand licensee, not Shell-wide adoption.", "departments": [ "Finance & accounting" ], "tags": [ "Operating result", "evidence limits" ], "why_useful": "What share of invoices qualifies, and how much staff time is actually released after exception handling?", "source_family": "Supplier customer case study", "publisher_group": "SpendConsole", "kind": "Supplier customer case study", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Client profile; solution; results; automated-scenario qualification.", "record_ids": [], "related_reports": [] }, { "id": "RESULTS-APOLLO", "title": "HSV-01 chemical treatment case study", "publisher": "Apollo Resources", "publisher_full": "Apollo Resources", "url": "https://apolloresourcesllc.net/apollo-resources-hsv-01-oil-treating-chemistry-effectively-eliminates-30000-ppm-of-h2s-in-senm/", "published_date": null, "display_date": "Unknown", "date_note": "Thirty-day observation; publication and operating calendar dates unknown.", "scope": "Unnamed facility, southeastern New Mexico.", "report_type": "Supplier customer case study", "ai_scope": "No AI identified; useful comparison for the same operating problem.", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "sample_method": "Prior treatment versus redesigned treatment; comparable throughput and pre-period observation length unknown.", "findings": [ "Apollo reports 65% lower chemical consumption and $210,000 lower monthly H2S treatment cost at an unnamed southeastern New Mexico facility.", "A flow-meter-linked controller, new chemistry and a relocated injection point changed treatment. Key operator personnel received online write access.", "Zero out-of-spec H2S events during a 30-day observation. The monthly cost claim is not annualized here." ], "metrics": [], "limitations": "Population, throughput, cost calculation, installation costs and longer-term durability unknown. New chemistry and equipment prevent attributing the result to software alone.", "departments": [ "Production & field operations" ], "tags": [ "Non-AI comparison", "evidence limits" ], "why_useful": "How much could treatment redesign achieve before adding AI, and what does it cost to sustain?", "source_family": "Supplier customer case study", "publisher_group": "Apollo Resources", "kind": "Supplier customer case study", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Solution and Results; 30-day observation.", "record_ids": [], "related_reports": [] }, { "id": "RESULTS-WEATHERFORD", "title": "Maximizing chemical efficiency with rigless precision-based capillary injection", "publisher": "Weatherford / SPE", "publisher_full": "Weatherford / SPE", "url": "https://www.weatherford.com/documents/papers/2025/maximizing-chemical-efficiency-and-reducing-environmental-footprint-with-rigless,-precision-based-ca/", "published_date": "2025-09-16", "display_date": "2025-09-16", "date_note": "Paper dated 16 September 2025. Trial dates and evaluation duration unknown.", "scope": "Unnamed field trials; geography unspecified.", "report_type": "Supplier customer case study", "ai_scope": "No AI identified; precision placement and engineering comparison.", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "sample_method": "Traditional cyclic treatment; raw volumes and cost baseline not supplied.", "findings": [ "A Weatherford-hosted technical-paper abstract reports 65% less chemical use per well and 52% lower operating expense in capillary-injection field trials.", "Simulation-guided placement, a capillary string and an injection valve deliver chemicals to the treatment point. No AI is identified.", "Results are from the public abstract of SPE-227643-MS; the full paper was not reviewed." ], "metrics": [], "limitations": "Operator, geography, well population, raw baseline, cost perimeter and installation costs unknown. This does not establish all-in ROI or scaled deployment.", "departments": [ "Production & field operations" ], "tags": [ "Non-AI comparison", "evidence limits" ], "why_useful": "Does targeted placement outperform surface optimization after installation and maintenance costs?", "source_family": "Supplier customer case study", "publisher_group": "Weatherford / SPE", "kind": "Supplier customer case study", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Public abstract, SPE-227643-MS; field-trial results.", "record_ids": [], "related_reports": [] }, { "id": "W23AM-CHEMISTRY", "title": "Permian field chemical automation case study", "publisher": "SLB", "publisher_full": "SLB", "url": "https://www.slb.com/resource-library/case-study-with-navigation/production-chemicals/permian-basin-field-automations-cs", "published_date": null, "display_date": "Unknown", "date_note": "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.", "scope": "Nearly 1,400 producing Permian wells; operator and individual assets unnamed.", "report_type": "Supplier earnings release / customer account", "ai_scope": "AI injection control plus ML anomaly detection; human review cadence disclosed, execution permissions and fallback unspecified.", "access": "Editor read the relevant original publisher page on 25 October 2026. Results remain source-reported; no independent outcome validation.", "sample_method": "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.", "findings": [ "SLB reports overtreatment falling from 17% to 3% across nearly 1,400 Permian wells over a 20-month program.", "Sensors feed ML detection; weekly team reviews support remote adjustments. Detailed approval and fallback permissions are not disclosed.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Production & field operations" ], "tags": [ "Chemical dosing", "evidence limits" ], "why_useful": "What costs were included, and did lower overtreatment persist without compromising corrosion or scale protection?", "source_family": "Supplier earnings release / customer account", "publisher_group": "SLB", "kind": "Supplier earnings release / customer account", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-25", "source_locator": "Results and solution sections; 20-month dosing comparison and first-nine-month savings are different periods.", "record_ids": [], "related_reports": [] }, { "id": "D26-TOPTAL", "title": "Mineral-title evaluation and regression testing", "publisher": "Toptal", "publisher_full": "Toptal", "url": "https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy", "published_date": null, "display_date": "Unknown", "date_note": "Publication, operating and evaluation dates unknown.", "scope": "US land records; anonymous technology provider, no named operator.", "report_type": "Technical evaluation", "ai_scope": "LLM title determination plus legal rules, SQL and knowledge-graph concepts.", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "sample_method": "Approximately 27% versus 2% false-positive rates; same-cohort comparison and denominator unspecified.", "findings": [ "Toptal reports false positives falling from about 27% to 2%, with recall about 93%; conveyance-party capture rose from about 85% to 96%.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Land & legal" ], "tags": [ "Technical test", "evidence limits" ], "why_useful": "How does the error rate hold on unseen deeds, and who signs off on ownership conclusions?", "source_family": "Technical evaluation", "publisher_group": "Toptal", "kind": "Technical evaluation", "period_group": "Date not established", "scope_group": "Oil & gas", "reviewed_at": "2026-10-26", "source_locator": "Challenge; Ground-truth AI Evaluation; Outcome. No publication/evaluation date displayed.", "record_ids": [], "related_reports": [] }, { "id": "D26-SLB-DDR", "title": "AI approach for drilling operation insights — US 20260063025 A 1", "publisher": "Schlumberger Technology Corporation / published US patent application", "publisher_full": "Schlumberger Technology Corporation / published US patent application", "url": "https://patents.google.com/patent/US20260063025A1/en", "published_date": "2026-03-05", "display_date": "2026-03-05", "date_note": "Application published 5 March 2026; exact underlying evaluation dates unknown.", "scope": "Drilling-report examples; no named customer or live asset tied to the 89%/44% row.", "report_type": "Technical evaluation", "ai_scope": "Classifies risks in daily drilling remarks using language models; engineering review and operating permissions not established.", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "sample_method": "Different metrics for one comparator, not a before/after improvement.", "findings": [ "An SLB patent application reports 89% overall accuracy but 44% equal-class recall for its oil-and-gas GPT-2 comparator.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Drilling & completions" ], "tags": [ "Technical test", "evidence limits" ], "why_useful": "Does the evaluation give rare, costly events enough weight?", "source_family": "Technical evaluation", "publisher_group": "Schlumberger Technology Corporation / published US patent application", "kind": "Technical evaluation", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-26", "source_locator": "Table 15 and section 5, Model Performance Comparative Analysis; published application, not a granted-patent or commercial-deployment claim.", "record_ids": [], "related_reports": [] }, { "id": "D26-SNAM", "title": "Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants", "publisher": "Snam / Engineering Group", "publisher_full": "Snam / Engineering Group", "url": "https://arxiv.org/html/2511.08609v1", "published_date": "2025-11-05", "display_date": "2025-11-05", "date_note": "Preprint 5 November 2025; evaluation and operating calendar dates unknown.", "scope": "Snam gas-plant documentation; Italy; approximately 1,500 ReMi in the provided collection.", "report_type": "Technical evaluation", "ai_scope": "OCR, vision/language models, scene-graph reasoning and constrained optimization support MGM users.", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "sample_method": "Table 2 relationship recall@100=.424; Table 3 final-task measures are different quantities.", "findings": [ "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.", "The authors say complex systems remain difficult, although simpler systems work better. This constrains claims about automatic plant digitization.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "IT, data & enterprise transformation" ], "tags": [ "Technical test", "evidence limits" ], "why_useful": "Which drawings still need manual reconstruction, and how long does correction take?", "source_family": "Technical evaluation", "publisher_group": "Snam / Engineering Group", "kind": "Technical evaluation", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-26", "source_locator": "Sections 3.1,4(Tables 1–3) and 5; top 100 relationship recall and complex-system limitation.", "record_ids": [ "AI-208" ], "related_reports": [] }, { "id": "D26-TOTAL", "title": "Honeywell AI pilot aids coker unit operations at TotalEnergies refinery", "publisher": "Control Global / Honeywell Users Group coverage", "publisher_full": "Control Global / Honeywell Users Group coverage", "url": "https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery", "published_date": "2026-06-11", "display_date": "2026-06-11", "date_note": "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.", "scope": "Port Arthur refinery delayed coker, Texas; alarm guidance separately covered more than 300 tags.", "report_type": "Source-reviewed project account", "ai_scope": "Advisory alerts and operator decisions; no autonomous process write authority established.", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "sample_method": "Different use cases have different populations; expected benefits are not measured achieved results.", "findings": [ "A named TotalEnergies engineer said a 75%-accuracy drum-level model was not retained because switching also depended on conditions outside its scope.", "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.", "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." ], "metrics": [], "limitations": "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.", "departments": [ "Refining, LNG & process operations" ], "tags": [ "Use case not retained", "evidence limits" ], "why_useful": "Did the prediction answer the actual operating decision?", "source_family": "Source-reviewed project account", "publisher_group": "Control Global / Honeywell Users Group coverage", "kind": "Source-reviewed project account", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-26", "source_locator": "Steam perturbance, drum cycling, heater spalling and alarm-decision-support sections; Thomas Klein remarks.", "record_ids": [ "AI-207" ], "related_reports": [] }, { "id": "D26-ESP-PAPER", "title": "ESP failure prediction — SPE-221571-MS", "publisher": "Baker Hughes authors / SPE", "publisher_full": "Baker Hughes authors / SPE", "url": "https://dam.bakerhughes.com/asset/2145a967-01f1-4ffa-8ee1-37b085e6fff1/spe-221571-ms-bkr-licensed.pdf", "published_date": null, "display_date": "2024-10-29/30", "date_note": "Conference 29–30 October 2024; historical records; exact publication day unknown.", "scope": "137 pumps at three unnamed Americas assets; failed-test subset inferred as 18.", "report_type": "Supplier-authored conference paper", "ai_scope": "ML/reliability/engineering ensemble; people choose interventions.", "access": "Editor read the original specified passage on 26 October 2026. Publisher/author-reported evidence; outcomes not independently validated.", "sample_method": "83 failed pumps trained; remaining pumps tested; source-specific event definitions.", "findings": [ "50% sensitivity; nine missed failures; daily deployment subsequently reported." ], "metrics": [], "limitations": "No prospective controlled benefit or complete net ROI. Test event definitions depend on intervening maintenance actions.", "departments": [ "Maintenance & reliability" ], "tags": [ "Technical test", "retrospective" ], "why_useful": "Ask whether the installed sensor resolution captures costly electrical failure modes.", "source_family": "Technical paper", "publisher_group": "Baker Hughes authors / SPE", "kind": "Technical paper", "period_group": "Current window", "scope_group": "Oil & gas", "reviewed_at": "2026-10-26", "source_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.", "record_ids": [], "related_reports": [] } ], "thesis": { "title": "Look at the work, and the market starts to make more sense.", "text": "Some companies use AI to run equipment and improve field operations. Others are trying language models for reading, writing and coordinating work. Both depend on good data, usable software and people who can put it to work. The reports help explain where each approach might be useful and what it still needs to prove." }, "start_here": [ { "id": "CON001", "lens": "VALUE POOLS", "reason": "McKinsey: the upstream model and its assumptions." }, { "id": "CON005", "lens": "BEYOND UPSTREAM", "reason": "BCG: refining and retail, with distinct margin baselines." }, { "id": "ADP002", "lens": "OPERATING REALITY", "reason": "IBM: a 2026 upstream survey with a disclosed sample." }, { "id": "ADP003", "lens": "BUYERS & VENDORS", "reason": "EY-Parthenon: digital spending and operator value." } ], "insights": [ { "title": "AI for equipment is further along than generative AI.", "observation": "EY’s upstream buyer research describes traditional AI as the largest current AI spending category, with generative and agentic AI at an earlier stage. Separately, IBM’s 2025 survey reports AI use in exploration among 44% of 105 respondents, without separating AI types. McKinsey’s May 2025 survey of 164 oilfield-services executives reports only 1% at significant GenAI scale. Different technologies, questions and samples explain why these figures can coexist.", "implication": "Treat improving field operations and introducing language tools as different jobs. They may need different buyers, connections to existing systems and tests. Limited use of generative AI does not mean companies have done little with other kinds of AI.", "report_ids": [ "ADP001", "ADP002", "ADP003", "CON002", "INS002" ], "company_ids": [ "AI-011", "AI-008" ], "company_connection": "ADNOC / SLB’s production optimization and SLB’s Tela catalog illustrate different product and operating boundaries. Their dossiers preserve what is deployed versus offered.", "counterpoint": "None of these surveys is a harmonized industry census. IBM’s measures are self-reported, while the Dallas Fed’s regional survey is historical context from June 2024. Do not draw a time series from them." }, { "title": "There is more to this market than upstream.", "observation": "BCG explicitly models refining and fuel retail as well as upstream. Its illustrative 2025-to-2030 EBIT bridges imply about 30% upstream, 50% refining and 70% fuel-retail uplift—but against different margin bases and different physical units. McKinsey’s $230 billion model covers upstream.", "implication": "Look at refinery maintenance, process settings, energy use and commercial work as well. Compare the dollars an asset might gain, what deployment costs and how likely the benefit is to last. The biggest percentage is not always the biggest opportunity.", "report_ids": [ "CON005", "CON001" ], "company_ids": [ "AI-060", "AI-140", "AI-181" ], "company_connection": "BP’s refining program, Shell’s equipment-health AI at Pearl GTL, and BPCL’s refinery alerts and video analytics show activity beyond the upstream exhibit. These examples do not validate BCG’s percentage estimates.", "counterpoint": "BCG’s bridge is illustrative, not measured sector performance. Refining and retail have different starting margins, and the reports provide no like-for-like absolute cross-segment ranking." }, { "title": "What an operator gains is different from what a supplier can sell.", "observation": "EY-Parthenon distinguishes a projected upstream digital-solutions market of $38 billion in 2035 from more than $790 billion of cumulative operator value over 2026–35. Both cover digital solutions broadly. McKinsey’s $230 billion is annual AI value at full potential, net of modeled implementation costs. These are different economic quantities.", "implication": "A sector-wide benefit is not a software sales forecast. For a supplier, ask what customers will actually pay for, when they can buy and how much support they will need. For an operator, ask what becomes cash after costs and overlapping projects are accounted for.", "report_ids": [ "ADP003", "CON001", "ADP005" ], "company_ids": [ "AI-009", "AI-010", "AI-038" ], "company_connection": "SLB’s Lumi, Shell / SLB’s collaboration and Aramco’s internal portfolio expose different roles in the value chain: seller, buyer / partner and operator. Public announcements do not disclose a comparable build-versus-buy allocation.", "counterpoint": "EY’s market is broader than AI software and its value estimate is survey-derived. These figures cannot establish vendor take rates, margins, or the split between internal development and external procurement." }, { "title": "An answer matters more when it reaches someone who can act.", "observation": "IEA’s application review spans subsurface work, maintenance and operating optimization. The technical and consulting literature adds two complementary routes: predictive models and controls acting on physical signals, and language-based tools helping teams retrieve information and coordinate work. IBM’s 2026 survey reports fragmented data as a barrier for 57% of its 100 upstream executives.", "implication": "Follow the whole job: what data is available, what the model recommends, who can act and what changes as a result. A fault prediction is more useful if it reaches the maintenance queue. A drilling suggestion needs to fit how engineers review and approve the work.", "report_ids": [ "INS001", "ADP002", "CON003", "CON008", "INS009" ], "company_ids": [ "AI-013", "AI-201", "AI-008" ], "company_connection": "OQ / SLB’s directional drilling, Chevron / OPX Ai’s Kaybob lift control and Tela’s workflow catalog illustrate different degrees of action and autonomy. Compare those boundaries before comparing ambition.", "counterpoint": "An integration or a large well count does not prove economic uplift. The public SPE gas-lift summary covers more than 1,300 wells but gives no comparable incremental return. Agentic orchestration can also create benefits before a field KPI changes." }, { "title": "People need time and support to change how they work.", "observation": "IBM’s upstream report treats workforce capability and operating redesign as scaling constraints. Accenture reports that 64% of surveyed energy respondents trust colleagues’ insights more than AI, although its public page omits sample details. Havtil / DNV’s review makes responsible use and human involvement part of the petroleum-sector implementation question.", "implication": "Budget for engineers, training and changes to people’s jobs alongside the software. Be clear about what AI can suggest, what it can do, who is responsible and when a person can override it. Counting people trained does not tell you whether they can use it well.", "report_ids": [ "ADP002", "ADP004", "INS004", "ADP010" ], "company_ids": [ "AI-071", "AI-012" ], "company_connection": "ADNOC’s enterprise training and bp / SLB’s remote drilling model are useful places to investigate whether capabilities and decision rights have changed with the technology.", "counterpoint": "These sources do not quantify a causal link between training and returns. Several workforce statistics pool sectors or have incomplete public methods; they do not support an industry-wide headcount-reduction thesis." }, { "title": "A few tools used well may be worth more than a long list of pilots.", "observation": "McKinsey’s bottom-up model assigns nearly half of its identified value to ten use cases, roughly two-thirds to twenty and about 95% to sixty. IBM emphasizes redesigning workflows to scale impact. Together, these point to depth of adoption as a potentially more useful priority than adding another disconnected pilot.", "implication": "Choose a few tasks that could make a meaningful difference. Test them on an asset, then see whether they work on similar assets. Before assuming savings across a portfolio, check the data, how each business operates and who benefits under the contracts.", "report_ids": [ "CON001", "ADP002", "CON002" ], "company_ids": [ "AI-011", "AI-001", "AI-141" ], "company_connection": "Production-system optimization, Chevron’s development planning and Aramco’s process control offer concrete workflow units for a replication test. Their scope and proof differ; none establishes universal transferability.", "counterpoint": "Concentration is a feature of McKinsey’s model, not an observed ranking of realized cash flow. Common platforms may still be prerequisites for many workflows, and asset differences can make replication expensive." }, { "title": "Who can explain where the value came from?", "observation": "The library mixes modeled sector potential, executive survey responses, illustrative margin bridges and company-attributed results. Aramco’s AI-specific Technology Realized Value is a company attribution within a defined benefit framework; IBM’s operating improvements are survey responses. Repeating either claim in several reports does not create independent replication.", "implication": "Keep each important result with its comparison, period, scope, costs and any projects that overlap. Ask whether it changes cash flow and whether the operator, supplier or customer gets the benefit. Saving time can also reduce the work a supplier can bill for.", "report_ids": [ "CON001", "CON005", "ADP001", "ADP004", "INS010" ], "company_ids": [ "AI-037", "AI-038", "AI-141" ], "company_connection": "Aramco’s portfolio, AI contribution to TRV and Fadhili process-control dossier allow readers to move from a corporate total toward operating evidence without treating them as three independent estimates.", "counterpoint": "Public disclosure is incomplete, and portfolio-level attribution can still be useful. An absent public counterfactual does not mean the benefit is zero; it limits how confidently another company can underwrite the same result." }, { "title": "A useful idea can still miss the requirements for funding.", "observation": "National Gas reports £91,153.23 spent on its AIASM Discovery study. Ofgem then declined Alpha funding because surveillance efficiency did not demonstrate the transition-planning outcomes its challenge required.", "implication": "Check what the sponsor is trying to fund, how a study would become a working system and what connecting it to existing systems would cost. A convincing prototype does not answer those questions.", "counterpoint": "One UK regulated-gas innovation award; no technical failure or permanent programme cancellation is established. Study costs are separate from implementation cost and operating return.", "report_ids": [ "GAPR003" ], "company_ids": [], "company_connection": "Kept outside company-deployment counts; the source describes feasibility and an unfunded proposal." } ], "metric_rules": [ { "title": "Annual potential", "text": "McKinsey’s full-potential annual AI value is a scenario, not a realized run rate or a software revenue forecast." }, { "title": "Cumulative digital value", "text": "EY’s 2026–35 value spans a decade and includes digital solutions beyond AI. It cannot be added to McKinsey’s annual estimate." }, { "title": "Illustrative EBIT uplift", "text": "BCG’s segment percentages use different profit baselines and volume units. Normalize economics before ranking assets." }, { "title": "Survey-reported improvement", "text": "IBM’s results describe what executives report. Samples, baselines and questions matter; a response is not a controlled causal estimate." } ], "coverage": { "summary": "A curated collection of 102 publications. Duplicate papers are merged; source populations, methods and publication dates remain separate.", "limits": "This is a broad research collection, not every report in the market. Public access varies: full studies, summaries, abstracts and bibliography leads are identified record by record. Selected high-signal claims received a second Luna review; that review is not an independent audit of the underlying operating results. Private-capital AI returns, comparable build-versus-buy budgets, and sustained project-level cash attribution remain material gaps.", "groups": [ { "group": "Consulting and strategy", "note": "McKinsey, BCG, EY-Parthenon, Bain, Kearney, Oliver Wyman / Quantrue and Roland Berger. Some framework and M&A articles are contextual, not quantified AI studies. BCG’s article and companion deck are one entry." }, { "group": "Adoption and operating models", "note": "IBM IBV, Accenture, Deloitte, PwC, KPMG and Capgemini. Cross-industry and broader-energy samples are labeled; they are not treated as O&G adoption benchmarks." }, { "group": "Institutions and technical evidence", "note": "IEA, Dallas Fed, DNV, Havtil, SPE/JPT and academic reviews. Historical survey evidence, simulation results, restricted full papers and technical literature are distinguished." }, { "group": "Remaining search gaps", "note": "No harmonized global O&G AI adoption panel, comparable PE-portfolio return study, or defensible sector-wide sourcing split was found in this pass. Regional-language and paywalled literature may add evidence." } ] }, "aliases": { "O14C-FORECAST": "GAPE004" } }